Automatic Selection of Best Optimizer
If you are unfamiliar with the relationship between problem characteristics and optimizer speed, you may prefer to let CPLEX determine the best algorithm to use to optimize your problem. Most models are well solved by the default optimizer selected by calling optimize / cplex.solve() / CPXlpopt(). We recommend using this option unless you wish to tune performance.
Under defaults, CPLEX solves an LP model using the dual simplex method. The primal simplex method is available as an alternative optimizer and can be faster on some models. To determine whether your problem contains a network, try the network optimizer as well. Additionally, if you are licensed to use it, we suggest that you try the primal-dual logarithmic barrier optimizer (that is, the ILOG CPLEX Barrier Optimizer); it is applicable to very large, sparse problems, particularly those with a block-matrix structure. The following sections say more about each linear optimizer.
