Systems of linear inequalities, the graph as the overlap of half-planes, systems whose regions do not meet, systems of three or more inequalities and their bounded regions, absolute value inequalities inside a system, and modelling a pair of constraints.
Subject: Algebra 2 · 74 slides · symbolic lesson
Open the interactive version of this deck
Title
Algebra 2 · Lesson 3.3
Two shadings on one set of axes, and the part they have in common
Objectives
Lesson 2.8 taught you to shade one inequality: draw its boundary, pick a test point off that boundary, and shade the side that works. Nothing about that procedure is being replaced here. You are about to run it twice on the same axes, or three or four times, and then ask one further question.
That further question is the whole lesson. When several conditions must all hold, which points survive all of them at once? The answer is always the part of the page that every shading covers, and it is usually much smaller than any one of them.
There is a reason this matters beyond the drawing. Real problems almost never arrive as one condition. A budget and a deadline, a protein target and a calorie ceiling, a minimum wage bill and a maximum number of hours: each is an inequality, and what you can actually do is the overlap.
Two habits will carry you through every slide. Shade each inequality separately before you look for the overlap, and finish by testing one point inside the region and one point outside it in every inequality you were given.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — Before, Now, Why and the Key Vocabulary
Warm-up
Before anything new, run the old procedure once. Write your answer down before reading on, because the comparison later is only worth something if you committed to something first.
\[ y \le x + 3 \]
Discussion prompt
Name the boundary equation, say whether you draw it dashed or solid and why, then test the origin and say which side gets shaded.
Hold on to the shape of what you just produced: a boundary, a decision about its style, and one test point that settled an entire half of the page. Every inequality in this lesson gets exactly that treatment, no more.
It is worth being clear about how strong that claim is. You checked a single pair of numbers and then shaded half the plane, which is millions of pairs you never tested. The licence for that is the fact that a line separates the plane into exactly two pieces, and the inequality cannot change its verdict without crossing the line.
What is new is only the last step. With two inequalities you will have two boundaries and two shadings, and you will then keep the points that both shadings cover. Everything before that step you already know how to do.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §2.8, p. 132 — the two-step procedure you are building on
Section
Part 1
Concept
A system of linear inequalities is just several linear inequalities in the same two variables, considered together. The book writes them stacked, one above the other, and numbers them so it can talk about them separately.
\[ \begin{array}{ll} x + y \le 8 & \text{Inequality 1} \\ 4x - y > 6 & \text{Inequality 2} \end{array} \]
solution of a system of inequalities — An ordered pair that is a solution of every inequality in the system. The graph of the system is the graph of all its solutions.
Read that definition twice, because the word every is doing all the work. A pair that satisfies one inequality and fails the other is not a partial solution or a near miss; it is simply not a solution, in the same way that a key that turns one of two locks does not open the door.
The book offers a pair to test, and you should test it rather than take it on trust. Notice you substitute the same two numbers into both inequalities: the pair travels together, because the point is a single place on the page.
\[ (5, -2) \]
The vocabulary is worth keeping straight from the start. The system is the list of inequalities; a solution is one ordered pair; and the graph of the system is the whole set of solutions, drawn. Those three things get confused constantly, and they are not the same.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — the definition of a solution of a system
Worked example
This is the book's own claim on page 168, and checking it takes two substitutions. Do them one at a time and write the verdict beside each, because the whole answer hangs on the word and.
\[ \begin{array}{ll} x + y \le 8 & \text{Inequality 1} \\ 4x - y > 6 & \text{Inequality 2} \end{array} \]
Substitute into Inequality 1
Why: put 5 for x and negative 2 for y, and evaluate the left side before comparing.
\[ 5 + (-2) = 3 \le 8 \;\checkmark \]
Three is comfortably at most eight, so the pair clears the first hurdle. Notice how little that tells you on its own: the whole half-plane under the line satisfies this one.
Substitute into Inequality 2
Why: the same two numbers again, this time into the second inequality. Watch the double negative in the subtraction.
\[ 4(5) - (-2) = 20 + 2 = 22 > 6 \;\checkmark \]
Subtracting negative two adds two. That small step is where most errors in this kind of check happen, so it is worth writing the intermediate line rather than doing it in your head.
Check a pair that fails, to see the difference
Why: the origin satisfies the first inequality easily but not the second, so it is not a solution of the system even though it is a solution of one of the inequalities.
\[ (0,0): \; 0 \le 8 \;\checkmark \qquad 4(0) - 0 = 0 > 6 \;\times \]
So the origin lies in one shading and outside the other, and that is exactly the situation the word and rules out. One failure is enough to disqualify a pair, no matter how well it did elsewhere.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — the worked check on (5, minus 2)
Picture it
Here is the system from Example 1 taken apart. The first two pictures are ordinary Lesson 2.8 graphs, each of one inequality on its own; the third keeps only what both of them cover.
\[ \begin{array}{ll} y > -2x - 5 & \text{Inequality 1} \\ y \le x + 3 & \text{Inequality 2} \end{array} \]
Figure (svg): Three copies of the same axes. The first shades everything above a dashed line, the second everything below a solid line, and the third shows only the wedge covered by both shadings.
Work left to right and notice what changes. The first two pictures each cover about half the page. The third covers a wedge, and every point of that wedge had to survive two separate decisions to get there.
This is why the book suggests coloured pencils. If you shade the first inequality red and the second blue, then the graph of the system is whatever came out purple, and you can see it rather than reason about it.
The edges carry information too. The rising edge is solid, so its points are solutions; the falling edge is dashed, so its points are not. The two decisions are independent, and one inequality's style never affects the other's.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — Example 1, steps 1 and 2
Concept
A system is a list of conditions joined by the word and. That single word is what turns two half-planes into one much smaller region, and it is worth naming exactly what it does.
Think of the inequalities as filters in a line, each one letting some points through. A point has to get through all of them to reach the far end, so adding a filter can only ever remove points. Nothing you add later can put back a point an earlier inequality threw out.
That has a useful consequence you can rely on when checking your work. If your shaded region got bigger when you added the third inequality, you have made a mistake, and you have made it in the last step rather than the first.
It also explains why systems so often have no solution while single inequalities almost never do. Every filter you add is another chance for the region to vanish entirely, which is the subject of Part 2.
Concept
One point is easy to check and tells you almost nothing. A system in two variables normally has infinitely many solutions, so listing them is out of the question and a picture is the only honest way to describe them.
\[ \text{graph of the system} = \{(x, y) : \text{every inequality holds}\} \]
That is why the shaded region is treated as the answer here rather than as an illustration of an answer. When a question says graph the system, the shaded region with its correctly styled boundaries is what is being asked for, complete.
It follows that your drawing has to be precise about three things, because each one changes which pairs you are claiming as solutions. Which side of each boundary is shaded, whether each boundary is dashed or solid, and where the boundaries cross.
Reading works in both directions too. Given a shaded region, you should be able to recover the inequalities that produced it, and you will do exactly that in the closing part of this lesson with a triangle from the exercise set.
One more habit that pays off immediately. Before shading anything, glance at the two boundaries and decide whether they cross, are parallel, or are the same line. That glance tells you in advance roughly what shape to expect.
Notation
The stacked layout is compact and easy to misread, so take one apart deliberately. Every symbol in it is either naming a boundary or making a decision about an edge.
Annotate
On: \( \begin{cases} y > -2x - 5 \\ y \le x + 3 \end{cases} \)
That last note is the one worth remembering and the one worth distrusting. Reading the side off the symbol works only when y really is alone on the left with a positive coefficient, and dividing by a negative number reverses the symbol, which is exactly how a shading ends up on the wrong side.
So use the shortcut when the inequality arrives in that form, and use a test point whenever you had to rearrange to get there. A test point costs one substitution and never lies to you.
Notice finally that the two symbols are different in this system. There is no rule saying the boundaries of a system match in style, and a region with one dashed edge and one solid edge is completely ordinary.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — the notation of Example 1
Worked example
Both inequalities are already solved for y, so each boundary is easy to draw and each side is easy to choose. The interesting work is the corner and the checking at the end.
\[ \begin{array}{ll} y > -2x - 5 & \text{Inequality 1} \\ y \le x + 3 & \text{Inequality 2} \end{array} \]
Graph the first boundary and shade above it
Why: the symbol is strictly greater, so the line is dashed and the solutions lie above it.
\[ y = -2x - 5 \quad \text{dashed, shade above} \]
Graph the second boundary and shade below it
Why: this symbol allows equality, so the line is solid, and less than or equal means at or below it.
\[ y = x + 3 \quad \text{solid, shade below} \]
Now find where the two boundaries cross, because that point is the corner of the region and you want it exactly rather than by eye. The boundaries are two equations in two unknowns, which is precisely Lesson 3.2.
Set the two boundary expressions equal
Why: at the crossing point the same pair of numbers satisfies both boundary equations, so the two right-hand sides must agree.
\[ -2x - 5 = x + 3 \]
Solve for x, then for y
Why: add two x to each side and subtract three, then substitute back into either boundary equation.
\[ -8 = 3x \;\Longrightarrow\; x = -\dfrac{8}{3}, \quad y = -\dfrac{8}{3} + 3 = \dfrac13 \]
Figure (svg): The wedge between a dashed line and a solid line, with the open corner where the boundaries cross marked, one green point inside it and two red points outside, one above the solid line and one below the dashed line.
The corner is drawn as an open circle because it sits on the dashed boundary, and the dashed boundary excludes its own points. So the region has a corner it does not contain, which sounds odd and is perfectly normal.
Check a point inside and two points outside
Why: the origin satisfies both, so it is a solution; the two red points each fail a different inequality, which is what makes the wedge the right wedge rather than one of its neighbours.
\[ \begin{array}{l} (0,0): \; 0 > -5 \;\checkmark \; \text{ and } \; 0 \le 3 \;\checkmark \\ (0,5): \; 5 > -5 \;\checkmark \; \text{ but } \; 5 \le 3 \;\times \\ (-1,-6): \; -6 \le 2 \;\checkmark \; \text{ but } \; -6 > -3 \;\times \end{array} \]
Two outside points chosen to fail different inequalities is worth the extra line, because it shows the region is pinned down from both directions rather than by accident. One sits just above the solid boundary, the other just below the dashed one.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — Example 1
Concept
Lesson 2.8 let you settle a whole half-plane with a single test point, and that licence still holds. What changes is how many times you use it, and people routinely use it once when they needed it twice.
\[ \text{two inequalities} \;\Longrightarrow\; \text{two boundaries} \;\Longrightarrow\; \text{two test points} \]
The reason is that each test point answers exactly one question: which side of this particular boundary works. A point that lies on the correct side of the first boundary says nothing whatever about the second, because the two boundaries cut the page differently.
There is a convenient exception. The origin can usually serve as the test point for every inequality in the system, provided it does not lie on any of the boundaries. One substitution per inequality, all with the easy numbers zero and zero.
Then, and only at the end, comes a different kind of test. Once you have the overlap, take a point from inside it and check it in every inequality at once. That is not choosing a side any more; it is confirming that the region you drew is the region you wanted.
Get into the habit of also testing a point just outside the region, across one boundary. If it fails exactly the inequality whose boundary you crossed, your shading is right; if it passes everything, your region is too small.
Sorting
Five pairs, tested against the system from Example 1. Substitute both coordinates into both inequalities before you place any of them, and notice as you go how often a pair passes one test and fails the other.
\[ \begin{array}{l} y > -2x - 5 \\ y \le x + 3 \end{array} \]
Sort into buckets
Sort each ordered pair by how it does against the two inequalities.
The middle bucket is the one that teaches something. Those pairs are the reason the region is a wedge and not a half-plane: they are the points one inequality was happy to keep and the other threw out.
Look also at where the buckets put the pairs on the page. The two solutions sit in the same wedge, the two near-misses sit in the two wedges either side of it, and the failure sits in the wedge opposite. Four wedges, four verdicts, which is the next slide.
Worked example
Same procedure, and this time the wedge opens to the right rather than upward. Notice before you start that the boundaries have slopes of opposite sign, so they certainly cross.
\[ \begin{array}{ll} y \le 3x - 2 & \text{solid boundary} \\ y > -x + 4 & \text{dashed boundary} \end{array} \]
Draw each boundary in its own style
Why: the first allows equality so it is solid; the second is strict so it is dashed.
\[ y = 3x - 2 \quad \text{and} \quad y = -x + 4 \]
Shade below the solid line and above the dashed line
Why: each symbol has y alone on the left with coefficient one, so at most means at or below and greater than means above.
\[ y \le 3x - 2: \text{ below} \qquad y > -x + 4: \text{ above} \]
Find the corner by solving the two boundary equations
Why: set the expressions equal, gather the x terms, and divide.
\[ 3x - 2 = -x + 4 \;\Longrightarrow\; 4x = 6 \;\Longrightarrow\; x = \dfrac32 \]
Substitute back for y
Why: either boundary equation will do, and using both is a free check on the arithmetic.
\[ y = 3\left(\tfrac32\right) - 2 = \tfrac52 \qquad y = -\tfrac32 + 4 = \tfrac52 \]
Figure (svg): A wedge opening to the right, bounded above by a solid line of steep positive slope and below by a dashed falling line, with its open corner and three test points marked.
Notice that the wedge opens to the right rather than upward, and that this was decided entirely by the two symbols. Reverse both of them and you would keep the wedge opening left instead, from the very same pair of boundaries.
The two boundary equations gave the same y, which is the first sign that the corner is right. Now look at the picture: the region is everything to the right of that corner, squeezed between a steep solid line above and a gentle dashed line below.
Check one point inside and two outside
Why: the point (4, 2) satisfies both; (1, 6) breaks the solid boundary and (0, 0) breaks the dashed one, so the wedge is pinned from both sides.
\[ \begin{array}{l} (4,2): \; 2 \le 10 \;\checkmark \; \text{ and } \; 2 > 0 \;\checkmark \\ (1,6): \; 6 \le 1 \;\times \\ (0,0): \; 0 > 4 \;\times \end{array} \]
The corner itself is open, because the dashed boundary passes through it. A pair like (1.5, 2.5) satisfies the first inequality but only equals the second, and equality is not allowed there.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Guided Practice 1
Prediction
Two crossing boundaries cut the page into four wedges, and the shading decisions pick out exactly one of them. Commit to a number before you look at the picture.
Predict first
Two boundaries cross, and each inequality shades one side. How many of the four wedges are the graph of the system?
Correct: Exactly one
Why: Each inequality divides the page in two and keeps one half. A wedge lies on a definite side of each boundary, so it is either kept by both, kept by one, or kept by neither. Exactly one of the four wedges is on the kept side of both, and that wedge is the graph of the system.
Figure (svg): Two crossing lines dividing the plane into four wedges, each labelled with how many of the two inequalities hold there, and only one of them shaded.
The three unshaded wedges are not all alike, and the labels say why. One of them sits on the wrong side of the dashed boundary only, one on the wrong side of the solid boundary only, and one is wrong about both.
That is a useful diagnosis when your answer disagrees with the book. If your region is next door to the right one, exactly one of your shading decisions is backwards; if it is diagonally opposite, both are.
The argument also tells you when this counting fails. It assumed the boundaries cross, so there are four wedges to choose from. Parallel boundaries make three strips instead of four wedges, which changes the answer completely and is the subject of Part 2.
Trap
Asked to graph this system, a very common answer shades everything that got shaded by either inequality.
\[ \begin{array}{l} y > -2x - 5 \\ y \le x + 3 \end{array} \]
It is an easy slip to make with pencil on paper. You shade the first, you shade the second, and what you see at the end is the total amount of pencil on the page.
\[ (0, 5): \; 5 \le 3 \;\times \]
The pair above lies inside that combined shading and fails the second inequality outright, so the combined shading is claiming a pair that is not a solution.
Keep only what both shadings cover. The graph of a system is the overlap, and it is smaller than either piece rather than larger than both.
\[ \text{system} = \text{intersection}, \; \text{not union} \]
The colour trick from the book is the cure. Shade the first in one colour and the second in another, then keep only what came out with both colours on it.
A quick way to catch this error in your own work: your shaded set should never reach past a boundary in the direction that boundary rejects. If pencil appears on both sides of a line, something has gone wrong.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 168 — the KEY CONCEPT box on identifying the common region
Fill the middle
Fill in the two values and read off the two verdicts. Substitute the same pair into both inequalities from Example 1, and evaluate the right-hand side fully before comparing.
Fill in the blanks
\begin-95 1 &>& -2(2) - 5 = ___ \\ 1 &\le& 2 + 3 = ___ \end___
Why: Negative 2 times 2 is negative 4, and negative 4 take away 5 is negative 9, so the first test reads 1 greater than negative 9, which is true. The second reads 1 at most 5, also true. The pair passes both, so (2, 1) is a solution of the system and lies inside the wedge.
Notice what made this easy: you evaluated each right-hand side to a single number and only then compared. Trying to judge the inequality while the arithmetic is half done is where sign slips come from.
Look at how much slack the first test has. One is greater than negative nine by ten whole units, which tells you the pair sits well inside that half-plane rather than near its edge. The second test is tighter, at four units of slack.
You can turn that observation into a rough sense of where a point sits. A pair that only just passes a test is near that boundary; a pair that passes by a wide margin is deep inside. When the margins are small, draw carefully.
Socratic
Definitions with the word every in them are easy to nod at and then quietly relax. Argue this one out instead, in a setting where getting it wrong costs money.
Discussion prompt
A delivery company will take a parcel only if it weighs at most 20 kilograms and its longest side is at most 120 centimetres. A parcel weighs 8 kilograms and is 150 centimetres long. Explain, in terms of the two conditions, why the parcel is refused, and then say what the corresponding statement is about an ordered pair and a system of inequalities.
The parcel example is worth keeping because it makes the asymmetry obvious. No amount of being light compensates for being long, and no amount of clearing one inequality compensates for breaking another.
It also shows why the overlap is usually so much smaller than either condition alone. Most parcels satisfy one of the two limits; far fewer satisfy both; and the ones that satisfy both are the only ones the company will carry.
Try inventing the third condition yourself, a girth limit, say, and ask what it does to the set of acceptable parcels. It can only shrink it, which is the same fact as the filters on the earlier slide.
Section
Part 2
Concept
A system whose boundaries are parallel behaves quite differently from one whose boundaries cross, so you want to notice parallel boundaries early. The reliable way is to put both inequalities in the same form and compare the coefficients of x.
\[ \begin{array}{ll} 2x + 3y < 6 & \text{Inequality 1} \\ y \ge -\tfrac23 x + 4 & \text{Inequality 2} \end{array} \]
The second is already solved for y, so rewrite the first to match. Divide with care, and remember the rule from Lesson 1.6: multiplying or dividing an inequality by a negative number reverses the symbol, while dividing by a positive number leaves it alone.
\[ 2x + 3y < 6 \;\Longrightarrow\; 3y < -2x + 6 \;\Longrightarrow\; y < -\tfrac23 x + 2 \]
Three is positive, so the symbol survives the division unchanged. Now the two inequalities are directly comparable, and they say something you could not see before: both boundaries have slope negative two thirds, so they are parallel and never meet.
With parallel boundaries there is no crossing point, so there is no corner to find and the four-wedge picture from Part 1 does not apply. What matters instead is which way each shading faces, and that is decided by the two symbols.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Example 2, before the graph
Worked example
Both inequalities are now in y-form, and the two boundaries are parallel. Read the symbols and you can predict the outcome before drawing anything.
\[ \begin{array}{l} y < -\tfrac23 x + 2 \\ y \ge -\tfrac23 x + 4 \end{array} \]
Graph the lower boundary dashed and shade below it
Why: the first inequality is strict and asks for y smaller than the lower line, so the shading heads downward, away from the other line.
\[ y = -\tfrac23 x + 2 \quad \text{dashed, shade below} \]
Graph the upper boundary solid and shade above it
Why: the second allows equality and asks for y at least the upper line, so this shading heads upward, also away from the first.
\[ y = -\tfrac23 x + 4 \quad \text{solid, shade above} \]
Figure (svg): Two parallel boundaries of equal negative slope. Everything below the lower dashed line is shaded one colour, everything above the upper solid line another, and the strip between them is left empty and labelled.
There is the answer, in a picture. The two shadings point in opposite directions from two lines that never meet, so no point on the page carries both colours, and the strip between the boundaries carries neither.
You do not have to rely on the drawing, though, and you should not. The two inequalities can be combined into a single claim, and that claim is visibly false.
Chain the two inequalities and read the contradiction
Why: if y is at least the upper expression and also less than the lower one, then the upper expression is less than the lower one; cancelling the matching x terms leaves a statement about numbers alone.
\[ -\tfrac23 x + 4 \le y < -\tfrac23 x + 2 \;\Longrightarrow\; 4 < 2 \]
Check by trying to produce a single solution
Why: pick any x you like, say 3, and work out what each inequality demands of y; the two demands do not overlap, so no pair exists and the system has no solution.
\[ x = 3: \quad y < 0 \quad \text{and} \quad y \ge 2 \]
The book stops at the picture, and the picture is correct. The chained line is worth adding because it turns no region appeared into a proof, and it works even when a sketch is too rough to trust.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Example 2
Concept
When the shadings miss each other, the honest answer to graph this system is a drawing with nothing shaded and the sentence the system has no solution. That is a complete answer, and it is worth saying out loud because it feels like an error.
\[ \text{solution set} = \varnothing \]
You have met this before in one variable. Asking for a number that is at least five and at most one is asking for nothing, and in Lesson 1.6 you wrote exactly that. Two variables change nothing about the logic; they only change the picture from a number line to a plane.
There is a second reason to be comfortable with an empty answer. In a modelling problem it is often the most informative outcome: it means the constraints you were handed cannot all be met, and somebody has to relax one of them.
So do not go hunting for a region that is not there, and do not quietly bend a shading to make one appear. Report the emptiness, and if you can, say which pair of inequalities is responsible for it.
Being able to prove it matters too. A sketch showing no overlap might just be a bad sketch, whereas the chained inequality from Example 2 reduced the whole question to whether 4 is less than 2, which settles it for good.
Picture it
Here is Example 2 again, reduced to a single value of x. Fix x at 3 and ask each inequality what it wants from y; the answer is two ranges that miss each other.
\[ x = 3: \quad 2(3) + 3y < 6 \;\Longrightarrow\; y < 0 \qquad y \ge -\tfrac23(3) + 4 = 2 \]
Figure (svg): A vertical number line for y at x equal to 3, with an arrow running down from an open circle at 0 and another running up from a filled circle at 2, and the interval from 0 to 2 marked as a gap that belongs to neither.
The open circle at zero records the strict inequality: y has to be below zero and is not allowed to equal it. The filled circle at two records the other symbol, which does allow equality.
Now notice the gap between them, from zero up to two. Any y in that gap is too big for the first inequality and too small for the second, so the column above x equals 3 contains no solutions at all.
The last step of the argument is the one worth keeping. Because the boundaries are parallel, the two ranges slide down together as x increases and the gap between them stays exactly two units wide forever. Every vertical line tells the same story, so the whole plane is empty of solutions.
Worked example
This pair looks nothing like Example 2 until you tidy it, and then it turns out to be the same situation in disguise. Clear the fraction first, because a fraction hides the resemblance.
\[ \begin{array}{ll} 2x - \tfrac12 y \ge 4 & \text{Inequality 1} \\ 4x - y \le 5 & \text{Inequality 2} \end{array} \]
Multiply the first inequality by 2
Why: two is positive, so the symbol is unchanged; doubling clears the one half and makes the left side match the second inequality exactly.
\[ 2\left(2x - \tfrac12 y\right) \ge 2(4) \;\Longrightarrow\; 4x - y \ge 8 \]
Now read the two lines together. The same expression, four x minus y, is required to be at least eight by one inequality and at most five by the other. No number is both.
\[ \begin{array}{ll} 4x - y \ge 8 & \text{rewritten Inequality 1} \\ 4x - y \le 5 & \text{Inequality 2} \end{array} \]
Figure (svg): A horizontal number line for the value of four x minus y, with a filled circle at 8 and an arrow running right, a filled circle at 5 and an arrow running left, and the interval from 5 to 8 shaded as a gap.
The picture is a number line rather than a coordinate plane, and that is deliberate. Once both inequalities are about the same expression, the two variables stop mattering and the whole question becomes whether one number can sit in two places at once.
Confirm that the boundaries are parallel
Why: solve each boundary equation for y and compare: the same slope, different intercepts, so the lines never meet.
\[ y = 4x - 8 \qquad y = 4x - 5 \]
Check by testing a vertical line
Why: take x equal to 3; the first inequality then needs y at most 4 and the second needs y at least 7, which cannot both happen, so the system has no solution.
\[ x = 3: \quad 12 - y \ge 8 \;\Rightarrow\; y \le 4, \qquad 12 - y \le 5 \;\Rightarrow\; y \ge 7 \]
Tidying an inequality before graphing it is not tidiness for its own sake. Here it turned a pair of unrelated-looking conditions into two bounds on one expression, and the contradiction became impossible to miss.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Guided Practice 2
Picture it
Parallel boundaries do not always mean no solution. Which way each shading faces decides the outcome, and there are three genuinely different things that can happen.
Figure (svg): Three copies of the same axes, each with two parallel boundaries. The first has shadings facing apart and nothing in common, the second a shaded strip between the lines, and the third a shaded half-plane above the upper line.
In the first picture the shadings face outward and share nothing, which is Example 2. In the second they face inward and share the strip between the lines, an infinite region between two parallel edges.
The third is the case people forget. Both shadings face the same way, so one of them sits entirely inside the other, and the overlap is simply the smaller of the two half-planes. The stricter inequality wins and the other one does no work at all.
So parallel boundaries are a signal to look at the symbols, not a verdict. Two arrows pointing apart give nothing, two pointing together give a strip, and two pointing the same way give a half-plane.
A fourth case exists as a curiosity, when the two boundaries are the same line. Then the region is the line itself if both inequalities allow equality and face opposite ways, and nothing at all if either of them is strict.
Sorting
Six systems, all with parallel boundaries. For each one, rewrite in y-form if you need to, then ask which way each shading faces before you place it.
Sort into buckets
Sort each system by what its graph turns out to be.
The pairs with horizontal boundaries, the ones about y alone, are the easiest place to see the pattern, and they are the ones you should reason from. A number cannot be above five and below one; it can easily be above one and below five.
The last bucket is the one that changes how you check your work. If one of your inequalities turns out to be redundant, your region will have one fewer edge than the number of inequalities you were given, and that is not a mistake.
Error analysis
Here is a piece of reasoning that gets the right answer on Example 2 and the wrong answer on half the systems in the exercise set. Read it looking for the step that does not follow.
Annotate
On: \( \begin{array}{l} y \ge x - 1 \\ y \le x + 1 \end{array} \quad \Longrightarrow \quad \text{no solution?} \)
The mistake is a real one and it is seductive because it uses a true fact. The lines really do never meet. What was smuggled in is the idea that the region lives near where the boundaries cross, when in fact the region is made of half-planes.
The cheapest possible guard against this is a single test point. Try the origin in both inequalities before you announce that a system is empty; if the origin passes both, the system plainly has solutions and your reasoning was wrong.
Notice also how little it costs to be wrong in the other direction. Claiming a strip where there is none is caught by the same test point, since it will fail one of the inequalities.
Counterexample
The strongest way to kill a false rule is to construct something it forbids. Do that here, deliberately, and then look at what your construction has in common with the strip from the previous slide.
Discussion prompt
Write a system of two linear inequalities whose boundaries are parallel and which has infinitely many solutions, then a second system, also with parallel boundaries, whose graph is one whole half-plane. Give a solution of each and say how you chose the symbols.
Compare the two answers and notice that the boundaries are identical in both; only the symbols changed. That is the point being made: parallel boundaries fix the geometry of the edges and leave the outcome entirely to the symbols.
If you want a third example for your own notes, make the two boundaries the same line with opposite symbols that both allow equality. The solution set is then the line itself, which is neither empty nor a region, and is a fair answer to a fair question.
Building counterexamples like this is a habit worth carrying into every topic. A rule you can break on demand is a rule you understand; a rule you can only recite is a rule that will fail you on an unfamiliar system.
Prediction
Two horizontal boundaries this time, which makes the reasoning almost purely arithmetic. Decide before you sketch, and be ready to say what the picture would look like.
\[ \begin{array}{l} y \ge 5 \\ y \le 1 \end{array} \]
Predict first
What is the graph of this system?
Correct: No points at all
Why: The first inequality asks y to be at least 5 and the second asks y to be at most 1. No number is both at least 5 and at most 1, so no ordered pair satisfies the system and nothing is shaded. The two shadings face away from each other, upward from the higher line and downward from the lower one.
The tempting wrong answer is the strip, because a strip is what you get from the same two boundaries with the symbols the other way round. Swap them and the region becomes every point with y between one and five, which is very much not nothing.
Notice that x never appeared in either inequality, and that this changes nothing about the method. A missing variable is an unrestricted variable, exactly as in Lesson 2.8, so each boundary is a full horizontal line and each shading a full half-plane.
A sanity check you can run in one second: try the pair with y equal to 3, or y equal to 0, or y equal to 9, and watch each one fail one of the two inequalities. If every candidate you can think of fails, the empty answer is the right one.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 171 — Exercise 6
Section
Part 3
Concept
Nothing in the two-step procedure mentioned the number two. Graph every inequality, then keep what all of them cover: the method reads the same for two, three or ten conditions, and the only thing that changes is how much is left at the end.
\[ \text{three inequalities} \;\Longrightarrow\; \text{three boundaries, three shadings, one common region} \]
Because each inequality is a filter, the region shrinks as the list grows. Two inequalities with crossing boundaries give a wedge, which runs off the page forever in one direction. A third inequality can cut that wedge off and close it.
bounded region — A region you could enclose in a large enough circle. Its boundary is a closed loop of straight pieces, and it has a finite number of corners, called vertices.
An unbounded region is one you could not: it runs off the page in at least one direction, however far you extend the axes. A half-plane is unbounded, and so is a strip, and so is the wedge from Example 1.
Roughly speaking, closing a region takes at least three inequalities, because you need a boundary facing you from every direction. Three can just manage it, as a triangle; four often give a quadrilateral. But having three inequalities does not guarantee a bounded region, as you will see two slides from now.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 170 — SYSTEMS OF THREE OR MORE INEQUALITIES
Worked example
Three inequalities, every one of them strict. Before drawing, notice that two of the boundaries are as simple as a boundary gets, a vertical line and a horizontal line, and only the third is slanted.
\[ \begin{array}{ll} x < 6 & \text{Inequality 1} \\ y > -1 & \text{Inequality 2} \\ y < x & \text{Inequality 3} \end{array} \]
Draw all three boundaries dashed
Why: every symbol here is strict, so no boundary contributes its own points and all three are drawn dashed.
\[ x = 6, \quad y = -1, \quad y = x \]
Shade each one
Why: left of the vertical line, above the horizontal line, and below the slanted line, where below y equals x means the side containing the point (1, 0).
\[ (1,0): \; 1 < 6 \;\checkmark, \quad 0 > -1 \;\checkmark, \quad 0 < 1 \;\checkmark \]
Now find the corners, and find them by algebra rather than by squinting. Each corner is where two of the boundaries meet, so take the boundaries two at a time and solve the resulting pair of equations.
Corner from the slanted and horizontal boundaries
Why: substitute y equal to negative 1 into y equals x.
\[ y = x \; \text{and} \; y = -1 \;\Longrightarrow\; (-1, -1) \]
Corner from the vertical and horizontal boundaries
Why: both coordinates are handed to you directly.
\[ x = 6 \; \text{and} \; y = -1 \;\Longrightarrow\; (6, -1) \]
Corner from the vertical and slanted boundaries
Why: substitute x equal to 6 into y equals x.
\[ x = 6 \; \text{and} \; y = x \;\Longrightarrow\; (6, 6) \]
Figure (svg): A triangle with all three sides dashed, bounded by a vertical line at 6, a horizontal line at minus 1 and the line y equals x, with its three open corners labelled and three test points marked.
All three corners are drawn open, because each one lies on two dashed boundaries. The region is the inside of the triangle and nothing else, which is why the exercise set calls it an open triangle.
Check one point inside and two outside
Why: the pair (2, 0) satisfies all three; (7, 2) breaks only the vertical boundary and (3, negative 2) breaks only the horizontal one, so each side of the triangle is doing real work.
\[ \begin{array}{l} (2,0): \; 2 < 6 \;\checkmark, \; 0 > -1 \;\checkmark, \; 0 < 2 \;\checkmark \\ (7,2): \; 7 < 6 \;\times \\ (3,-2): \; -2 > -1 \;\times \end{array} \]
Testing an outside point just across each edge is the cheapest way to confirm you shaded the right side of every boundary, and it takes three substitutions.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 171 — Exercise 17
Concept
Finding a vertex of a region has nothing to do with inequalities. Two boundaries meet at a single point, and that point is the solution of the two boundary equations, which is exactly the problem Lessons 3.1 and 3.2 were about.
\[ \begin{array}{l} \text{boundary 1: } a_1 x + b_1 y = c_1 \\ \text{boundary 2: } a_2 x + b_2 y = c_2 \end{array} \]
So use substitution when one equation already has a variable alone, and elimination when the coefficients line up conveniently. Nothing about the surrounding inequalities changes the method; you simply replace each inequality symbol by an equals sign for the duration of that calculation.
Two warnings come with it. First, not every pair of boundaries gives a vertex of the region: two boundaries may cross at a point that some third inequality rejects, in which case the crossing exists but the corner does not belong to the region.
Second, whether a corner belongs to the region depends on the symbols, not on the algebra. If either of the two boundaries through it is dashed, the corner is excluded, and you record that with an open circle.
Corners are worth this trouble for a reason you will meet next year. In linear programming the largest or smallest value of a quantity over a region like this always occurs at a vertex, so the whole method reduces to listing the corners and testing them.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.2, p. 160 — the substitution and elimination methods used here
Fill the middle
Exercise 24 gives four inequalities whose boundaries make a parallelogram. Find the corner where two particular boundaries meet, using elimination exactly as in Lesson 3.2.
\[ \begin{array}{l} x + y < 5, \quad x + y > -5 \\ x - y < 4, \quad x - y > -2 \end{array} \]
Fill in the blanks
\begin94.5 x + y &=& 5 \\ x - y &=& 4 \\ 2x &=& 0.5 \\ x &=& ___ \\ y &=& ___ \end___
Why: Adding the two equations cancels the y terms and leaves 2x equal to 9, so x is four and a half. Substituting back into the first equation gives y equal to 5 take away four and a half, which is one half. The corner is at four and a half, one half, and it is open because both of those boundaries come from strict inequalities.
The other three corners come the same way, from the other three pairs of boundaries: the two sums with the two differences, taken in all four combinations. Each pair takes one line of elimination.
Notice that you never paired the two sums with each other. Those two boundaries are parallel, so they do not meet, and attempting the elimination would give you the false statement that 5 equals negative 5, which is Lesson 3.2 telling you there is no crossing point.
Corners with halves in them are perfectly normal and are worth leaving as exact fractions rather than rounding. A corner is a precise point, and writing it as four point five is fine while writing it as about four is not.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 24
Worked example
This system is mine rather than the book's, written so that every corner is a whole number and so that you meet the shape that matters most in applications. Two of its inequalities are the ones a real problem almost always adds silently.
\[ \begin{array}{ll} x \ge 0 & \text{stay right of the vertical axis} \\ y \ge 0 & \text{stay above the horizontal axis} \\ 2x + 3y \le 12 & \text{a budget}\end{array} \]
Draw the two axis boundaries solid and shade inward
Why: both inequalities allow equality, so the axes themselves belong to the region, and together they restrict everything to the first quadrant.
\[ x = 0 \quad \text{and} \quad y = 0 \]
Find the intercepts of the third boundary
Why: set y to zero and then x to zero in the boundary equation, which is the quickest way to draw a line in standard form.
\[ 2x = 12 \;\Rightarrow\; x = 6, \qquad 3y = 12 \;\Rightarrow\; y = 4 \]
Shade the side of the third boundary containing the origin
Why: the origin makes the left side zero, and zero is at most twelve, so the origin's side is the shaded one.
\[ (0,0): \; 2(0) + 3(0) = 0 \le 12 \;\checkmark \]
Figure (svg): A right triangle in the first quadrant bounded by the two axes and the line two x plus three y equals twelve, with corners at the origin, at six on the horizontal axis and at four on the vertical axis.
Three solid boundaries, so this region owns its edges and its corners, and all three corners are whole-number points. The two axes supply two of them for free and the intercepts supply the third.
The corners are where two boundaries meet, as always. The origin comes from the two axes; the point six, zero comes from the horizontal axis and the budget line; the point zero, four comes from the vertical axis and the budget line.
Check one point inside and one outside
Why: the pair (1, 1) satisfies all three inequalities, while (5, 1) stays in the first quadrant but spends thirteen when only twelve was available.
\[ \begin{array}{l} (1,1): \; 2 + 3 = 5 \le 12 \;\checkmark \\ (5,1): \; 10 + 3 = 13 \le 12 \;\times \end{array} \]
Read the budget line as a price list if it helps: each unit of x costs 2 and each unit of y costs 3, and the region is every purchase you can afford with 12. The corner six, zero is spending it all on x.
Ranking
Five steps, deliberately shuffled. Two of them can feel interchangeable, so as you order them ask what each step needs to already exist before it can be carried out.
Put in order
Put the five steps of graphing a system of inequalities into the order you would actually carry them out.
Why: You cannot draw a boundary until you know its equation, and you cannot shade a side until the boundary is on the page. Only once every shading exists can you look for what they all cover, and the final test is a check on the finished region, so it has to come last of all.
The step most often skipped is the rewriting, and skipping it is what produces shadings on the wrong side. An inequality in standard form does not tell you which way to shade until you either solve it for y or test a point.
The step most often omitted altogether is the last one. It is the only step that can catch an error, and it costs one substitution per inequality, so it is the best value in the whole list.
Notice also that drawing all the boundaries before shading any of them is worth doing in that order on paper. The crossings tell you where the corners will be, and seeing them first makes it much harder to shade a wrong side without noticing.
Worked example
Three inequalities again, but this time the region does not close, and one of the three turns out to be doing no work at all. Rewrite the slanted ones for y before you shade.
\[ \begin{array}{ll} x \ge 2 & \text{Inequality 1} \\ -3x + y < -1 & \text{Inequality 2} \\ 4x + 3y < 12 & \text{Inequality 3} \end{array} \]
Solve Inequality 2 for y
Why: add three x to each side; the coefficient of y is already positive one, so nothing reverses.
\[ -3x + y < -1 \;\Longrightarrow\; y < 3x - 1 \]
Solve Inequality 3 for y
Why: subtract four x, then divide by the positive number 3, so again the symbol survives.
\[ 4x + 3y < 12 \;\Longrightarrow\; y < \dfrac{12 - 4x}{3} \]
Both slanted inequalities put a ceiling on y, and nothing anywhere puts a floor on it. That is worth noticing before you draw: a region with two ceilings and no floor cannot possibly be bounded.
Find the one corner
Why: the vertical boundary meets the lower of the two ceilings; substitute x equal to 2 into the third boundary equation.
\[ 4(2) + 3y = 12 \;\Longrightarrow\; 3y = 4 \;\Longrightarrow\; y = \dfrac43 \]
Figure (svg): A region to the right of a solid vertical line at 2 and below a dashed falling line, shaded downward with arrows showing it never closes, and a third faint boundary drawn well away from it.
Compare the two ceilings at x equal to 2. One allows y below four thirds, the other allows y below five, so the first is far stricter, and as x grows the difference only widens because one ceiling rises while the other falls.
Show that the second inequality is redundant
Why: for every x at least 2 the third inequality's ceiling is the lower one, so any pair satisfying the third automatically satisfies the second.
\[ \dfrac{12 - 4x}{3} < 3x - 1 \;\Longleftrightarrow\; 12 - 4x < 9x - 3 \;\Longleftrightarrow\; x > \dfrac{15}{13} \]
Check one point inside and one outside
Why: the pair (2, 1) satisfies all three, while (3, 0) satisfies the first two and spends twelve exactly on the third, which the strict symbol refuses.
\[ \begin{array}{l} (2,1): \; 2 \ge 2 \;\checkmark, \; 1 < 5 \;\checkmark, \; 8 + 3 = 11 < 12 \;\checkmark \\ (3,0): \; 12 + 0 = 12 < 12 \;\times \end{array} \]
Fifteen thirteenths is a little over one, and the region only exists for x at least two, so the redundancy holds everywhere that matters. Three inequalities, but the picture has only two edges.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 171 — Exercise 21
Prediction
Deciding this before you draw saves real time, and the reasoning is short. Ask whether every direction you could walk in is eventually blocked by a boundary.
Predict first
Which of these systems has a bounded region, one you could enclose in a circle?
Correct: x ≥ 0 and y ≥ 0 and x + y ≤ 5
Why: Only the system with three inequalities closes. Walking right along the horizontal axis you are eventually stopped by the sum being at most 5, walking left you are stopped by x being at least 0, and walking down you are stopped by y being at least 0, so the region is the triangle with corners at the origin, at 5 on one axis and at 5 on the other. Each of the other systems leaves at least one direction open forever.
The two-inequality systems are quick to dismiss once you look for an escape route. The first quadrant on its own runs away up and to the right; the strip between two parallel lines runs away along its own length; and the fourth system runs away down and to the left.
Counting inequalities is not the test, though it is a useful first glance. Exercise 21 had three inequalities and no floor at all, so it stayed unbounded, and you can build systems with five inequalities that are still open in one direction.
The test that actually works is to walk. Pick a direction, ask which inequality eventually stops you, and if you can find a direction where nothing does, the region is unbounded. In a modelling problem that usually means a constraint has been left out.
Matching
Corners come in pairs of boundaries, and matching them is the skill that makes a region describable. Three of these come from Exercise 17 and one from the budget triangle.
Match the pairs
Why: A horizontal boundary hands you the y coordinate and a vertical one hands you the x coordinate, so those pairings are read off directly. When a slanted boundary is involved you substitute the known coordinate into its equation: y equals x with y equal to negative 1 gives x equal to negative 1, x equal to 6 in y equals x gives y equal to 6, and x equal to 0 in the budget equation gives three y equal to twelve, so y is 4.
Notice how much easier the vertical and horizontal boundaries make life. Each one fixes one coordinate outright, which is why a system built from axis-parallel boundaries and one slanted line has corners you can find almost without writing anything down.
The general case is no harder, only longer. Two slanted boundaries need a genuine substitution or elimination, exactly the work you did on the parallelogram corner two slides ago.
One habit to carry away: when you list the corners of a region, write beside each one the two boundaries that produced it. That record is what lets you check later that every corner really belongs to the region and that none has been invented.
Real world
A region full of possibilities usually comes with a question about the best one, and it turns out you almost never have to search the whole region. Think about why before reading the answer.
Discussion prompt
A workshop can make x tables and y chairs, and its constraints give exactly the budget triangle from earlier: x at least 0, y at least 0, and 2x plus 3y at most 12, where the numbers are hours of machine time. Each table earns 40 dollars of profit and each chair 30. Work out the profit at each of the three corners, and explain why you do not need to test any point in the middle.
Notice that the reasoning did not depend on the particular profits. Whatever the two numbers are, moving in the direction that improves the total keeps improving it until a boundary stops you, and boundaries stop you at edges and edges end at corners.
There is one interesting exception, and it is worth knowing about. If the profit happens to be constant along an entire edge, then every point of that edge is equally good, and the two corners at its ends are still among the best answers. So testing corners never loses.
This also explains why an unbounded region is a warning sign in a modelling problem. If the region runs off forever in a profitable direction, the model is claiming you can earn without limit, which means a real constraint has been left out.
Section
Part 4
Concept
Go back and read the two steps: graph each inequality, then keep what all the graphs cover. Neither step mentions a line. You have already graphed an absolute value inequality on its own in Lesson 2.8, and putting one into a system changes nothing about the method.
\[ y > \left|x + 4\right| \]
What does change is the shape of the boundary. The graph of the absolute value of x plus 4 is a V with its point at negative four on the horizontal axis, made of two straight arms of slope one and negative one. So the boundary comes with a corner of its own, before any other inequality is involved.
The vocabulary carries over unchanged. Strict symbols still give a dashed boundary, and that includes the point of the V, which is why you will draw open circles at some very natural-looking corners. Symbols allowing equality still give a solid boundary.
The shading decision carries over too. With y alone on the left, greater than means the side where y is larger, which for a V opening upward is the region inside the V, and less than means the region below it, outside.
One new possibility appears, and it is the reason this section exists. A V faces you from two directions at once, so a line and a V together can enclose a region, which two straight boundaries never could.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — the Review note pointing back to absolute value graphs
Notation
Before you graph anything, take the boundary expression apart. Everything you need in order to draw the V is visible in it, and none of it needs a table of values.
Annotate
On: \( y > \left|x + 4\right| \)
The middle note is the one to slow down on. A plus sign inside the bars moves the V in the negative direction, and a minus sign moves it in the positive direction. If that feels backwards, it is: the safe test is to find the value of x making the inside zero.
It is worth reading off the vertex as a pair rather than a single number, because you will need both coordinates when you list the region's corners. Here the inside is zero at x equal to negative four and the absolute value is then zero, so the vertex is negative four, zero.
Nothing here is special to absolute value either. Any boundary you can graph can go into a system, and in later chapters you will put parabolas and circles into exactly this procedure.
OpenStax Algebra and Trigonometry 2e, §11.3 Systems of Nonlinear Equations and Inequalities: Two Variables §11.3 — systems of inequalities whose boundaries are curves
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Example 3
Picture it
Example 3 pairs a horizontal line with a V. Here is each inequality shaded on its own, before any overlap is taken, so that you can see which half of the page each one is claiming.
\[ \begin{array}{ll} y \le 3 & \text{Inequality 1} \\ y > \left|x + 4\right| & \text{Inequality 2} \end{array} \]
Figure (svg): Two copies of the same axes. The first shades everything at or below a solid horizontal line at three; the second shades everything strictly above a dashed V whose point sits at minus four on the horizontal axis.
The first is as simple as an inequality gets: a solid horizontal line and everything under it, an unbounded half-plane with a flat lid. On its own it goes on forever left, right and downward.
The second is unbounded too, but in a different direction: it is the inside of the V, opening upward forever, and its two arms pinch together at the point negative four, zero, which the dashed boundary excludes.
Now imagine laying the second picture on top of the first. The V opens upward and the lid stops it at height three, so what survives is a triangle, and that is the whole content of Example 3.
Notice that both regions are unbounded on their own and their overlap is bounded. That is exactly what Part 3 said about filters: the overlap can be much smaller than either piece, and here it is finite when neither piece was.
Worked example
A line and a V, and the region they leave between them. Take the V's vertex and the two crossing points before you draw, so that the triangle is placed by algebra rather than by eye.
\[ \begin{array}{ll} y \le 3 & \text{solid boundary} \\ y > \left|x + 4\right| & \text{dashed boundary} \end{array} \]
Locate the point of the V
Why: the inside of the bars is zero when x is negative 4, and the absolute value is then 0.
\[ x + 4 = 0 \;\Longrightarrow\; x = -4, \quad y = 0 \]
Find where the V meets the ceiling
Why: set the boundary expressions equal; the absolute value equation splits into two cases, one for each arm.
\[ \left|x + 4\right| = 3 \;\Longrightarrow\; x + 4 = 3 \;\text{ or }\; x + 4 = -3 \]
Solve both cases
Why: subtract 4 from each, giving the two ends of the triangle's top edge.
\[ x = -1 \quad \text{or} \quad x = -7 \]
Shade inside the V and below the line
Why: test the point negative four, two, which lies above the vertex and under the ceiling.
\[ (-4,2): \; 2 \le 3 \;\checkmark, \quad 2 > \left|0\right| = 0 \;\checkmark \]
Figure (svg): A triangle with a solid top edge at height three and two dashed sides coming down to a point at minus four on the horizontal axis, with its three open corners labelled and three test points marked.
So the region is a triangle with a solid top edge and two dashed sides. All three corners are drawn open, because each one sits on the dashed V, and the corner at the bottom is the V's own vertex.
It is worth saying why the top corners are excluded even though the top edge is solid. A corner has to satisfy both inequalities, and each top corner lies exactly on the V, which the strict symbol rejects. One failure is enough.
Check one point inside and two outside
Why: the pair negative four, two satisfies both; negative four, four breaks the ceiling and zero, two lies outside the V, so the triangle is pinned from above and from the side.
\[ \begin{array}{l} (-4,4): \; 4 \le 3 \;\times \\ (0,2): \; 2 > \left|4\right| = 4 \;\times \end{array} \]
The second failure is the instructive one. The pair zero, two is comfortably under the ceiling and still fails, because being inside the V is a separate condition that has nothing to do with height alone.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Example 3
Intuition
In Part 1 every system of two straight inequalities gave a wedge or a strip or a half-plane, and every one of those runs off the page forever. Two straight boundaries simply cannot enclose anything.
\[ \text{two lines} \;\Longrightarrow\; \text{at most a wedge} \qquad \text{a line and a V} \;\Longrightarrow\; \text{a triangle} \]
The reason is that a line blocks you from one direction only. Two lines block two directions and leave the other two open, which is why a wedge always has an open mouth.
A V blocks you from two directions at once, because it has two arms pointing different ways. Add a line across its mouth and every direction is covered, so the region closes into a triangle with three corners.
That is why Example 3 and Guided Practice 4 can produce bounded regions from only two inequalities, while Exercise 17 needed three. Counting inequalities is a poor guide to whether a region is bounded; counting the directions that are blocked is the right question.
You can push the idea further. Two V's facing each other block four directions and close a four-cornered region with no straight boundaries at all, which is exactly Exercise 30 later in this part.
Worked example
The same pairing as Example 3, with one important difference in the symbols: both of them allow equality this time, so the region will own every one of its edges and corners.
\[ \begin{array}{ll} y \le 4 & \text{solid boundary} \\ y \ge \left|x - 5\right| & \text{solid boundary} \end{array} \]
Locate the point of the V
Why: the inside of the bars is zero when x is 5, and a minus inside shifts the V to the right.
\[ x - 5 = 0 \;\Longrightarrow\; x = 5, \quad y = 0 \]
Find where the V meets the ceiling
Why: set the absolute value equal to 4 and split into the two arms.
\[ \left|x - 5\right| = 4 \;\Longrightarrow\; x - 5 = 4 \;\text{ or }\; x - 5 = -4 \]
Solve both cases
Why: add 5 to each, giving the two ends of the top edge.
\[ x = 9 \quad \text{or} \quad x = 1 \]
Figure (svg): A triangle with a solid horizontal top edge at height four and two solid sides descending to a point at five on the horizontal axis, with all three corners filled in and two test points marked.
Every boundary is solid here, so all three corners are filled in and belong to the region. Compare that with Example 3, where the identical shape had three open corners, and notice that only the symbols were different.
The region is the closed triangle with vertices five, zero and one, four and nine, four. Being able to name a region that precisely is worth aiming for: it is a complete answer with no picture required.
Check one point inside and one outside
Why: the pair five, two sits above the vertex and below the ceiling, while one, one is under the ceiling and below the left arm of the V.
\[ \begin{array}{l} (5,2): \; 2 \le 4 \;\checkmark, \; 2 \ge \left|5-5\right| = 0 \;\checkmark \\ (1,1): \; 1 \le 4 \;\checkmark, \; 1 \ge \left|1-5\right| = 4 \;\times \end{array} \]
Notice that the absolute value is always evaluated at the point's own x. Using the wrong x there, or forgetting to take the absolute value at all, is the most common slip in this whole topic, and writing the substitution out stops it.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Guided Practice 4
Trap
Graphing the second inequality of Example 3, a very common answer shades everything above the left arm of the V and calls it done.
\[ y > \left|x + 4\right| \]
The reasoning sounds fine: greater than means above, and the left arm is a line, so shade above it.
\[ (0, 2) \]
But the pair above lies well above the left arm extended, and it is not a solution: at x equal to zero the V is four units high.
The boundary is the whole V, not one of its arms, so above means above both arms at once, which is the region inside the V.
\[ (0,2): \; 2 > \left|4\right| = 4 \;\times \]
One test point settles it properly. Try the point above the vertex, negative four, five: five is greater than zero, so the inside of the V is the shaded set.
\[ (-4,5): \; 5 > 0 \;\checkmark \]
The cure is to evaluate the absolute value at the test point's own x rather than thinking about arms. That turns a shape question back into the arithmetic you can already do.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Example 3, step 1
Worked example
This time the absolute value carries a minus sign in front, which turns the V upside down, and the second inequality provides a floor rather than a ceiling. Expect the triangle to point downward.
\[ \begin{array}{ll} y > -2 & \text{dashed boundary} \\ y \le -\left|x + 2\right| & \text{solid boundary} \end{array} \]
Locate the point of the upside-down V
Why: the inside is zero at x equal to negative 2, and the minus in front makes every other value of the right side negative, so the V opens downward from its highest point.
\[ x + 2 = 0 \;\Longrightarrow\; x = -2, \quad y = 0 \]
Find where the V meets the floor
Why: set the boundary expressions equal and split the absolute value into its two cases.
\[ -\left|x + 2\right| = -2 \;\Longrightarrow\; \left|x + 2\right| = 2 \;\Longrightarrow\; x = 0 \;\text{ or }\; x = -4 \]
Shade below the V and above the floor
Why: test the point negative two, negative one, which sits just under the vertex.
\[ (-2,-1): \; -1 > -2 \;\checkmark, \quad -1 \le -\left|0\right| = 0 \;\checkmark \]
Figure (svg): A downward pointing triangle whose apex sits at minus two on the horizontal axis, bounded below by a dashed horizontal line at minus two, with the apex filled in and the two lower corners open.
The apex at negative two, zero is filled in, because the inequality that owns that boundary allows equality. The two lower corners are open, because they sit on the dashed floor.
So one region can have some corners included and others not, and the rule is always the same: a corner belongs to the region only if every inequality accepts it, and a dashed boundary through it means one does not.
Check one point inside and one outside
Why: negative three, negative one is inside, since the V is one unit below the axis there; zero, negative one is above the floor but the V has already dropped to negative two at that x.
\[ \begin{array}{l} (-3,-1): \; -1 \le -\left|-1\right| = -1 \;\checkmark \\ (0,-1): \; -1 \le -\left|2\right| = -2 \;\times \end{array} \]
The first of those sits exactly on the V and is still a solution, because that boundary is solid. Points on a solid boundary are the easiest check you can run on whether you drew the right style.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 169 — Guided Practice 5
Worked example
No straight boundaries at all in this one, and yet it closes a four-cornered region. Both V's have their point at x equal to 3, one opening down and one opening up.
\[ \begin{array}{ll} y \le -\left|x - 3\right| + 2 & \text{solid, opens downward from } (3,2) \\ y > \left|x - 3\right| - 1 & \text{dashed, opens upward from } (3,-1) \end{array} \]
Read off the two vertices
Why: in both, the inside of the bars is zero at x equal to 3; the constant outside then gives the height of each point.
\[ (3, 2) \quad \text{and} \quad (3, -1) \]
Find where the two boundaries cross
Why: set the expressions equal and gather the absolute values on one side.
\[ -\left|x-3\right| + 2 = \left|x-3\right| - 1 \;\Longrightarrow\; 3 = 2\left|x-3\right| \]
Solve the absolute value equation
Why: divide by 2, then split into the two cases and add 3 to each.
\[ \left|x-3\right| = \tfrac32 \;\Longrightarrow\; x = 4.5 \;\text{ or }\; x = 1.5 \]
Find the matching y values
Why: substitute either x into either boundary equation; both give the same height, which is a free check.
\[ y = \left|1.5-3\right| - 1 = 1.5 - 1 = 0.5 \]
Figure (svg): Two V shapes with the same point of the horizontal axis below them, one opening upward and dashed and one opening downward and solid, enclosing a four-cornered region with one filled corner at the top and three open corners.
Both boundaries have the same two slopes, one and negative one, so the four sides of the diamond come in parallel pairs. That is what makes the shape a rhombus rather than an irregular quadrilateral, and it is a consequence of both absolute values being plain, with no number multiplying them.
Four corners, arranged as a diamond. The top one is filled in because it lies only on the solid boundary; the other three are open, because each of them lies on the dashed one.
Check one point inside and one outside
Why: the pair three, one half sits between the two vertices, while three, three is above the downward V's own peak.
\[ \begin{array}{l} (3,0.5): \; 0.5 \le 2 \;\checkmark, \; 0.5 > -1 \;\checkmark \\ (3,3): \; 3 \le 2 \;\times \end{array} \]
Notice that both inequalities were easy to check at x equal to 3, because the absolute value is zero there. Choosing a test point at a V's own vertex line is a small trick worth using.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 30
Prediction
Two V's again, but this time they share a vertex. Think about what the system is asking before you predict, and in particular about what it says at the origin.
\[ \begin{array}{l} y < \left|x\right| \\ y > -\left|x\right| \end{array} \]
Predict first
What does the graph of this system look like?
Correct: Two wedges, one to the left and one to the right, touching only at the origin
Why: The system asks y to be smaller than the absolute value of x and bigger than its negative, which together say that the size of y is smaller than the size of x. To the right of the vertical axis that is the wedge between the two arms opening rightward, and to the left it is the matching wedge opening leftward. The origin fails both strict inequalities, so the two wedges touch there without joining.
The answer is worth restating in words, because it is the neatest description: a pair is a solution exactly when the size of y is less than the size of x. Every point closer to the horizontal axis than to the vertical one, in that sense.
The origin is the point that does the interesting work. Both inequalities are strict, and at the origin both sides are zero, so both fail, and the single point that would have joined the two wedges is missing.
This is also a reminder that a solution set need not be one connected piece. Nothing in the definition of a system says the overlap has to be a single blob, and a bent boundary is the easiest way to see that.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 28
Picture it
Here is the picture. Two dashed V's sharing a vertex, and the shading is the pair of wedges that open left and right between them.
Figure (svg): Two dashed V shapes, one opening upward and one downward, shading the left and right wedges between them, with the origin left open where the two wedges touch.
Test the labelled points against the description. At three, one the size of y is one and the size of x is three, so the pair qualifies; the same is true at negative three, one, which is why the left wedge is shaded too.
At one, three the sizes are the other way round, so the pair fails. That point sits in the upper region, between the arms of the upward V, and the same happens by symmetry in the lower region.
The open circle at the origin is the whole reason to call this two pieces rather than one. Walking from the right wedge to the left one, you would have to pass through the origin, and the origin is not in the set.
Worth noticing for later: if both inequalities had allowed equality, the origin would have been included, the two wedges would have joined there, and the answer would be one connected region with a pinch point. A single symbol changes the topology.
Comparison
Fill in the missing entries. Every row is something you have used in the last few slides, and the point of the table is that only two rows really change.
Comparison matrix
| what you are comparing | a straight boundary | a V-shaped boundary |
|---|---|---|
| the boundary equation | y = mx + b | y = |x − h| + k |
| straight pieces in the boundary | 1 | 2 |
| corners the boundary brings with it | 0 | 1 |
| what the inequality shades | one side of the line | the inside or the outside of the V |
| dashed or solid is decided by | the symbol | the symbol, for the whole V at once |
| can two inequalities bound a region? | no | yes |
The last row is the one with real consequences. It is why Example 3 could produce a triangle from two conditions, and why you should not assume that a bounded region needs three inequalities.
The row about corners matters for your list of vertices. A V arrives with a corner already built in, so a system with one V can have a vertex that no pair of boundaries produced; it is the V's own point.
Everything else on the table is unchanged, and that is the real message. You did not learn a new procedure in this part; you applied the old one to a boundary with a bend in it.
Section
Part 5
Concept
Real conditions arrive as sentences, and almost every sentence that mentions a range is two inequalities wearing one coat. Splitting them apart is the first move in every modelling problem in this section.
| what the sentence says | the inequalities | why |
|---|---|---|
| the price is between 20 and 80 dollars | x ≥ 20 and x ≤ 80 | a range gives a floor and a ceiling |
| you may work at most 20 hours | x + y ≤ 20 | at most is a ceiling on the total |
| you must earn at least 92 dollars | 7.5x + 6y ≥ 92 | at least is a floor |
| the discount is 10 to 60 percent | y ≥ 0.4x and y ≤ 0.9x | paying 40 to 90 percent of the price |
| the number of hours cannot be negative | x ≥ 0 and y ≥ 0 | a constraint nobody bothers to state |
Read the last row twice. Nothing in the wording of a real problem tells you that hours, prices and quantities are not negative, and yet the region would be nonsense without those inequalities. You are expected to add them yourself.
The fourth row is the one that trips people. A discount of sixty percent means you pay forty percent, so the smallest the sale price can be is four tenths of the regular price. Percent off and percent paid add to one hundred, and reading one as the other is the single commonest modelling error in this section.
One more habit before the examples. Say out loud what x and y stand for, including units, and write it down. A region is meaningless until you know which axis is which, and half the difficulty of these problems disappears once the two letters are pinned to real quantities.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 170 — Example 4, step 1
Worked example
A discount shoe shop advertises ten to sixty percent off all footwear, on items regularly priced from twenty to eighty dollars. Let x be the regular price in dollars and y the sale price in dollars, and turn each phrase of the advertisement into an inequality.
Write the two inequalities about the regular price
Why: the advertisement restricts which items are on sale at all, and that is a range on x alone.
\[ x \ge 20 \quad \text{and} \quad x \le 80 \]
Turn the largest discount into a floor on the sale price
Why: sixty percent off means you pay one hundred take away sixty, which is forty percent, so the sale price is at least four tenths of the regular price.
\[ y \ge 0.4x \]
Turn the smallest discount into a ceiling on the sale price
Why: ten percent off means you pay ninety percent, so the sale price is at most nine tenths of the regular price.
\[ y \le 0.9x \]
Notice the direction of each translation. The biggest discount gives the lowest price and therefore the floor, and the smallest discount gives the highest price and therefore the ceiling. Getting that pairing the wrong way round flips the region.
Find the four corners
Why: each corner is one of the vertical boundaries meeting one of the two rays, so substitute x equal to 20 and x equal to 80 into both ray equations.
\[ \begin{array}{ll} (20,\,8) & (20,\,18) \\ (80,\,32) & (80,\,72) \end{array} \]
Figure (svg): A four-cornered region between two rays through the origin and two vertical lines at twenty and eighty, with a dashed vertical slice at seventy meeting the region between twenty-eight and sixty-three.
The region is a four-cornered strip fanning out to the right, and every point in it is a legitimate pair of prices: a regular price the shop stocks, together with a sale price the advertisement permits for it.
Read off the answer for a seventy dollar item
Why: the question asks for the possible sale prices when x is 70, which is the vertical slice through the region at that value.
\[ 0.4(70) = 28 \quad \text{and} \quad 0.9(70) = 63 \;\Longrightarrow\; 28 \le y \le 63 \]
Check the answer against the advertisement
Why: the low end should be sixty percent off seventy dollars and the high end ten percent off, and both come out right.
\[ 70 - 0.6(70) = 70 - 42 = 28, \qquad 70 - 0.1(70) = 70 - 7 = 63 \]
So footwear regularly priced at seventy dollars sells for between twenty-eight and sixty-three dollars inclusive. The check matters because it computes the answer a completely different way, by subtracting the discount rather than multiplying by what is left.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 170 — Example 4
Estimation
A graph you have built is a machine for answering the same question about any price. Take a fifty dollar pair of shoes and read the slice, or work it out arithmetically, whichever you trust more.
\[ x = 50 \]
Predict first
Under the same advertisement, what sale prices are possible for footwear regularly priced at 50 dollars?
Correct: Between 20 and 45 dollars
Why: The floor is four tenths of fifty, which is twenty dollars, and the ceiling is nine tenths of fifty, which is forty-five. So the slice through the region at fifty runs from twenty up to forty-five, and both ends are included because both of those boundaries are solid.
The last of those options is the answer to the previous slide's question rather than this one, which is a useful reminder to check what x you are standing at before reading a slice.
The option running from five dollars comes from taking ninety percent off instead of ten, and the one stopping at thirty comes from using sixty percent of the price rather than forty. Both are the percent-off confusion, and both are caught instantly by asking whether the answer looks like a sensible discount.
Notice how the width of the slice grows with the price. At twenty dollars the possible sale prices span ten dollars, at fifty they span twenty-five, and at eighty they span forty. That widening is what makes the region a fan rather than a strip of constant width.
Worked example
Now suppose the advertisement offers twenty to fifty percent off, on regular prices from forty to one hundred dollars. Rebuild the system from scratch, and notice how little of the work is actually new.
Write the range on the regular price
Why: the two ends of the stated price range give a floor and a ceiling on x.
\[ x \ge 40 \quad \text{and} \quad x \le 100 \]
Turn the largest discount into a floor on the sale price
Why: fifty percent off leaves you paying fifty percent, so the sale price is at least half of the regular price.
\[ y \ge 0.5x \]
Turn the smallest discount into a ceiling
Why: twenty percent off leaves eighty percent, so the sale price is at most eight tenths of the regular price.
\[ y \le 0.8x \]
Find the four corners
Why: substitute each end of the price range into each ray equation.
\[ \begin{array}{ll} (40,\,20) & (40,\,32) \\ (100,\,50) & (100,\,80) \end{array} \]
Figure (svg): A four-cornered region between the rays y equals half x and y equals four fifths x, cut off at forty and one hundred on the horizontal axis, with a dashed slice at sixty meeting it between thirty and forty-eight.
Compare this picture with the previous one. The shape is identical in kind: two rays through the origin, cut off by two vertical lines. Only the steepness of the rays and the positions of the cuts have changed.
That is what a model buys you. Having built the region once, a changed advertisement is not a new problem but a new set of four numbers, and the reasoning is already done.
Read off the answer for a sixty dollar item
Why: take the slice at x equal to 60, whose two ends are half of sixty and eight tenths of sixty.
\[ 0.5(60) = 30 \quad \text{and} \quad 0.8(60) = 48 \]
Check both ends against the discounts
Why: thirty dollars should be fifty percent off sixty and forty-eight should be twenty percent off, and both do.
\[ 60 - 0.5(60) = 30, \qquad 60 - 0.2(60) = 48 \]
So a sixty dollar item sells for between thirty and forty-eight dollars. Sixty does lie inside the stated price range of forty to one hundred, which is worth confirming before answering, since a question about a thirty dollar item would have no answer at all under this advertisement.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 170 — Guided Practice 7
Error analysis
Here is a model of the original shoe sale that a careful-looking student produced. The four inequalities have the right shape and two of them are wrong, in the way that is easiest to miss.
Annotate
On: \( \begin{array}{l} x \ge 20, \quad x \le 80 \\ y \ge 0.1x, \quad y \le 0.6x \end{array} \)
Notice that the error also reverses which percentage is the floor and which is the ceiling. The biggest discount gives the smallest price, so the sixty must end up in the lower boundary, and in the faulty version it is in the upper one.
The cheapest guard is a single sanity check with round numbers. On a hundred dollar item, sixty percent off is forty dollars and ten percent off is ninety, so your boundaries should give forty and ninety and not sixty and ten.
More generally, test every model you write on one easy number before you graph it. A region built from a wrong inequality is drawn just as convincingly as a region built from a right one, and the picture will not warn you.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 170 — the percentages in Example 4, step 1
Concept
Word problems leave out the conditions that are obvious to a human reader and invisible to a graph. A region drawn without them will contain answers that make no sense, and your job is to put them back.
The two non-negativity inequalities do real geometric work: they put the two axes into the system as boundaries and confine everything to the first quadrant. Without them a region often runs off into a corner of the plane that means nothing.
The whole-number condition is different in kind, because no system of linear inequalities can say it. The honest procedure is to graph the region as it stands and then read off the lattice points inside it, which is what a problem means when it asks how many plans are possible.
So the habit is: write down every inequality the sentences give you, then read the list back and ask what a nonsensical answer would look like. Every nonsensical answer you can imagine is a missing inequality.
Matching
Four phrases from real problems, four inequalities. Decide each one by asking whether the phrase sets a floor or a ceiling, and on what combination of the two quantities.
Match the pairs
Why: At most is a ceiling and at least is a floor, so the hours phrase gives a less-than-or-equal inequality on the sum and the earnings phrase a greater-than-or-equal inequality on the money. Taking 40 to 60 percent off means paying 40 to 60 percent, so the sale price is squeezed between four tenths and six tenths of the regular price. The pH phrase restricts one variable only, which is a pair of vertical boundaries.
The phrase about the discount is the one to dwell on. Here forty to sixty percent off happens to leave forty to sixty percent paid, which is a coincidence of these particular numbers and not a rule. Always subtract from one hundred and see what you get.
Notice also that two of the four inequalities involve both variables and two involve only one. Both kinds are ordinary members of a system, and an inequality in one variable still draws a full line and shades a full half-plane, exactly as in Lesson 2.8.
A phrase using between is really two inequalities, and writing it as a double inequality is a compact way to record both. When you graph it, split it back into two conditions and draw two boundaries.
Real world
This is the shape of problem that made linear programming worth inventing. Read the constraints, write every one of them including the ones the question does not mention, and then look at what the region tells you.
You can work at most twenty hours next week and need to earn at least ninety-two dollars. Walking dogs pays seven dollars fifty an hour and the car wash pays six dollars an hour.
Discussion prompt
Let x be the hours spent walking dogs and y the hours at the car wash. Write the full system of inequalities, including the two conditions the problem does not state, and then say whether twelve hours of dog walking and nothing else would do.
Figure (svg): A four-cornered region in the first quadrant, bounded above by the line x plus y equals twenty and below by the earnings line, with its four corners marked and one rejected plan shown in red.
The region has four corners and they come in two kinds. Two of them sit where the earnings line crosses the axes, at about twelve point two seven hours of dog walking or about fifteen point three three hours at the car wash; the other two are the ends of the twenty-hour line.
The red point at four hours of each is the kind of plan the region rules out: well inside the twenty hour limit and earning only fifty-four dollars. The green point at eight hours of each earns one hundred and thirty-two dollars in sixteen hours, comfortably inside.
Notice that the earnings boundary is the only one that is really binding for a short week, and the hours boundary the only one that binds for a long one. Between them they describe every workable plan, and the corners are where you would look if you also wanted to minimise the hours.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 34
Worked example
A book on tropical fish says the water should be between eight point zero and eight point three pH units, and between seventy-six and eighty degrees Fahrenheit. Let x be the pH and y the temperature.
Write the pH range as two inequalities
Why: a range on one variable gives a floor and a ceiling, drawn as two vertical boundaries.
\[ x \ge 8.0 \quad \text{and} \quad x \le 8.3 \]
Write the temperature range as two more
Why: another range on one variable, this time giving two horizontal boundaries.
\[ y \ge 76 \quad \text{and} \quad y \le 80 \]
Four one-variable inequalities, two vertical boundaries and two horizontal ones, so the region is a rectangle. That is the shape you should expect whenever every inequality restricts a single variable, and it is the easiest region in the whole lesson to describe.
Convert the two temperatures to Celsius
Why: the conversion formula from Lesson 1.4 subtracts thirty-two and then multiplies by five ninths.
\[ C = \tfrac59 (F - 32): \quad \tfrac59(76-32) = \tfrac{220}{9} \approx 24.44 \]
Convert the other end the same way
Why: eighty degrees Fahrenheit is forty-eight above freezing, and five ninths of forty-eight is a little under twenty-seven.
\[ \tfrac59(80-32) = \tfrac{240}{9} \approx 26.67 \]
Figure (svg): Two panels of the same pH range. The left shades a rectangle from seventy-six to eighty degrees Fahrenheit; the right shades the same conditions from twenty-four point four to twenty-six point seven degrees Celsius, a visibly shorter rectangle.
The two pictures describe exactly the same set of tanks. What changed is the scale on the vertical axis, and because a Celsius degree is larger than a Fahrenheit degree, the same band of temperatures occupies a shorter interval, about two point two two units instead of four.
Check the conversion by going forward again
Why: turn twenty-four point four four degrees Celsius back into Fahrenheit with the reverse formula and confirm it gives seventy-six.
\[ F = \tfrac95 C + 32: \quad \tfrac95\left(\tfrac{220}{9}\right) + 32 = 44 + 32 = 76 \]
The rectangle also gets slightly narrower in appearance only, because the pH axis did not change at all. Nothing about which tanks are acceptable depends on the units you happen to measure them in, which is a good thing to be able to see rather than to be told.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 36
Socratic
A shaded region answers one question completely and refuses to answer several others. Being clear about which is which is what stops a model being oversold.
Discussion prompt
Look back at the summer jobs region. Name two questions the region answers exactly, and two questions it cannot answer at all. Then say what extra information would be needed to answer one of the ones it cannot.
The distinction is between describing a set of possibilities and choosing among them. Everything in this lesson does the first, and the second needs one more ingredient that nobody has given you yet.
That is also why the corners keep getting singled out. They are the points where the region is most extreme in some direction, so as soon as somebody names a quantity to push to its limit, the corners are where the answer will be.
Worth carrying into any modelling you do: a region is a claim about what is allowed, not a recommendation. Confusing the two is how a graph ends up being blamed for a decision it never made.
Section
Part 6
Pattern
Everything in this lesson is one procedure with four steps, and each step contains a decision that can be made carelessly. The steps are easy; the decisions are where marks are lost.
Figure (svg): A four-row flow diagram. Each row names a step of the procedure on the left and the decision hidden inside it on the right.
The last line is the only one that can catch a mistake, and it is the one people skip. Two extra substitutions turn a drawing you hope is right into a drawing you know is right.
Notice that nothing in the list depends on how many inequalities there are, or on whether their boundaries are straight. That is why the same six lines handled a wedge, an empty system, an open triangle, a diamond between two V's and a fan of shoe prices.
Worked example
Everything so far has gone from inequalities to a picture. This exercise runs the other way: you are given a shaded triangle with vertices at one comma three, negative two comma negative three and two comma negative two, and asked for a system that produces it.
Figure (svg): A shaded triangle with vertices at one comma three, minus two comma minus three and two comma minus two, its three sides drawn solid and extended slightly past the corners.
Find the equation of the side through (1, 3) and (−2, −3)
Why: the slope is the rise over the run, and then use point-slope form from Lesson 2.4.
\[ m = \dfrac{3 - (-3)}{1 - (-2)} = \dfrac63 = 2 \;\Longrightarrow\; y = 2x + 1 \]
Find the equation of the side through (−2, −3) and (2, −2)
Why: the same two moves, with a much gentler slope.
\[ m = \dfrac{-2 - (-3)}{2 - (-2)} = \dfrac14 \;\Longrightarrow\; y = \tfrac14 x - \tfrac52 \]
Find the equation of the side through (1, 3) and (2, −2)
Why: a steep negative slope this time.
\[ m = \dfrac{-2 - 3}{2 - 1} = -5 \;\Longrightarrow\; y = -5x + 8 \]
Three vertices, three side equations. Now each inequality has to face the right way, and one interior point decides all three at once, which is why the picture has a test point marked inside it.
Test an interior point in all three
Why: the average of the three vertices is one third, negative two thirds, and it lies inside the triangle; substitute it into each side expression and see which way the inequality must point.
\[ \begin{array}{l} -\tfrac23 \le 2\left(\tfrac13\right) + 1 = \tfrac53 \\ -\tfrac23 \ge \tfrac14\left(\tfrac13\right) - \tfrac52 \approx -2.42 \\ -\tfrac23 \le -5\left(\tfrac13\right) + 8 \approx 6.33 \end{array} \]
Write the system
Why: keep the direction each test gave, and draw every boundary solid since the sides of the shaded triangle are part of it.
\[ \begin{array}{l} y \le 2x + 1 \\ y \ge \tfrac14 x - \tfrac52 \\ y \le -5x + 8 \end{array} \]
Check every vertex in every inequality
Why: each vertex should satisfy all three with equality in exactly two of them, since a vertex lies on two sides and strictly inside the third.
\[ (2,-2): \; -2 \le 5 \;\checkmark, \; -2 = \tfrac14(2) - \tfrac52 \;\checkmark, \; -2 = -5(2) + 8 \;\checkmark \]
Two equalities and one strict inequality is exactly the signature of a vertex, so that single line confirms both the side equations and the directions. Run it on the other two vertices and the system is proved.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 33
Tweak it
Here is the Part 2 question made live. One boundary stays where it is and the other slides, and you can watch the strip between them narrow, close and vanish.
Parameter explorer
The system is y at least 2x minus 4, whose boundary stays put, together with y at most 2x plus b, whose boundary you can slide. Move b and say for which values of b the system still has solutions, and describe the solution set at the smallest of those values.
\[ y \le 2x + b, \quad b = {b} \]
Watch what happens as you bring the curve down. While the sliding boundary is above the fixed one, the solutions fill a strip whose vertical width is exactly b plus four, and the strip gets thinner as b falls.
At the value negative four the strip has zero width, and the two inequalities together say that y equals two x minus four. The solution set is the line itself, which is neither empty nor a region, and which both inequalities accept because both allow equality.
Below that value the shadings face apart and there is nothing. That is Example 2, with the gap width under your control instead of fixed, and it shows that no solution is not a strange special case but simply what lies past the moment the strip closes.
Try the same experiment in your head with a strict symbol on the sliding inequality. The strip closes at the same value of b, but at that value the solution set is empty rather than a line, because the boundary is then excluded.
Check
Three inequalities, so three substitutions per candidate. Work through the options rather than guessing, and notice which single inequality each wrong answer breaks.
Check your understanding
Which ordered pair is a solution of the system x < 6, y > −1, y < x?
Answer: A
Why: Substituting the pair 2 and 0 gives 2 less than 6, then 0 greater than negative 1, then 0 less than 2, and all three are true, so this pair lies inside the open triangle. Every other option fails exactly one of the three inequalities, which is enough to disqualify it.
Notice that each wrong option satisfies two of the three inequalities. That is deliberate, and it is the whole lesson in miniature: doing well on most of the conditions earns nothing when the word joining them is and.
It is also a diagnosis you can use on your own drawing. If a point you expected to be inside turns out to fail exactly one inequality, then exactly one of your shadings is on the wrong side, and you know which one.
Worth doing once on paper: place all four of these points on the triangle from Exercise 17 and see that each wrong one sits just across a different edge. The failures are not random, they are geographical.
Check
This is Exercise 26, and it rewards thinking before substituting. Rewrite the second inequality for y, then ask what each condition allows in each quadrant.
Check your understanding
Which quadrant of the coordinate plane contains no solutions of the system y ≤ −|x − 3| + 2 and 4x − 5y ≤ 20?
Answer: B
Why: Rewriting the second inequality gives y at least four fifths of x take away four. In the quadrant where x is negative and y is positive, the first inequality is the binding one: with x negative the absolute value of x take away 3 is more than 3, so the first inequality forces y below negative 1, which cannot also be positive. Every other quadrant does contain solutions, for instance the pair 3 and 1, the pair negative 1 and negative 3, and the pair 3 and negative 1.
The key move is realising that the upside-down V drops fast once you leave its vertex at three, two. Four units to the left of the vertex the boundary is already at negative two, and it keeps falling, so the whole left of the plane is pushed below the horizontal axis.
The other inequality is a half-plane above a gently rising line, and it is generous: it accepts everything above a line that passes through five, zero and zero, negative four. It is not the one that rules the second quadrant out.
A quicker route, if you spot it: the first inequality alone allows a positive y only when the absolute value of x take away three is less than two, which means x lies between one and five. Every such x is positive, so no solution can have a negative x and a positive y.
McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.3 Graph Systems of Linear Inequalities §3.3, p. 172 — Exercise 26
Check
Four systems, one of them empty. Rewrite anything that is not already solved for y, then compare the slopes and the directions of the two shadings.
Check your understanding
Which of these systems has no solution?
Answer: A
Why: In the empty one, both boundaries have slope 1, the first shading goes upward from the higher line and the second goes downward from the lower line, so they face apart across a gap of vertical width 2. Chaining the two inequalities gives x plus 1 at most x take away 1, which reduces to 1 at most negative 1, and that is false.
Three of these four systems have parallel or axis-parallel boundaries, and only one of them is empty. That is the point of Part 2: parallel boundaries are a reason to look at the symbols, never a verdict on their own.
The test that settles all four quickly is the origin. It satisfies both inequalities of the strip system and both of the quarter-plane system, fails one inequality of the system that reduces to y greater than five, and fails one of the empty system too, so where the origin fails you have to look a little harder.
When you do have to look harder, the chaining argument is the reliable tool. Write the two bounds on y in one line and cancel the matching terms, and you are left with a statement about numbers that is either true always or false always.
Explain it to yourself
One idea has carried the whole lesson, and putting it into your own sentences is what makes it stick. Write your answer out before reading the one provided.
Discussion prompt
In three or four sentences, explain to someone who can already graph a single linear inequality what changes when they are given a system, and what stays exactly the same. Include why the region can never grow when an inequality is added, and why a corner might not belong to the region it corners.
If your explanation talked mostly about drawing, try it again talking about testing. The picture is a way of recording the answer to infinitely many tests at once, and understanding the tests is what makes the picture trustworthy.
A good measure of whether you have it: could you explain why the answer to a system is usually a wedge, a strip, a triangle or nothing, without drawing anything? If yes, you understand the overlap rather than just the procedure.
Try the explanation once out loud on the shoe sale. Four sentences, each condition named, and then the region described as the set of price pairs the advertisement permits. Modelling problems get much easier once you can narrate them.
Exit ticket
One system that uses everything: a V, a line, corners found by algebra, and a decision about whether each corner belongs.
\[ \begin{array}{l} y \le 4 \\ y > \left|x + 1\right| \end{array} \]
Discussion prompt
Graph this system. Give the three corners of the region, say for each one whether it belongs to the region and why, then test one point inside and one point outside in both inequalities.
Work through it in the order the lesson taught. Find the vertex of the V first, then the two places where the V meets the ceiling, and only then draw, because those three points fix the whole triangle.
The top edge here is solid and the two sides are dashed, so the region owns its lid and not its sides. Saying which corners are in and which are out is part of a complete answer, not a decoration.
If you finish early, change the first inequality to allow equality only on the V instead of the line and ask what happens to the three corners. Two of them change status and one does not, and working out which is a good test of whether the rule has landed.
Recap
One procedure, run once per inequality, and then one final question. Graph each inequality on the same axes, keep only what every shading covers, and check the region you are left with.
| what you see in the system | what the region turns out to be | example from this lesson |
|---|---|---|
| two boundaries that cross | a wedge, unbounded | Example 1 |
| parallel boundaries, shadings apart | no solution at all | Example 2 |
| parallel boundaries, shadings inward | a strip between the lines | the middle panel of the three fates |
| parallel boundaries, same direction | the stricter half-plane; one inequality is idle | Exercise 21 |
| three or four boundaries, all directions blocked | a bounded region with corners | Exercise 17, Exercise 24 |
| a V and a line | a triangle, from only two inequalities | Example 3, Guided Practice 4 |
| two ranges in words | a rectangle or a fan of allowed pairs | Example 4, Exercise 36 |
\[ \begin{array}{l} \text{corner} = \text{solution of the two boundary equations} \\ \text{open if either boundary through it is dashed} \end{array} \]
That line is the piece to carry forward. Corners are found by the algebra of Lesson 3.2, and whether a corner belongs to the region is decided by the symbols, not by the algebra.
Look back at the table and notice that only the last three rows are genuinely new. The first four are all the same reasoning about two half-planes, applied to boundaries that cross or do not, and the rest is that reasoning repeated.
Lesson 3.4 goes on to systems of three equations in three unknowns, and the chapter after it turns these regions into a method. Once you attach a quantity to maximise, the corners you have been finding all lesson become the only candidates worth testing, and that is linear programming.
OpenStax Algebra and Trigonometry 2e, §11.3 Systems of Nonlinear Equations and Inequalities: Two Variables §11.3 — the same overlap idea when the boundaries are curves
Want this taught 1-on-1? Alexander tutors Algebra 2 — $55/session, free consultation.