|
Integer programming is a quite desirable formulation technique. But, integer problems can be hard to solve due to search nature of solution process
Three approaches can help
| • | Reformulate reflecting as much problem knowledge in the formulation as possible improving the depiction of the |
| — | Way the integer variables are tied together by the constraints. For example, entering constraints that reflect that if one size of machine is chosen in the first stage of an assembly line that it must be matched with a comparable machine in the second stage. |
| — | Way the integer and continuous are tied together by the constraints. For example, one could achieve benefits by entering constraints indicating that the sum of the continuous variables depicting volume through a warehouse whose construction is depicted by integer variable representing warehouse must be no more than capacity times the integer variable and no less than say 75% of capacity times the integer variable. |
| — | Restricting the values of the integer variables eliminating "unnecessary" cases of integer variables. For example, in a warehouse location problem if only one warehouse can practically be built require the sum across the variables to be one. Similarly if experience says at least 2 machines are necessary but no more than 5 then enter such constraints. |
| • | Use MIP solver features through options and GAMS. Sometimes spectacular reductions in solution time can be achieved in very little time by changing solver branch and bound procedures. |
| • | Expand available resources or allowable iterations. |
Also these solves are generally slower so one must be patient.
|