Linear Programming

A furniture dealer has Rs 50,000 to spend and storage for at most 60 pieces. Tables cost Rs 2500 and earn a profit of Rs 250 each; chairs cost Rs 500 and earn Rs 75 each. Buy only tables and the profit is Rs 5000. Buy only chairs and it is Rs 4500. Buy 10 tables and 50 chairs and it jumps to Rs 6250. With infinitely many combinations available, how does the dealer find the best one without checking every possibility by hand? This is the question that opens Chapter 12, and its answer — Linear Programming — turns a vague "maximise my profit" wish into a precise, solvable piece of mathematics. The chapter's big insight is almost shocking in its economy: among the infinitely many points satisfying a set of linear constraints, the best one is never buried in the interior — it always sits at a corner point of the feasible region. That single fact, proved as a theorem here (without proof, as NCERT states, since the proof needs convexity theory beyond this book), converts an infinite search into a finite checklist: list the corners, plug each into the objective function, and read off the winner. This is the Corner Point Method, the working engine of the whole chapter. You will learn to translate a word problem into decision variables, an objective function, and a system of linear constraints; to graph that system and shade the feasible region using the same techniques from Class XI's linear inequalities; to identify corner points either by inspection or by solving pairs of boundary equations; and to handle the subtlety of unbounded regions, where a naive "largest value in my table" can be dangerously wrong. Along the way you meet infeasible problems (no feasible region at all) and problems with infinitely many optimal solutions (an entire edge of the region ties for best). By the end, the furniture dealer's problem — and dozens like it in diet planning, manufacturing, transportation, and allocation of resources — become routine five-step exercises. This is applied mathematics at its cleanest: a real decision, made rigorous by algebra and geometry working together.

Key takeaways

A linear programming problem (LPP) seeks the maximum or minimum value of a linear objective function $Z = ax + by$ subject to linear constraints (inequalities or equations) and non-negativity restrictions $x \geq 0,\ y \geq 0$ .
The feasible region is the set of all points satisfying every constraint simultaneously; every point in it is a feasible solution, and points outside are infeasible.
The feasible region of a linear programming problem is always a convex region — a bounded or unbounded polygon with no dents.
Theorem 1 (Fundamental theorem of LPP): if $Z$ has an optimal value on the feasible region, that value must occur at a corner point (vertex) of the region.
Theorem 2: if the feasible region is bounded, $Z$ has both a maximum and a minimum, and both occur at corner points — this guarantees the Corner Point Method always succeeds on bounded regions.
The Corner Point Method: find every corner point, evaluate $Z$ at each, and the largest/smallest value found is the maximum/minimum (when the region is bounded).
When the feasible region is unbounded, a value that looks maximum (or minimum) in the corner-point table must be checked by testing whether the open half-plane $ax + by > M$ (or $< m$ ) meets the feasible region; if it does, no such optimal value exists.
If two corner points give the same optimal value of $Z$ , then every point on the segment joining them is also optimal — the problem has infinitely many optimal solutions.
A feasible region can be empty (no point satisfies every constraint at once); such an LPP has no feasible solution and hence no optimal solution.
Key idea: the optimal value of a linear objective function over a convex feasible region always occurs at a corner point, so solving any LPP graphically reduces to listing corner points, evaluating Z at each, and picking the best — with an extra unbounded-region check when needed.

The core insight

Because a linear objective function's extreme values over a convex feasible region always occur at a vertex, every linear programming problem graphically reduces to finding the corner points and comparing $Z$ at each of them.

What this chapter covers

The chapter builds in five stages, each depending on the one before it. It opens with the furniture dealer problem, used throughout as the running example, to motivate why real allocation decisions are optimisation problems. From this example the chapter extracts the vocabulary of linear programming: decision variables, objective function, constraints, non-negative restrictions, and the general mathematical formulation of an LPP. Next comes the graphical method: plotting the constraint inequalities exactly as in Class XI, shading the common region, and naming it precisely — feasible region, feasible solution, infeasible solution, optimal solution, bounded and unbounded regions. The heart of the chapter is the pair of theorems that justify the Corner Point Method — that an optimum (if it exists) sits at a vertex, and that a bounded feasible region guarantees both a maximum and a minimum. These are stated without proof in NCERT (their rigorous proof needs convexity results beyond Class XII), but this article gives you the reasoning behind why they must be true, and a worked derivation of the method's mechanics. The chapter then works through NCERT's five core solved examples, escalating in difficulty: a straightforward bounded maximisation (Example 1), a bounded minimisation (Example 2), a problem with multiple optimal solutions (Example 3), an unbounded region requiring the half-plane test (Example 4), and a case with no feasible region at all (Example 5). It closes with Exercise 12.1, ten LPPs to solve graphically, and a Miscellaneous Exercise extending to word problems in diet planning, manufacturing, and transportation — all solved by the identical five-step recipe you master here.

The core ideas

Formulating a linear programming problem

Every LPP begins the same way: identify what you control (the decision variables), what you want to optimise (the objective function), and what limits you (the constraints). For the furniture dealer, let $x$ be the number of tables and $y$ the number of chairs bought. Since you cannot buy a negative number of tables or chairs,

x \geq 0, \qquad y \geq 0. \qquad \text{(non-negative restrictions)}

where: these two conditions are compulsory in every LPP and are stated separately from the other constraints because they arise from the physical meaning of the variables, not from the specific problem's resources. The dealer's money and storage space limit his choices. Investment cannot exceed Rs 50,000, and total pieces cannot exceed 60:

2500x + 500y \leq 50000 \quad \Longleftrightarrow \quad 5x + y \leq 100, \qquad x + y \leq 60.

where: dividing through by the common factor 500 turns the first inequality into smaller, easier-to-graph numbers — a standard simplification that preserves the inequality direction because 500 is positive. Finally the goal itself, the profit Z, stated as a linear function of the decision variables — this is the objective function:

Z = 250x + 75y.

where: the coefficients $250$ and $75$ are the profit per table and per chair; $Z$ is to be maximised. Collecting all of this gives the complete mathematical formulation:

\text{Maximise } Z = 250x + 75y \ \text{ subject to } \ 5x + y \leq 100,\ \ x + y \leq 60,\ \ x \geq 0,\ y \geq 0.

where: the phrase "subject to" introduces the constraint set; a solution to the LPP is any pair $(x, y)$ satisfying every one of these conditions, and the optimal solution is the one among them giving the best $Z$ . In general, a linear programming problem is one that finds the optimal (maximum or minimum) value of a linear objective function of several variables, subject to constraints that are linear inequalities or equations together with non-negativity restrictions. The term linear signals that no variable is squared, multiplied by another variable, or appears under a root; programming refers to planning a course of action, not to computer code.

The feasible region and its geometry

Once the constraints are graphed as half-planes (using the Class XI technique — plot the boundary line, then test a convenient point such as the origin to decide which side satisfies the inequality), their common intersection is the feasible region, also called the solution region. Every point in it — interior or on the boundary — is a feasible solution; every point outside it is an infeasible solution. A feasible region may be bounded (it fits entirely inside some sufficiently large circle) or unbounded (it stretches out indefinitely in some direction). This distinction matters enormously, because Theorem 2 below only guarantees an optimal value when the region is bounded. A crucial structural fact, visible in every example NCERT works, is that the feasible region of a linear programming problem is always convex: the line segment joining any two points of the region lies entirely within the region. Convexity is what makes the corner-point idea work at all — a non-convex region could hide an interior local optimum that no vertex captures.

Figure

The furniture dealer feasible region OABC bounded by the lines 5x plus y equals 100 and x plus y equals 60 with corner points O A B and C marked

Corner points and the two fundamental theorems

A corner point (vertex) of a feasible region is a point of the region lying at the intersection of two of its boundary lines. Once the region is drawn, the two theorems that anchor the whole chapter tell us exactly where to look for an optimal value of $Z$ .

\textbf{Theorem 1: } \text{if } Z = ax + by \text{ has an optimal value on the feasible region } R, \text{ it occurs at a corner point of } R.

where: $R$ must be the region determined by the constraints of an LPP (in particular, a convex region); this theorem does not by itself guarantee that an optimal value exists — it only says where to look if one does.

\textbf{Theorem 2: } \text{if } R \text{ is bounded, then } Z \text{ attains both a maximum and a minimum on } R, \text{ each at a corner point.}

where: boundedness is the extra condition that upgrades Theorem 1's conditional "if an optimum exists" into an unconditional guarantee — a bounded convex region always has finitely many corner points, and comparing $Z$ at each of them is guaranteed to find both extremes. NCERT states both without proof, since a full proof requires convexity theory beyond Class XII; the derivation below instead builds the intuition for why they must be true and turns that intuition into the working algorithm — the Corner Point Method.

Figure

Parallel level lines of Z equals ax plus by sliding across a convex feasible region and last touching it at a single vertex, showing the maximum occurs at a corner point

Worked derivations

Derivation AWhy the optimum of a linear function over a convex region must sit at a corner

The goal here is to see, using only elementary reasoning about straight lines, why a linear objective function $Z = ax + by$ can never attain its maximum or minimum at a genuinely interior point of a convex polygon, and why any point on the boundary that is not a vertex is likewise ruled out — forcing every optimum onto one of the finitely many corner points. This is the geometric content behind Theorem 1, made concrete rather than left as an assertion, and it explains why the Corner Point Method is not a lucky shortcut but a logical necessity.

Step 1. Look at what the level sets of $Z$ look like. For any fixed value $k$ , the equation $ax + by = k$ is a straight line, and as $k$ varies these lines form a family of parallel lines (all with the same slope $-a/b$ ), since only the constant term changes. Increasing $Z$ corresponds to sliding this line in one fixed direction — the direction of steepest increase, perpendicular to the lines — and decreasing $Z$ corresponds to sliding it the opposite way.

Step 2. Suppose, for contradiction, that $Z$ attains its maximum at a point $P$ strictly inside the feasible region (not on the boundary at all). Because $P$ is interior, a small disc around $P$ lies entirely within the region. But the line $ax + by = Z(P)$ through $P$ has points on both sides of it within that disc, and on one of those two sides $Z$ is strictly larger than $Z(P)$ (since $Z$ is not constant unless $a = b = 0$ , a trivial case excluded from real problems). That nearby point is still feasible, so $Z(P)$ was not the maximum after all — a contradiction. Hence no genuinely interior point can be optimal.

Step 3. So the optimum must lie on the boundary. Consider any boundary edge of the convex region that is not just a single point — a straight segment joining two corners, say $V_1$ and $V_2$ . Along this whole edge, $Z = ax + by$ varies linearly as you move from $V_1$ to $V_2$ (because the edge itself is a straight segment and $Z$ is a linear function of position along it). A linear function of one parameter attains its extreme values only at the endpoints of the interval it is evaluated on, never in the interior of the interval, unless it is constant throughout. So on this edge, the maximum (and minimum) of $Z$ occurs at $V_1$ or $V_2$ — the corner points bounding it — or, in the special tied case, equally at every point of the edge.

Step 4. Since the feasible region's entire boundary is made up of finitely many such edges, and each edge pushes its extreme values to its own two endpoints, the overall maximum (and minimum) of $Z$ over the whole boundary — and hence over the whole region, by Step 2 — must occur at one of the finitely many corner points where two edges meet. This is exactly Theorem 1. Theorem 2 follows because a bounded convex polygon has only finitely many edges and vertices, all of them actual points of the region, so comparing $Z$ at each vertex is guaranteed to catch both the true maximum and the true minimum — there is no way for the optimum to "escape" toward infinity as it might in an unbounded region. This reasoning is the geometric justification NCERT omits; the two theorems it produces are exactly what license the Corner Point Method used throughout the chapter.

Go deeper · Derivation B · 2 minThe Corner Point Method, step by step, and the unbounded-region check

Derivation BThe Corner Point Method, step by step, and the unbounded-region check

The goal of this derivation is to turn Theorems 1 and 2 into the precise, repeatable algorithm NCERT calls the Corner Point Method, and — crucially — to derive the extra test needed when the feasible region is unbounded, where the theorems alone are not enough to guarantee an optimal value exists at all. This matters because a student who only memorises "find corners, compare Z" will get Example 4 of the chapter wrong: the largest number in the table is not automatically the maximum when the region is unbounded.

Step 1. Graph every constraint as a half-plane exactly as in Class XI: draw the boundary line of each inequality (treating it as an equation), then test a convenient point not on the line — the origin $(0,0)$ when it is not itself on the boundary — to decide which side of the line satisfies the inequality, and shade that side. The intersection of all shaded half-planes, together with $x \geq 0, y \geq 0$ , is the feasible region $R$ .

Step 2. Locate every corner point of $R$ . A corner point is the intersection of two of the boundary lines forming $R$ 's edges. Some corners are read off directly from the axes intercepts of a single constraint; others require solving two boundary equations simultaneously (elimination or substitution), since a corner is where two lines meet.

Step 3. Evaluate $Z = ax + by$ at every corner point found in Step 2, and record the results in a table. Let $M$ denote the largest of these values and $m$ the smallest.

Step 4. If $R$ is bounded (Theorem 2 applies directly), $M$ is the true maximum of $Z$ and $m$ is the true minimum, full stop — no further check is needed, because a bounded convex region cannot let $Z$ escape past every corner value.

Step 5. If $R$ is unbounded, Theorem 2 no longer applies, and Theorem 1 only says an optimum, if it exists, is at a corner — it does not guarantee existence. So each candidate must be tested separately: $M$ is genuinely the maximum only if the open half-plane $ax + by > M$ shares no point with $R$ ; if it does share a point, $Z$ can be pushed past $M$ somewhere in $R$ and no maximum exists at all. Symmetrically, $m$ is the minimum only if the open half-plane $ax + by < m$ shares no point with $R$ ; otherwise $Z$ has no minimum. This half-plane test is simply asking, geometrically, whether the region keeps extending into territory where $Z$ is still improving.

Step 6. Two edge cases complete the method: if two corner points give the same optimal value of $Z$ , then by the "linear function on an edge is extreme only at endpoints, or constant throughout" fact from the previous derivation, every point of the segment joining them gives that same optimal value too — infinitely many optimal solutions. And if the constraints are mutually contradictory, the feasible region is empty, so there is no feasible solution and consequently no optimal solution to report. This six-step procedure — graph, find corners, evaluate, then apply Step 4 or Step 5 depending on boundedness — is executed verbatim in the three worked examples below.

Worked examples

Board level - Maximise Z = 4x + y (NCERT Example 1)

Step 1. The problem is Maximise $Z = 4x + y$ subject to $x + y \leq 50$ , $3x + y \leq 90$ , $x \geq 0, y \geq 0$ . Graph the two boundary lines $x + y = 50$ and $3x + y = 90$ along with the axes; testing the origin in each inequality confirms the region toward the origin's side is feasible, so the feasible region OABC is bounded.

Step 2. Find the corner points. $O = (0,0)$ is the origin. $A$ is where $3x + y = 90$ meets the x-axis ( $y = 0$ ): $3x = 90 \Rightarrow x = 30$ , so $A = (30, 0)$ . $C$ is where $x + y = 50$ meets the y-axis ( $x = 0$ ): $y = 50$ , so $C = (0, 50)$ . $B$ is the intersection of the two lines: solving $x + y = 50$ and $3x + y = 90$ by subtraction gives $2x = 40 \Rightarrow x = 20$ , then $y = 30$ , so $B = (20, 30)$ .

Step 3. Evaluate $Z = 4x + y$ at each corner: at $O(0,0)$ , $Z = 0$ ; at $A(30,0)$ , $Z = 120$ ; at $B(20,30)$ , $Z = 80 + 30 = 110$ ; at $C(0,50)$ , $Z = 50$ .

Step 4. Since the region is bounded, Theorem 2 applies directly — no half-plane check is needed. Comparing all four values, the largest is at $A$ .

The maximum value of Z is 120, attained at (30, 0).

Board level - Minimise Z = 200x + 500y (NCERT Example 2)

Step 1. The problem is Minimise $Z = 200x + 500y$ subject to $x + 2y \geq 10$ , $3x + 4y \leq 24$ , $x \geq 0, y \geq 0$ . Here the first constraint is a "greater than or equal to," so the feasible side of that line is away from the origin, while the second constraint's feasible side is toward the origin; graphing both gives a bounded triangular region ABC.

Step 2. Find the corner points. $A$ is where $x + 2y = 10$ meets the y-axis: $x = 0 \Rightarrow y = 5$ , so $A = (0, 5)$ . $C$ is where $3x + 4y = 24$ meets the y-axis: $x = 0 \Rightarrow y = 6$ , so $C = (0, 6)$ . $B$ is the intersection of the two lines: from $x + 2y = 10$ , $x = 10 - 2y$ ; substituting into $3(10-2y) + 4y = 24$ gives $30 - 6y + 4y = 24 \Rightarrow -2y = -6 \Rightarrow y = 3$ , then $x = 10 - 6 = 4$ , so $B = (4, 3)$ .

Step 3. Evaluate $Z = 200x + 500y$ : at $A(0,5)$ , $Z = 2500$ ; at $B(4,3)$ , $Z = 800 + 1500 = 2300$ ; at $C(0,6)$ , $Z = 3000$ .

Step 4. The region is bounded, so Theorem 2 guarantees these are the true extremes without further checking; the smallest value is at $B$ .

The minimum value of Z is 2300, attained at (4, 3).

Competitive level (JEE / NEET) - Unbounded region with a half-plane existence check (NCERT Example 4, extended)

Step 1. The problem is Determine the minimum value of $Z = -50x + 20y$ subject to $2x - y \geq -5$ , $3x + y \geq 3$ , $2x - 3y \leq 12$ , $x \geq 0, y \geq 0$ . Graphing all three lines together with the axes shows the feasible region is unbounded, extending indefinitely — this is the signal that Theorem 2 does not apply and the half-plane check will be required.

Step 2. Find the corner points by solving adjacent boundary pairs. Intersecting $2x - y = -5$ with the y-axis ( $x=0$ ) gives $y = 5$ , so one corner is $(0,5)$ . Intersecting $2x - y = -5$ with $3x + y = 3$ : adding the two equations gives $5x = -2$ ... instead intersecting $3x+y=3$ with the y-axis gives $(0,3)$ , and intersecting $3x + y = 3$ with the x-axis gives $(1, 0)$ ; intersecting $2x - 3y = 12$ with the x-axis gives $(6, 0)$ . The relevant corner points bounding the feasible region are $(0,5),\ (0,3),\ (1,0),\ (6,0)$ .

Step 3. Evaluate $Z = -50x + 20y$ at each: at $(0,5)$ , $Z = 100$ ; at $(0,3)$ , $Z = 60$ ; at $(1,0)$ , $Z = -50$ ; at $(6,0)$ , $Z = -300$ . The smallest table value is $-300$ at $(6,0)$ .

Step 4. Since the region is unbounded, $-300$ is only a candidate minimum, so test whether the open half-plane $-50x + 20y < -300$ (equivalently $-5x + 2y < -30$ ) has any point in common with the feasible region. Graphing this half-plane against the feasible region shows they do overlap (points further along the unbounded direction satisfy both). Because a common point exists, $Z$ can be pushed below $-300$ somewhere in the feasible region.

Step 5. Conclusion: since the test half-plane does meet the feasible region, $Z = -50x + 20y$ has no minimum value subject to these constraints — the naive smallest table entry is a trap. This is exactly the subtlety JEE examiners like to probe: always check boundedness before declaring a corner-point winner final. Timing note: on an unbounded-region question, spend the first 20 seconds just classifying bounded vs unbounded before computing anything — if unbounded, budget an extra 30-40 seconds for the half-plane existence check on whichever extreme (max or min) the question actually asks for; do not waste time checking the one it doesn't ask about.

Common traps

A very common error is forgetting the non-negative restrictions $x \geq 0, y \geq 0$ when shading the feasible region — this silently drags in the second, third, and fourth quadrants and produces a wrong (usually much larger and wrong-shaped) region. Another frequent slip is applying the Corner Point Method's "largest table value is the maximum" shortcut to an unbounded region without running the half-plane test — Example 4 above shows this can manufacture a false answer where in fact no maximum or minimum exists. Students also mix up which side of a boundary line is feasible: always test a convenient point (the origin, if it is not on the line) directly in the original inequality rather than guessing from the picture. When finding a corner point that is the intersection of two lines, a common arithmetic trap is solving the wrong pair of equations — always check that the point you compute is actually a vertex of the shaded region, not the intersection of two lines that happens to fall in the infeasible zone. Finally, when two corner points tie for the optimal value of $Z$ , students often report only one of them; the correct statement is that every point of the line segment joining the two tied corners is equally optimal, giving infinitely many optimal solutions, not just two.

You should be able to…

Translate a word problem into decision variables, a linear objective function, and a full set of linear constraints including non-negative restrictions.
Graph a system of linear inequalities in two variables and correctly shade the common feasible region using the test-point method.
State precisely what is meant by feasible region, feasible solution, infeasible solution, bounded region, unbounded region, and optimal solution.
State Theorem 1 and Theorem 2 of this chapter and explain, in your own words, why an optimum of a linear function must occur at a corner point of a convex region.
Execute the Corner Point Method completely: find every corner point, evaluate Z at each, and correctly conclude the maximum and minimum on a bounded region.
Recognise when a feasible region is unbounded and correctly apply the open half-plane existence test before declaring a maximum or minimum.
Identify and correctly interpret a problem with multiple optimal solutions (a tied edge) rather than reporting only a single corner point.
Recognise when a problem has no feasible region and explain why, by pointing to the specific conflicting constraints.
Solve a full linear programming word problem end to end within typical exam time, including a clearly labelled sketch of the feasible region.

Ask the tutor: Want to work through Exercise 12.1's transportation or diet-planning word problems together, step by step?

Why must the objective function and constraints be linear, not quadratic or otherwise nonlinear?

The name "linear programming" is precise: both the objective function $Z = ax + by$ and every constraint must be linear expressions in the decision variables — no squares, products of variables, or roots. This restriction is what makes the geometry so clean: linear inequalities always produce half-planes with straight boundary lines, their intersection is always a convex polygon, and a linear objective function always has its extremes at a vertex, as proved in this chapter. If the objective function were quadratic, for instance, its level curves would be parabolas or ellipses rather than straight lines, the "extreme only at endpoints of an edge" argument would fail, and an interior point genuinely could be optimal — the Corner Point Method would simply not work. Problems with nonlinear objectives or constraints belong to the broader field of nonlinear programming, which needs calculus-based methods beyond this chapter.

How do I know which side of a boundary line to shade for each constraint?

Pick any point that is clearly not on the boundary line — the origin $(0,0)$ is almost always the most convenient choice, provided the line does not itself pass through the origin. Substitute that point's coordinates into the original inequality (not the boundary equation). If the inequality holds true at that point, shade the side of the line containing that point; if it is false, shade the opposite side. For example, for $5x + y \leq 100$ , testing $(0,0)$ gives $0 \leq 100$ , which is true, so the origin's side is feasible. Repeat this test independently for every constraint, then take the region common to all of them — that intersection, further restricted to $x \geq 0, y \geq 0$ , is the feasible region.

What is the difference between a feasible solution and an optimal solution?

A feasible solution is any point $(x, y)$ that satisfies every constraint of the problem, including the non-negativity conditions — there are typically infinitely many of them, filling the entire feasible region. An optimal solution is a special feasible solution: the one (or ones, in the tied case) among all feasible solutions that gives the objective function its best possible value, maximum or minimum depending on what the problem asks for. Every optimal solution is necessarily a feasible solution, but the reverse is false — most feasible solutions are not optimal. The whole point of the Corner Point Method is to skip checking every one of the infinitely many feasible solutions and instead check only the finitely many corner points, which by Theorem 1 are guaranteed to include the optimal one.

Why does an LPP sometimes have no feasible region at all, like NCERT Example 5?

A feasible region is empty exactly when the constraints are mutually contradictory — no point in the plane can satisfy all of them simultaneously. In Example 5, the constraints $x + y \geq 8$ and $3x + 5y \leq 15$ (together with $x, y \geq 0$ ) conflict: the first demands $x$ and $y$ be reasonably large, while the second caps a weighted sum of them at a small value, and no non-negative pair can satisfy both at once. Graphically, the two half-planes required by these constraints simply do not overlap anywhere in the first quadrant. When this happens, the problem has no feasible solution, and consequently no optimal solution exists to report — the correct exam answer is to state clearly that the feasible region is empty and explain briefly why (usually by exhibiting the conflicting requirement).

Do I always have to solve two equations simultaneously to find every corner point?

No — only for the corners that lie at the intersection of two slanted constraint lines. Many corner points are found more quickly by inspection: wherever a constraint line crosses an axis, just set the other variable to zero directly in that line's equation. For instance, in Example 1, $C = (0, 50)$ came straight from setting $x = 0$ in $x + y = 50$ , no simultaneous equations needed. Reserve the elimination-or-substitution method for the corner where two non-axis boundary lines genuinely meet, such as point $B$ in that same example. Correctly identifying which corners come from axis intercepts versus line intersections, and sketching the region first, saves significant time under exam conditions.