[FlopCpp] Using FLOPC++ with GLPK MathProg examples

[timsmail at chippingtonderrick.co.uk]

Hi

A perhaps obvious suggestion - try imposing the solution from one model in the other model. This can be quite simply done by adding constraints which directly force the values for each variable. You may need to make sure that you have a good enough precision in the numbers that you get from one model to feed into the other, but it should work. If the models are compatible, then they should both work with both solutions. If one solution is not compatible with the other model, try removing the fixed values constraints in turn or in groups until you find the origin of the problem.

It can be a bit tedious, but it should find the problem.

Tim

>This does indeed look like it is the same model.
>
>My guesses would be, in this order:
>1) You looked wrong and changed 3 and 6
>2) There is some small difference in the models we havent noticed
>3) Your version of GLPK MathProg does not work correctly.