Framing a Design as an Objective Under Constraints
Atlas stands at a large whiteboard in a sunlit engineering lab, sketching a bridge cross-section with one hand while pointing to a numbered list of constraints chalked beside it, a scale model of a truss bridge and a stack of material datasheets spread across the workbench in front of him.
- Identify the single objective function in a given engineering design problem.
- Distinguish between objectives and constraints in a real design scenario.
- Explain why every engineering problem involves trade-offs among competing performance metrics.
- Formulate a simple design problem by stating one objective and at least two quantified constraints.
- Predict how tightening a constraint changes the achievable value of the objective.
Key terms
- Objective function
- The single quantity an engineer chooses to minimize or maximize in a design problem.
- Constraint
- An inequality or equality the design must satisfy; violating any one makes the design infeasible.
- Feasible region
- The set of all designs that satisfy every constraint simultaneously and are therefore legal to consider.
- Decision variable
- A parameter the engineer is free to adjust, such as a beam thickness or a material choice.
- Scalarization
- Collapsing multiple competing objectives into one weighted score so a single optimum can be sought.
Objective Versus Constraint
An objective is the one metric you push as far as possible, while a constraint is a pass-fail line the design must respect. The cleanest test is to ask whether better is always better: a quantity you would happily improve without limit is an objective, whereas a quantity that only needs to clear a threshold is a constraint. Minimize helmet mass is an objective because lighter is always preferred; keep impact force below 400 G is a constraint because once it passes, further reduction earns no extra credit.
The Feasible Region and Infeasibility
Every constraint carves the design space into allowed and forbidden zones, and the intersection of all allowed zones is the feasible region. Only designs inside this region may be ranked by the objective; a design with a brilliant objective value but a single violated constraint is simply illegal. When constraints conflict so severely that no design satisfies all of them, the feasible region is empty and the problem is infeasible, signaling that a requirement must be relaxed before any solution can exist.
Why One Objective at a Time
Two genuinely conflicting objectives, such as low cost and high strength, have no single best solution, because improving one degrades the other along a Pareto frontier of equally defensible designs. Engineers resolve this by promoting the primary goal to the objective and demoting the rest to constraints with acceptable thresholds, or by scalarizing into a weighted combination. Either way the conflict is made explicit, which prevents the impossible promise of optimizing everything at once.
Worked examples
Formulate a design for a shelf bracket as an objective under constraints, then state what happens to the optimum if a constraint is tightened.
- Choose the single objective: minimize the bracket's material mass, since lighter is always preferred.
- List the constraints as pass-fail thresholds: must support at least 150 N without yielding, must fit within a 100 mm projection, manufacturing cost under $5.
- Write the formulation: Minimize mass, subject to load capacity ≥ 150 N, projection ≤ 100 mm, cost ≤ $5.
- Tighten one constraint: raise the required load capacity from 150 N to 250 N.
- Reason about the effect: the feasible region shrinks, so the minimum achievable mass rises (more material is needed) or, if no design can meet 250 N within the other limits, the problem becomes infeasible.
Answer: Minimize mass subject to load ≥ 150 N, projection ≤ 100 mm, cost ≤ $5; tightening load to 250 N raises the minimum mass or renders the problem infeasible.
Activity
Sort each item below into either the Objective or Constraint column for this bridge design problem: minimize total steel mass, span must be at least 20 m, minimize total construction time, deflection under load must not exceed 15 mm, factor of safety must be at least 2.0, construction cost must stay under $500,000.
Practice
For a lightweight bicycle helmet, classify minimize mass, peak force below 400 G, and unit cost below 18 dollars as objective or constraint.
Formulate a phone-stand design as one objective subject to at least two quantified constraints.
Common mistakes to avoid
- You can optimize cost and strength at the same time.Conflicting objectives have no unique optimum, so one must be turned into a constraint or the two scalarized into a weighted score.
- Tightening a constraint can improve the objective.Tightening a constraint only shrinks the feasible region, so the best achievable objective value can stay the same or worsen, never improve.
Check your understanding
An engineer is designing a lightweight bicycle helmet. She wants to minimize helmet mass while keeping the peak impact force transmitted to the head below 400 G and keeping unit manufacturing cost below $18. In this formulation, what is the objective?
A student claims that because both cost and structural strength matter for a bridge design, the engineer should set both as objectives and optimize them simultaneously without treating either as a constraint. What is the strongest argument against this approach?
An engineer tightens a constraint in a bridge design problem, reducing the maximum allowable deflection from 20 mm to 10 mm, while keeping the objective (minimize steel mass) unchanged. What is the most likely effect on the optimal design?
Recap
Engineering design is framed as one objective function to minimize or maximize, subject to constraints that are pass-fail thresholds defining a feasible region. Only feasible designs may be ranked, conflicting objectives must be reduced to one via constraints or scalarization, and tightening any constraint shrinks the feasible region so the optimum can only hold steady or worsen.
Reflect
For a goal you care about, identify the single thing you are truly optimizing and the hard limits you cannot cross.