Constrained optimisation
Write one objective in one variable, enforce its domain and compare every candidate before interpreting the result
- Separate objective and constraint Name the quantity to maximise or minimise. Use the given relation to remove other variables; differentiating the constraint itself does not optimise the objective
- State the valid domain Use physical conditions and every additional restriction. A stationary input outside this domain cannot be selected
- Compare all admissible candidates On a closed interval, evaluate every stationary input inside the interval and both included endpoints. Compare exact objective values
- Recover the original quantities The optimiser is an input, while the optimum is an output. Substitute back into the original constraint for any other requested dimensions and state their units
- Use the right geometric model A closed square-base box has surface area 2x² + 4xH and volume x²H. An open cut-out box instead has a different surface constraint and cubic volume
- Squared distance preserves the minimiser Distance is nonnegative and square root is increasing. Minimising squared distance gives the same input as minimising distance, while avoiding a square root during differentiation
Worked example
A rectangle has perimeter 30 m. Find both side lengths and the maximum area
- \(\displaystyle y=15-x,\quad A(x)=15x-x^2,\quad0\le x\le15\)Eliminate the companion length and include zero-area endpoints for comparison
- \(\displaystyle A'(x)=15-2x=0\quad\Rightarrow\quad x=15/2\)The stationary input is permitted
- \(\displaystyle A(0)=A(15)=0,\quad A(15/2)=225/4\)The stationary area exceeds both endpoint areas
- \(\displaystyle x=y=15/2\text{ m},\quad A_{\max}=225/4\text{ m}^2\)Recover the other side and distinguish length from area
Common mistakes
Difficulty levels
Questions about this topic
Can the optimum be an endpoint?+
Yes. A restriction can exclude an otherwise optimal stationary input
Why recover a second dimension?+
The chosen input determines it through the original constraint; reporting both verifies the contextual solution
Are physical zero dimensions always permitted?+
No. Follow the stated domain. Some models include degenerate endpoints for comparison; a positive-dimension box requires positive lengths
Can I use a calculator?+
No. This topic is practised without a calculator
Which year level and chapter is this?+
Year 11, VCE Units 1 & 2, Mathematical Methods, chapter 13: Differentiation and applications