You are on page 1of 19

Practice Problem Set 2

OA4201 Nonlinear Programming


Solve problems 1-8 using min/max decomposition. For practice, try solving each of them
from a couple of different starting points (i.e., initial guesses for y).

1. (This one is a slight variation an example from lecture.)


min max (1  y1 )u1  (3  2 y2 )u2  (5  3 y3 )u3
yY u

s.t. u1  2u2  4u3  4


u1 , u2 , u3  {0,1}

where Y  { y1 , y2 , y3  {0,1} | y1  y2  y3  1}

2.
min max yu1  (3  3 y )u2
0  y 1 u

s.t. u1  u2  1
u1 , u2  0

3.
Practice Problem Set 2 Solution

OA4201 Nonlinear Programming


Solve problems 1-8 using min/max decomposition. For practice, try solving each of them
from a couple of different starting points (i.e., initial guesses for y).

1. (This one is a slight variation an example from lecture.)

min max (1  y1 )u1  (3  2 y2 )u2  (5  3 y3 )u3


yY u

s.t. u1  2u2  4u3  4


u1 , u2 , u3  {0,1}

where Y  { y1 , y2 , y3  {0,1} | y1  y2  y3  1}

Solution: start with an initial guess of y = [1 0 0]/. The subproblem is:

max 3u2  5u3


u

s.t. u1  2u2  4u3  4


u1 , u2 , u3  {0,1}

The optimal solution to the subproblem is u1 = [0 0 1]/, with optimal objective value 5
(our new upper bound). The master problem is then:
min z
y,z

s.t. z  5  3 y3
y1  y2  y3  1
y1 , y2 , y3  {0,1}

The optimal solution to the master problem is y* = [0 0 1]/ with objective value z* = 2
(our new lower bound).

You might also like