Linear programming is a method for finding the best possible outcome (maximum profit, minimum cost) in a situation with linear constraints — limits expressed as linear inequalities, like a factory having a maximum number of labour hours or raw materials available. Graphing the constraints identifies a feasible region — the area satisfying every constraint simultaneously — and the key insight of linear programming is that the optimal solution (the maximum or minimum) will always occur at one of the feasible region's corner points (vertices), not somewhere in the middle. This turns a potentially complex optimisation problem into a manageable one: just test the objective function at each corner point.
Example
A small business making two products, each requiring different amounts of labour and materials (both limited resources), can graph these constraints as inequalities to find a feasible region of possible production combinations — then, rather than testing every single possible combination, checking the profit function at just the handful of corner points of that region reveals the exact combination that maximises profit.
Key terms
Constraint:
A limit on a variable, expressed as a linear inequality.
Feasible region:
The area on a graph satisfying every constraint simultaneously.
Questions
1. Linear programming is used to find:
The best possible outcome given linear constraints
A completely random result with no optimisation
Only the average of a data set
Something unrelated to constraints or optimisation
2. A constraint is:
A limit expressed as a linear inequality
Something with no effect on a problem
A type of graph colour
Always exactly equal to zero
3. The feasible region is:
The area satisfying every constraint simultaneously
A single, isolated point only
An area with no constraints applied at all
Something unrelated to graphing constraints
4. The optimal solution in linear programming occurs at:
A corner point (vertex) of the feasible region
A completely random point with no pattern
The exact centre of the feasible region always
A point outside the feasible region
5. Linear programming can be used to find:
Maximum profit or minimum cost
Only random, unoptimised outcomes
Nothing related to business or resources
Only the size of a shape
6. A factory's limited labour hours or raw materials are examples of:
Constraints
The objective function itself
Something unrelated to production limits
A type of profit calculation only
7. Testing the objective function at corner points is a strategy used because:
The optimal solution is guaranteed to occur at one of these points
It is always slower than testing every single point in the region
Corner points are always irrelevant to finding an optimal solution
This approach never actually finds the correct answer
8. Why does graphing multiple linear constraints together create a feasible region rather than a single line?
Each linear inequality defines a half-plane, and the overlap of all of them together forms a bounded (or unbounded) region of all jointly valid solutions
Graphing multiple linear constraints together always produces exactly one single line with no region
Constraints in linear programming never actually overlap or combine to form any kind of region
A feasible region can only ever be created using a single constraint, never multiple combined
9. Why is it more efficient to test the objective function only at the corner points of the feasible region, rather than at every point within it?
The mathematical properties of linear functions guarantee the optimal value occurs at a vertex, so checking every interior point would be unnecessary extra work
Testing every single point within the feasible region is always faster than only checking the corner points
The optimal solution in linear programming is never actually found at any of the feasible region's corner points
There is no mathematical reason why corner points would be more likely to contain the optimal solution
10. Why might a business have both a minimum requirement (like needing at least 10 units of a product) and a maximum limit (like available labour hours) as separate constraints in the same problem?
Real business situations often involve multiple, sometimes competing, requirements or limits that all need to be satisfied at the same time
A linear programming problem can only ever include a single constraint, never combine multiple types together
Minimum requirements and maximum limits always describe exactly the same type of constraint with no distinction
Business situations never actually involve more than one relevant constraint at the same time
11. Why might a feasible region sometimes be empty (no solution satisfies every constraint at once)?
If constraints directly conflict with each other (like requiring both at least 10 and at most 5 of something), no combination can satisfy all of them simultaneously
A feasible region is always guaranteed to contain at least one valid solution, no matter what the constraints are
Constraints in a linear programming problem can never actually conflict with each other
An empty feasible region has no real mathematical or practical explanation
12. Why might a feasible region sometimes be unbounded (extending infinitely in some direction), and why could this be a problem for maximising an objective function?
If constraints don't fully enclose the region, values could increase without limit in that direction, meaning a maximum might not actually exist, even though a minimum still could
A feasible region is always fully bounded and enclosed, regardless of what constraints are included in the problem
An unbounded feasible region never actually causes any issue when trying to find a maximum value for the objective function
Whether a feasible region is bounded or unbounded has no bearing on whether an optimal solution can be found
13. Why might the objective function in a linear programming problem (like "maximise profit = 3x + 5y") need to be a linear expression, consistent with the term "linear" programming?
The method's guarantee that the optimum occurs at a corner point specifically relies on both the constraints and the objective function being linear, not curved or more complex
The objective function in a linear programming problem can always be any type of expression, linear or otherwise, with identical results
Linear programming works identically well regardless of whether the objective function or constraints are linear or non-linear
The term "linear" in linear programming has no actual connection to any requirement about the objective function's form
14. Why might a company use linear programming to decide how to allocate a limited advertising budget across different platforms?
It provides a systematic way to maximise a measurable outcome (like reach or sales) while respecting a fixed budget constraint and other real limitations
Linear programming has no genuine application to real-world resource allocation decisions like advertising budgets
A limited advertising budget can never actually be modelled using linear constraints
Businesses always allocate resources like advertising budgets through pure guesswork, with no systematic method
15. Why might real-world optimisation problems sometimes require additional techniques beyond basic linear programming, such as when relationships between variables aren't actually linear?
Linear programming specifically assumes straight-line relationships, so genuinely non-linear real-world relationships (like diminishing returns) would require different mathematical tools
Every real-world optimisation problem can always be accurately modelled using only linear relationships and constraints
Non-linear relationships between variables never actually occur in real-world optimisation problems
Linear programming techniques work identically well regardless of whether real relationships are linear or non-linear
16. Why might a linear programming problem with three or more variables (rather than just two) be significantly harder to solve using a simple graph?
Graphing becomes impractical beyond two dimensions, so three-or-more-variable problems typically require algebraic or computational methods instead of visual graphing
Problems with three or more variables are always exactly as easy to graph and solve as two-variable problems
The number of variables in a linear programming problem never has any bearing on how difficult it is to solve
Three-or-more-variable optimisation problems can never actually be solved by any method
17. Why might understanding which constraint is "binding" (limiting the optimal solution) at the final answer be valuable information for a business beyond just knowing the optimal solution itself?
Knowing which specific constraint is limiting the outcome reveals where additional resources (like more labour hours) would have the most impact if the business could obtain them
Identifying which constraint is binding provides no additional useful information beyond simply knowing the final optimal solution
Every constraint in a linear programming problem is always equally binding and equally limiting on the final solution
Binding constraints have no real practical relevance to business decision-making beyond the mathematics itself
18. Why might linear programming be considered a foundational technique that underlies more advanced optimisation methods used in fields like logistics, supply chain management or economics?
The core concepts (constraints, feasible regions, optimising an objective) extend into more complex, real-world optimisation problems, even when those problems require more advanced mathematical tools
Linear programming has no genuine connection to any more advanced optimisation techniques used in other fields
Fields like logistics and supply chain management never actually make use of any optimisation techniques
More advanced optimisation methods in other fields share no conceptual foundation with basic linear programming
19. Why might a delivery company use linear programming to decide how to allocate a fleet of trucks across different routes, subject to constraints like fuel budget and driver hours?
It provides a systematic way to minimise cost or maximise deliveries while respecting genuine operational limits, rather than relying on guesswork for a complex, multi-constraint decision
Linear programming has no genuine application to real-world logistics or fleet allocation decisions
Fuel budget and driver hours can never actually be modelled as constraints in a linear programming problem
Delivery companies always allocate their fleets through pure trial and error, with no systematic optimisation method
20. Why might changing just one constraint (like increasing available labour hours) potentially shift which corner point of the feasible region produces the optimal solution?
Altering a constraint changes the shape and boundaries of the feasible region itself, which can move where the optimal vertex is located and change the resulting best outcome
Changing a single constraint never actually has any effect on the shape of the feasible region or the location of the optimal solution
The feasible region's corner points always remain in exactly the same position regardless of any change to a constraint
Only changing the objective function itself, never a constraint, can ever shift the optimal solution
21. Understanding linear programming and optimisation mainly helps you to:
Systematically find the best outcome within a set of linear constraints using feasible regions and corner points
Assume the optimal solution to a linear programming problem could occur anywhere within the feasible region
Ignore the possibility that a feasible region could be empty due to conflicting constraints
Treat linear programming as having no real-world business or resource allocation application
Answer key (parent copy)
1. The best possible outcome given linear constraints
2. A limit expressed as a linear inequality
3. The area satisfying every constraint simultaneously
4. A corner point (vertex) of the feasible region
5. Maximum profit or minimum cost
6. Constraints
7. The optimal solution is guaranteed to occur at one of these points
8. Each linear inequality defines a half-plane, and the overlap of all of them together forms a bounded (or unbounded) region of all jointly valid solutions
9. The mathematical properties of linear functions guarantee the optimal value occurs at a vertex, so checking every interior point would be unnecessary extra work
10. Real business situations often involve multiple, sometimes competing, requirements or limits that all need to be satisfied at the same time
11. If constraints directly conflict with each other (like requiring both at least 10 and at most 5 of something), no combination can satisfy all of them simultaneously
12. If constraints don't fully enclose the region, values could increase without limit in that direction, meaning a maximum might not actually exist, even though a minimum still could
13. The method's guarantee that the optimum occurs at a corner point specifically relies on both the constraints and the objective function being linear, not curved or more complex
14. It provides a systematic way to maximise a measurable outcome (like reach or sales) while respecting a fixed budget constraint and other real limitations
15. Linear programming specifically assumes straight-line relationships, so genuinely non-linear real-world relationships (like diminishing returns) would require different mathematical tools
16. Graphing becomes impractical beyond two dimensions, so three-or-more-variable problems typically require algebraic or computational methods instead of visual graphing
17. Knowing which specific constraint is limiting the outcome reveals where additional resources (like more labour hours) would have the most impact if the business could obtain them
18. The core concepts (constraints, feasible regions, optimising an objective) extend into more complex, real-world optimisation problems, even when those problems require more advanced mathematical tools
19. It provides a systematic way to minimise cost or maximise deliveries while respecting genuine operational limits, rather than relying on guesswork for a complex, multi-constraint decision
20. Altering a constraint changes the shape and boundaries of the feasible region itself, which can move where the optimal vertex is located and change the resulting best outcome
21. Systematically find the best outcome within a set of linear constraints using feasible regions and corner points