Homework #2 Optimization

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

MATH 319 / ENGR 419

Homework #2
Questions
1. Mathematical Analysis [Luenberger, 10.8 (see also, Chaps. 6 and 10); Vanderplaats, 1.6]
State the optimality conditions for a constrained optimization problem. (In other words, what
conditions are necessary and sufficient to ensure that a minima has been found?) Your answer
should include both the mathematical equations that define optimality as well as a brief discussion
of what the equations mean. (You may find it helpful to include a sketch or two in your discussion.)
2. Hand Calculations
Consider the following optimization problem. We wish to minimize the function, f = x12 + x22, by choosing
the values of x1 and x2. The solution must also satisfy the relationships: x1 + x2 4 and x1 + x2 2. The
parameter, , has a value of 0.5.
a. Rewrite this problem in standard form.
b. Solve this problem by hand using the 1st-order KKT (i.e., optimality) conditions.
Confirm that your solution satisfies the KKT conditions.
3. Software Implementation
Instructions: You may partner with one classmate for part (a). You must complete part (b) on your own.
List your project partner and submit your code to the D2L dropbox.
a. Method Implementation
i. Write a Matlab script to minimize univariate functions using Newtons method.
ii. Write a Matlab script to minimize univariate functions using the golden section method.
iii. Write a Matlab script to call Matlabs line search algorithm, FMINBND.
b. Method testing
(This section should be completed on your own.)
i. Using all three methods, minimize f(x) = ex-5x. Start from initial points x0=-4 and x0 = 4.
ii. Using all three methods, minimize f(x) = -x cos(x) from initial points x0=0 and x0 = 1.
iii. Using your favorite method, solve the application problem listed below.
For parts (i)-(iii), decide when x* has converged. Count (and compare) the number of function
evaluations required for each method, including each calculation of f, f or f. Discuss your
results. Which method was most efficient, and why? What combinations of starting points and
methods did not converge, and why?
Application Problem: Climb performance is an important design criterion for aircraft. An aircrafts rate of
climb (R/C) is dependent on both engine performance (i.e., power available, PA) and aerodynamic
characteristics (i.e., power required, PR). In specific, R/C = (PA-PR) / W. A Cessna 152 has the following
characteristics: PA = 41,580[1-2900V-2.35] ft*lb/s, PR = [0.0049V3 + (1675)2 / (3.243V)] ft*lb/s, and W =
1675 lb. What is its maximum rate of climb, and at what velocity does it occur?

You might also like