Jump to content

Math340/Solution6

From UBC Wiki

Question 1 (Tuesday June 16)

There are two different trucks T1,T2 and three routes r1,r2,r3. T1 can only be used for the first two routes, while T2 can be used for all three, so we have the decision variables (corresponding to edges):

x11,x12,x21,x22,x23. These are all binary variables, that is:

xij={1Truck i is assigned to route j,0otherwise.

The objective:

minimizez=i,jci,jxij.

The constraints:

jxij1 for i=1,2,

ixij1 forj=1,2,3.

Question 2 (Wednesday, June 17)

Our problem here would be:

maximize5x1+4x2+3x3subject to2x1+3x2+x36,4x1+x2+2x311,3x1+4x2+2x38x1,x2,x30,x1,x2,x3 integers.

The LP relaxation is the same thing without the integer constraints:

maximize5x1+4x2+3x3subject to2x1+3x2+x36,4x1+x2+2x311,3x1+4x2+2x38x1,x2,x30.

And we computed in the previous homework that the solution to this is:

z=1313 and x1=223,x2=0,x3=0.

This isn't integer, so we branch into problem 2: maximize5x1+4x2+3x3subject to2x1+3x2+x36,4x1+x2+2x311,3x1+4x2+2x38x1,x2,x30,x12.

And problem 3: maximize5x1+4x2+3x3subject to2x1+3x2+x36,4x1+x2+2x311,3x1+4x2+2x38x1,x2,x30,x13.

These you can solve in your favourite method (I used a web solver) to get:

The solution for problem 2 is z=13, x=(2,0,1), while problem 3 is infeasible. So, the optimal solution is the solution to problem 2, x=(2,0,1).