Overview: INTPOINT Procedure

The INTPOINT procedure solves the Network Program with Side Constraints (NPSC) problem (defined in the section Mathematical Description of NPSC) and the more general Linear Programming (LP) problem (defined in the section Mathematical Description of LP). NPSC and LP models can be used to describe a wide variety of real-world applications ranging from production, inventory, and distribution problems to financial applications.

Whether your problem is NPSC or LP, PROC INTPOINT uses the same optimization algorithm, the interior point algorithm. This algorithm is outlined in the section The Interior Point Algorithm.

While many of your problems may best be formulated as LP problems, there may be other instances when your problems are better formulated as NPSC problems. The section Network Models describes typical models that have a network component and suggests reasons why NPSC may be preferable to LP. The section Getting Started: NPSC Problems outlines how you supply data of any NPSC problem to PROC INTPOINT and call the procedure. After it reads the NPSC data, PROC INTPOINT converts the problem into an equivalent LP problem, performs interior point optimization, then converts the solution it finds back into a form you can use as the optimum to the original NPSC model.

If your model is an LP problem, the way you supply the data to PROC INTPOINT and run the procedure is described in the section Getting Started: LP Problems.

You can also solve LP problems by using the OPTLP procedure. The OPTLP procedure requires a linear program to be specified by using a SAS data set that adheres to the MPS format, a widely accepted format in the optimization community. You can use the MPSOUT= option in the INTPOINT procedure to convert typical PROC INTPOINT format data sets into MPS-format SAS data sets.

The remainder of this chapter is organized as follows: