Algebra › Section D · Systems of linear equations › Topic D.8
Systems of linear inequalities
One inequality shades half the plane; a system keeps only what every inequality allows. That surviving region — the feasible region — is where planning, budgeting, and optimisation live.
From half-planes to overlap
A single two-variable inequality graphs as a half-plane: the boundary line, shaded on the satisfying side. A system of inequalities asks for points that satisfy every inequality at once, and its graph is therefore the overlap of all the half-planes. Real planning problems arrive exactly this way — a budget cap, a labour limit, a minimum quality standard — each one a half-plane, and the workable plans all sitting inside the shared region.
Graphing each inequality, one at a time
The machinery carries over unchanged from single inequalities. Draw each boundary line: solid for \(\le\) or \(\ge\), whose points count as solutions, and dashed for strict \(<\) or \(>\), whose boundary points do not. Choose a test point not on the line — the origin whenever the line avoids it — and shade the side that makes the inequality true. Repeat per inequality on the same axes; the final picture is read from the common overlap, not from any single shading.
The solution is the feasible region
The overlap where every shading agrees is called the feasible region: every point inside satisfies all inequalities simultaneously, and every point outside fails at least one. To confirm any candidate point, substitute it into every inequality in the system — one failure disqualifies it. For \(x \ge 0\), \(y \ge 0\) and \(x + y \le 8\), the point \((3, 4)\) passes all three since \(7 \le 8\), while \((4, 5)\) fails the third since \(9 > 8\), and \((-2, 10)\) fails the first two.
Worked example: a region with three corners
Graph the system \(x \ge 0\), \(y \ge 0\), \(x + y \le 8\). The first two inequalities restrict attention to the first quadrant; the third shades below the line from \((8, 0)\) to \((0, 8)\), drawn solid since the boundary is included. The feasible region is the triangle with corners \((0, 0)\), \((8, 0)\) and \((0, 8)\). Testing the interior point \((2, 3)\): both coordinates are nonnegative and \(2 + 3 = 5 \le 8\), so it is a solution; the nearby point \((4, 5)\) is not, since \(9 > 8\).
Corners: where boundaries cross
The corners (vertices) of a feasible region matter more than any other points, and they are found by algebra: each corner is the crossing of two boundary lines, so solve the corresponding pair of equations as a system. The corner of \(x + y = 8\) with the axis \(y = 0\) is \((8, 0)\); with \(x = 0\) it is \((0, 8)\); and the axes cross at \((0, 0)\). In regions bounded by more constraints, test which crossings actually survive every inequality — a crossing that fails one constraint is not a corner of the region.
Linear programming: optimising over a region
Linear programming asks for the best value of a linear objective — maximise profit, minimise cost — subject to a system of constraints. Its central result is the corner principle: a linear objective on a polygonal feasible region takes its maximum and minimum at corners. The reason is geometric: sweeping a level line of the objective across the region, the last point it touches is a corner. So the recipe is short — graph the region, locate its corners, evaluate the objective at each, and pick the winner.
Worked example: a bakery's best batch
A bakery makes cookie batches (profit $5 each, 2 hours of oven time) and brownie batches (profit $3 each, 1 hour), with 14 oven hours available and at most 9 batches total. With \(x\) cookie and \(y\) brownie batches, the constraints are \(2x + y \le 14\), \(x + y \le 9\) and \(x, y \ge 0\). The corners: \((0, 0)\); \((7, 0)\) where \(2x + y = 14\) meets \(y = 0\); \((5, 4)\) where \(2x + y = 14\) crosses \(x + y = 9\) (subtracting gives \(x = 5\)); and \((0, 9)\). Evaluating the profit \(P = 5x + 3y\): $0, $35, $37 and $27. The maximum is \(37\) dollars at 5 cookie batches and 4 brownie batches.
Checking the winning corner
The optimal point must survive every constraint, so verify it as though the optimisation were a rumour. For the bakery: oven time \(2(5) + 1(4) = 14\), exactly the available hours; batches \(5 + 4 = 9\), exactly the limit; both variables nonnegative. The plan is feasible, and by the corner principle no other feasible point beats it. A full sweep of the profit line \(5x + 3y = k\) across the region touches the plane last at \((5, 4)\) — the corner principle at work, with the level line's slope lying between the two boundary slopes, which is exactly why a corner wins.
Common mistakes to avoid
First, shading the wrong side of a boundary — always run a test point, every time. Second, drawing a dashed boundary for \(\le\) or \(\ge\), or solid for strict inequalities; the line's style states whether boundary points count. Third, reporting the union of the shadings rather than the overlap: a system's region is where all inequalities hold, not where any holds. Fourth, accepting a boundary crossing as a corner without checking it against every inequality. Fifth, in optimisation, testing interior points instead of corners — a linear objective's extremes never hide in the interior.
Key takeaways
A system of linear inequalities graphs as the overlap of half-planes, one per inequality, each drawn with a solid boundary for included lines and dashed for excluded ones, shaded via a test point. The surviving region is the feasible region, and a point belongs exactly when it satisfies every inequality. Corners come from solving pairs of boundary equations, and only crossings that pass all constraints count. Linear programming optimises a linear objective over the region, and the corner principle guarantees the best value occurs at a corner — the recipe behind Dantzig's simplex method and decades of industrial scheduling.
Flashcards
Flashcard
A point must satisfy every inequality
Test your understanding
Question 1 of 5
Which point lies in the feasible region of the system x >= 0, y >= 0, x + y <= 8?
Question 2 of 5
What are the corners of the region defined by x >= 0, y >= 0, x + y <= 8?
Question 3 of 5
At what point do the boundary lines 2x + y = 14 and x + y = 9 cross?
Question 4 of 5
A linear objective is maximised over a polygonal feasible region. Where can the maximum occur?
Question 5 of 5
Profit is P = 5x + 3y and the corners of the feasible region are (0, 0), (7, 0), (5, 4) and (0, 9). Which corner maximises profit?
