The LP Procedure
- Overview
-
Getting Started
-
Syntax
-
Details
Missing Values Dense Data Input Format Sparse Data Input Format Converting Any PROC LP Format to an MPS-Format SAS Data Set Converting Standard MPS Format to Sparse Format The Reduced Costs, Dual Activities, and Current Tableau Macro Variable ORLP Pricing Scaling Preprocessing Integer Programming Sensitivity Analysis Range Analysis Parametric Programming Interactive Facilities Memory Management Output Data Sets Input Data Sets Displayed Output ODS Table and Variable Names Memory Limit
-
Examples
An Oil Blending Problem A Sparse View of the Oil Blending Problem Sensitivity Analysis: Changes in Objective Coefficients Additional Sensitivity Analysis Price Parametric Programming for the Oil Blending Problem Special Ordered Sets and the Oil Blending Problem Goal-Programming a Product Mix Problem A Simple Integer Program An Infeasible Problem Restarting an Integer Program Alternative Search of the Branch-and-Bound Tree An Assignment Problem A Scheduling Problem A Multicommodity Transshipment Problem with Fixed Charges Converting to an MPS-Format SAS Data Set Migration to OPTMODEL: Assignment Migration to OPTMODEL: Multicommodity Transshipment
- References
| Input Data Sets |
In addition to the DATA= input data set, PROC LP recognizes the ACTIVEIN= and the PRIMALIN= data sets.
ACTIVEIN= Data Set
The ACTIVEIN= data set contains a representation of the current active tree. The format is identical to that of the ACTIVEOUT= data set.
PRIMALIN= Data Set
The format of the PRIMALIN= data set is identical to the PRIMALOUT= data set. PROC LP uses the PRIMALIN= data set to identify variables at their upper bounds in the current solution and variables that are basic in the current solution.
You can add observations to the end of the problem data set if they define cost (right-hand-side) sensitivity change vectors and have PRICESEN (RHSSEN) types. This enables you to solve a problem, save the solution in a SAS data set, and perform sensitivity analysis later. You can also use the PRIMALIN= data set to restart problems that have not been completely solved or to which new variables have been added.