Linear Inequalities & Feasible Regions

SECTION 1

Historical Context & Motivation

The study of linear inequalities and their geometric counterparts, feasible regions, arose from humanity's longstanding need to allocate scarce resources under competing constraints. While ancient civilizations solved specific allocation problems through trial and error, the formal mathematical framework for reasoning about inequality constraints did not crystallize until the twentieth century, when wartime logistics and industrial planning demanded systematic methods for optimization. The resulting theory — anchored in the geometry of half-planes and their intersections — now pervades economics, operations research, engineering design, and data science.

1826

Fourier's Inequality Theory

Joseph Fourier published early work on systems of linear inequalities, establishing foundational ideas about the solution sets of inequality systems and methods for eliminating variables — what would later be called Fourier–Motzkin elimination.

1939

Kantorovich's Linear Programming

Soviet mathematician Leonid Kantorovich formulated resource-allocation problems as linear programs, recognizing that optimal solutions live on the boundary of a feasible region defined by linear inequality constraints.

1947

Dantzig's Simplex Method

George Dantzig developed the simplex algorithm, which traverses the vertices of the feasible region (a convex polytope) to find optimal solutions efficiently — bringing feasible-region geometry to practical computation.

1975

Finite Math Curricula Emerge

Universities began offering Finite Mathematics courses integrating linear inequalities, linear programming, and matrix methods as essential quantitative reasoning tools for students in business, social science, and the life sciences.

The central question this lesson addresses is deceptively simple: given a set of linear inequality constraints, what does the collection of all solutions look like geometrically, and how do we find it? Answering this question provides the visual and algebraic foundation for linear programming, where one seeks the best outcome (maximum profit, minimum cost) within a feasible region.

SECTION 2

Core Principles & Definitions

Before graphing systems of inequalities, it is essential to establish the vocabulary and the fundamental geometric ideas that connect algebraic expressions to regions of the coordinate plane. A linear inequality in two variables is any statement of the form ax + by ≤ c (or with <, ≥, >) where a, b, and c are real constants and at least one of a or b is nonzero. Each such inequality partitions the plane into two half-planes separated by the boundary line ax + by = c. The solution set of the inequality is one of those half-planes, possibly including the boundary line itself (when the inequality is non-strict, i.e., ≤ or ≥).

Half-Plane

A linear inequality divides ℝ² into two half-planes. The solution set is the half-plane containing all (x, y) that satisfy the inequality. A solid boundary line indicates ≤ or ≥; a dashed line indicates strict inequality (< or >).

System of Inequalities

A system of linear inequalities is a collection of two or more inequalities considered simultaneously. A point (x, y) is a solution of the system if and only if it satisfies every inequality in the system.

Feasible Region

The feasible region (or solution region) is the intersection of all half-planes defined by the system's inequalities. Geometrically, it is a convex polygonal region (or unbounded convex set) in the plane.

Corner Points (Vertices)

Corner points are the vertices of the feasible region, formed where boundary lines intersect. In linear programming, optimal values of an objective function always occur at one or more corner points.

Bounded vs. Unbounded

A feasible region is bounded if it can be enclosed in a circle of finite radius, and unbounded otherwise. Systems with constraints like x ≥ 0 and y ≥ 0 without upper limits on both variables often produce unbounded regions.

✦ KEY TAKEAWAY

Think of each linear inequality as a fence on a field: the fence itself is the boundary line, and only one side of the fence is "in bounds." When you have multiple fences (a system), the feasible region is the patch of ground that lies inside every fence simultaneously. The shape of that patch — and particularly its corners — is where all the action happens in optimization.

SECTION 3

Graphing a Single Linear Inequality

The diagram below illustrates the solution set of the inequality 2x + 3y ≤ 12. The boundary line 2x + 3y = 12 is drawn as a solid line (because the inequality is non-strict), and the half-plane below and to the left is shaded to indicate all points satisfying the inequality. A test point such as the origin (0, 0) confirms the shading direction: 2(0) + 3(0) = 0 ≤ 12 is true, so the origin's side is the solution side.

The solid cyan line represents the boundary 2x + 3y = 12, with intercepts at (0, 4) and (6, 0). The shaded region below the line contains all points (x, y) for which 2x + 3y ≤ 12. The gold dot at the origin confirms the shading direction via the test-point method.

The procedure for graphing any single linear inequality in two variables can be summarized in three steps. First, graph the boundary line by converting the inequality to an equation and plotting it (solid for ≤ or ≥, dashed for < or >). Second, choose a convenient test point not on the boundary — the origin works unless the line passes through it — and substitute its coordinates into the inequality. Third, shade the half-plane containing the test point if the inequality is satisfied, or the opposite half-plane if it is not.

SECTION 4

Mathematical Framework

The algebraic treatment of linear inequalities rests on a small number of manipulation rules that parallel those for equations, with one critical difference involving multiplication or division by negative numbers. Mastery of these rules is essential for rearranging inequalities into slope-intercept or standard form for graphing.

GENERAL LINEAR INEQUALITY (TWO VARIABLES)

ax + by ≤ c (or <, ≥, >)

where a, b, c ∈ ℝ, and (a, b) ≠ (0, 0). The boundary line is ax + by = c.

SLOPE-INTERCEPT FORM (SOLVED FOR y)

y ≤ (−a/b)x + c/b when b > 0

If b > 0, dividing by b preserves the inequality direction: the solution is the half-plane at or below the line y = (−a/b)x + c/b. If b < 0, dividing by b reverses the inequality, and the solution lies at or above the resulting line.

INTERSECTION OF HALF-PLANES (FEASIBLE REGION)

F = H₁ ∩ H₂ ∩ … ∩ Hₙ

The feasible region F is the intersection of half-planes H₁, H₂, …, Hₙ defined by the system's n inequalities. Each Hᵢ is a convex set, and the intersection of convex sets is convex; hence F is always convex (or empty).

⚠️ Inequality Reversal Rule

When you multiply or divide both sides of an inequality by a negative number, the direction of the inequality must be reversed. For example, −2x ≥ 6 becomes x ≤ −3 after dividing by −2. Forgetting this reversal is one of the most common algebraic errors in inequality manipulation.

To find the corner points of a feasible region, solve pairs of boundary-line equations simultaneously. Each pair of intersecting boundary lines may produce a candidate vertex; a candidate is an actual vertex of the feasible region if and only if it satisfies all other inequalities in the system. The set of vertices, together with the edges connecting them, defines the polygonal boundary of the feasible region.

SECTION 5

Systems of Inequalities & Feasible Region Geometry

When multiple linear inequalities act simultaneously, their solution set is the region of the plane that lies in every individual half-plane at once. The diagram below depicts the feasible region for the system: x + y ≤ 5, x ≥ 1, y ≥ 0. Each constraint carves away a portion of the plane, and what remains — the intersection — is the feasible region. Notice that the resulting region is a triangle with three corner points, each formed by the intersection of two boundary lines.

The feasible region (shaded in amber-to-emerald gradient) is the triangle with vertices at (1, 0), (5, 0), and (1, 4). The violet vertical line represents x = 1, the emerald horizontal line represents y = 0 (the x-axis), and the pink diagonal line represents x + y = 5. Only the region satisfying all three constraints simultaneously is shaded.

Corner-point computation for the system

Boundary Pair	System Solved	Intersection Point	In Feasible Region?
x + y = 5 and x = 1	1 + y = 5 → y = 4	(1, 4)	Yes ✓
x + y = 5 and y = 0	x + 0 = 5 → x = 5	(5, 0)	Yes ✓
x = 1 and y = 0	Direct substitution	(1, 0)	Yes ✓

The convexity of the feasible region is a fundamental geometric property: if you pick any two points inside the region and draw the line segment connecting them, every point on that segment also lies inside the region. This property is not incidental — it follows from the fact that each half-plane is convex, and the intersection of convex sets is always convex. Convexity is precisely the structural property that makes the simplex method and other linear programming algorithms efficient.

SECTION 6

Worked Example: Finding & Graphing a Feasible Region

Consider the system of linear inequalities arising from a simple production scenario: a small bakery produces cakes (x) and pies (y), subject to constraints on flour, oven time, and non-negativity.

Determine the Feasible Region

Step 1 — State the SystemThe bakery's constraints translate to the following system: (1) 2x + y ≤ 8 (flour constraint, in kg) (2) x + 2y ≤ 10 (oven-time constraint, in hours) (3) x ≥ 0 (non-negativity) (4) y ≥ 0 (non-negativity)

Step 2 — Graph Boundary LinesConvert each inequality to an equation. For 2x + y = 8: the x-intercept is (4, 0) and the y-intercept is (0, 8). For x + 2y = 10: the x-intercept is (10, 0) and the y-intercept is (0, 5). Draw both lines as solid lines, along with the axes x = 0 and y = 0.

Step 3 — Shade Half-PlanesTest the origin in each inequality: 2(0) + 0 = 0 ≤ 8 ✓ and 0 + 2(0) = 0 ≤ 10 ✓. So the origin is on the solution side of both inequalities. Shade the region that is below/left of both lines and in the first quadrant (x ≥ 0, y ≥ 0).

Step 4 — Find Corner PointsSolve each pair of boundary-line equations: • (0, 0): intersection of x = 0 and y = 0. • (4, 0): intersection of 2x + y = 8 and y = 0 → 2x = 8, x = 4. • (0, 5): intersection of x + 2y = 10 and x = 0 → 2y = 10, y = 5. • Intersection of 2x + y = 8 and x + 2y = 10: multiply the first equation by −2 to get −4x − 2y = −16, then add to x + 2y = 10 to obtain −3x = −6, so x = 2. Then y = 8 − 2(2) = 4, giving (2, 4).

Check (2, 4): 2(2) + 4 = 8 ≤ 8 ✓, and 2 + 2(4) = 10 ≤ 10 ✓. Check (0, 5): 2(0) + 5 = 5 ≤ 8 ✓, and 0 + 2(5) = 10 ≤ 10 ✓.

Step 5 — Identify the Feasible RegionThe feasible region is the quadrilateral with vertices (0, 0), (4, 0), (2, 4), and (0, 5). It is a bounded, convex polygon in the first quadrant. Any point strictly inside this polygon satisfies all four inequalities, and the bakery could produce that combination of cakes and pies without exceeding its resource limits.

Feasible region vertices: (0, 0), (4, 0), (2, 4), (0, 5)

💡 Why Corner Points Matter

The Corner-Point Theorem (a consequence of convexity) states that if a linear objective function has a maximum or minimum value on a bounded feasible region, that optimum occurs at a corner point. This theorem is the theoretical justification for the simplex method and means that, in practice, one need only evaluate the objective function at the vertices rather than at infinitely many points.

SECTION 7

Strengths, Limitations & Common Pitfalls

Aspect	Strengths	Limitations / Pitfalls
Visual Clarity	In two variables, the feasible region can be drawn on a standard coordinate plane, making the solution set immediately tangible and aiding intuition.	Beyond three variables, graphical methods are infeasible; algebraic or computational approaches (simplex, interior-point) become necessary.
Algebraic Manipulation	Inequalities follow rules similar to equations, making them accessible to students with algebra proficiency.	The sign-reversal rule for multiplication/division by negatives is frequently forgotten, leading to incorrect solution sets.
Corner-Point Theorem	Reduces an infinite search to a finite check of vertices — enormously efficient for bounded regions.	Requires the region to be bounded for guaranteed max/min; unbounded regions may lack an optimum for certain objective functions.
Test-Point Method	A quick, reliable way to determine which side of a boundary line to shade.	If the boundary line passes through the origin, the origin cannot serve as the test point; an alternative must be chosen.
Modeling Power	Linear constraints model a wide range of real-world scenarios: budgets, capacities, blending, scheduling.	Real constraints may be nonlinear; linear inequalities approximate but do not capture all scenarios (e.g., economies of scale).

🔍 PUTTING IT IN PERSPECTIVE

Linear inequalities and feasible regions occupy a pivotal place in the broader optimization landscape. They provide the entry point into linear programming, which is itself the simplest member of a family of constrained optimization methods that includes integer programming, nonlinear programming, and semidefinite programming. Mastering the graphical two-variable case builds the geometric intuition needed to understand algorithmic approaches in higher dimensions.

SECTION 8

Connection to Linear Programming & Advanced Theory

The feasible region studied in this lesson is precisely the domain over which linear programming optimizes a linear objective function Z = c₁x + c₂y. Once you can identify and graph the feasible region, the next step is to evaluate Z at each corner point and determine which vertex yields the maximum or minimum value. This progression — from graphing inequalities to optimizing — represents the natural arc of a finite mathematics course.

Feature	This Lesson (Intro)	Linear Programming (Next Step)
Goal	Identify and graph the set of all feasible solutions	Find the point in the feasible region that maximizes or minimizes an objective function
Output	Shaded region + list of corner points	Optimal point + optimal value of Z
Number of Variables	Typically 2 (graphical method)	2 (graphical) to thousands (simplex/interior-point)
Key Theorem	Each inequality defines a convex half-plane	Corner-Point Theorem: optimum is attained at a vertex
Computational Tool	Graph paper or graphing software	Simplex algorithm, interior-point methods, LP solvers

In higher dimensions, the feasible region generalizes from a polygon to a convex polytope — a higher-dimensional analogue bounded by hyperplanes. The simplex method navigates from one vertex of this polytope to an adjacent vertex of higher objective value, exploiting the same convexity and corner-point logic you have learned here. Duality theory, sensitivity analysis, and integer programming further extend the framework, but all rest on the foundational understanding of how linear inequalities carve out feasible regions.

SECTION 9

Practice Problems

PROBLEM 1 — CONCEPTUAL

Explain geometrically why the feasible region formed by a system of linear inequalities is always convex. What property of individual half-planes guarantees this?

PROBLEM 2 — BASIC CALCULATION

Graph the inequality 3x − 2y ≥ 6. Identify the intercepts of the boundary line, state whether the boundary is solid or dashed, and determine which side of the line to shade using a test point.

PROBLEM 3 — INTERMEDIATE

Find all corner points of the feasible region defined by: x + y ≤ 6, 2x + y ≤ 10, x ≥ 0, y ≥ 0. Determine whether the region is bounded or unbounded.

PROBLEM 4 — APPLIED

A furniture shop builds tables (x) and chairs (y). Each table requires 4 hours of labor and 3 units of wood; each chair requires 2 hours of labor and 1 unit of wood. The shop has at most 24 hours of labor and 15 units of wood per day, and must produce at least 2 chairs. Write the system of inequalities, find the corner points of the feasible region, and determine the feasible combination that maximizes total items produced (x + y).

PROBLEM 5 — CRITICAL THINKING

Prove that if the feasible region F of a system of linear inequalities is nonempty and bounded, then F has at least three corner points (i.e., it is at least a triangle). Under what conditions could F have exactly one point?

SUMMARY

Summary & Key Concepts

A linear inequality in two variables partitions the coordinate plane into two half-planes separated by a boundary line. The solution set is determined by graphing the boundary (solid for ≤ or ≥, dashed for < or >), selecting a test point, and shading the appropriate side. When manipulating inequalities algebraically, remember the critical rule: multiplying or dividing by a negative number reverses the inequality direction.

A system of linear inequalities produces a feasible region — the intersection of all individual half-planes — which is always a convex set. The corner points (vertices) of this region are found by solving pairs of boundary-line equations and verifying each candidate against all remaining constraints. These vertices are the key to linear programming, where the Corner-Point Theorem guarantees that any linear objective function attains its optimum at a vertex of a bounded feasible region.