Overview

The MIP solution pool (XPRSmipsolpool) stores the column solution values for multiple solutions. No slack or dual information is stored with the solutions. The solutions will usually be for MIP problems although the pool will store any arbitrary length vector of double values. Solutions can be read into the MIP solution pool by the user via the file reader routine XPRS_msp_readslxsol or from memory via the XPRS_msp_loadsol routine. Solutions can also be captured automatically from problems (XPRSprob) as they are found during a search.

Since the XPRSprob problems only store the best MIP solution the MIP solution pool is useful when the user is interested in more than one MIP solution for a problem. This can happen when there are constraints or costs not reflected in the problem that the user wants to use to select a solution from a set that have been found by the FICO Xpress Optimizer.

Once a MIP solution pool contains some solutions the user can query for the solution values and also for more abstract attributes such as the objective value of the solution with respect to a given problem. Apart from querying information for a single solution the user has various options for managing the set of stored solutions. The user can query for lists of stored solutions. They may delete solutions from the pool and output solutions to file. Also, a set of policies are available to the user for automatically handling the exclusion of solutions that are duplicated.

The following sections discuss the MIP solution pool functionality and provide some hints on usage.