XA

This document describes the GAMS/XA linear and mixed-integer programming solver. The GAMS/XA solver (here also simply referred to as XA) is based on Sunset Software Technology’s XA Callable Library, an implementation of high performance solvers for LP and MIP problems. XA implements primal simplex, dual simplex, and barrier algorithms for solving linear problems. The primal/dual simplex method is very robust, and in most cases you should get good performance, especially from a warm start. The barrier method is particularly efficient on large models. Both algorithms benefit from XA’s presolver, which reduces the size of the model by removing redundant contraints, subsituting constraints, etc. In most cases, GAMS/XA should perform satisfactorily without using any options. However, if fine-tuning is necessary or desired, XA provides many options and parameters designed for this purpose. These options are accessible via GAMS option statements or via an XA-specific option file.