Concise cplex

Abstract

This paper is a concise guide to CPLEX, the leading solver for linear and convex quadratic optimisation problems. The paper is self contained and includes information for first time CPLEX users as well as code snippets and lemmas that may be of referential value to experienced users. The paper starts with a brief explanation of how to run CPLEX on departmental servers at Imperial and on standalone machines in section 1, how to create and solve simple Linear Programs in section 2 and how to obtain detailed solution results in section 3. The paper then moves on to discuss several CPLEX issues and quirks that may confuse first time users including: anomalous objective values caused by big-M scaling, the implications of long MILP solution times and removing memory limitations for problems with large MILP solution trees. The paper concludes with logical equivalence proofs in section 9 that can be used as a starting point for complex problem translation and references are provided for additional reading.