Originally posted by: Eumpfenbach

I solve a problem in Matlab, write an MIP start using Cplex.writeMipStart (the idea was that I was trying to write the optimal solution as an MIP start, to test how quickly I could prove optimality if I already had the optimal solution), read this MIP start into the same problem, and Cplex finds a solution that is better than the optimal found with no MIP starts.

Here is my model (too big to attach) http://www.mediafire.com/myfiles.php

The file I use to call cplex is miptestd.m. If I toggle on or off line 47, it is giving me a different optimal solution. I know that it will give an error at line 54. This is just part of a larger algorithm. The problem I am describing occurs before line 54. The optimal solution without an MIP start is -3.86 x 10 ^ 9. With the MIP start, I am seeing an optimal solution that is over -4 x 10 ^ 9.

Please let me know what is going on!
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: Eumpfenbach

MIP start file
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: Eumpfenbach

Proper link for my model

http://www.mediafire.com/?k7vbkif0651xy6i

Also, I am using V12.2.
#CPLEXOptimizers
#DecisionOptimization