Question 1Multiple choiceOptimization & Lagrange multipliers
Use Lagrange multipliers to find the maximum value of subject to the constraint .
- A.
14
- B.
42
- C.
49
- D.
77
Worked solution
Let . At a constrained extremum the gradients are parallel, , which together with the constraint gives four equations in four unknowns:
None of the components of is zero, so and each equation can be solved for a coordinate: , , . Substitute into the constraint to pin down :
gives the point and gives . The sphere is closed and bounded and is continuous, so the maximum and minimum exist and must be among these candidates:
As a cross-check, the Cauchy–Schwarz inequality gives , with equality exactly when is parallel to .
The tempting wrong answer is 42, from putting the whole "budget" into the variable with the largest coefficient, ; that point satisfies the constraint, but the gradients are not parallel there, so it is not an extremum. The answer 77 multiplies the sum of the coefficients by the radius, which has no geometric meaning here, and 14 is the value at .
The maximum value is 49, attained at (option C).