FICO Xpress Optimization is a software suite for developing and deploying optimization models. It comprises a set of solvers, including linear, mixed-integer, and various nonlinear solvers (local and global) as well as a constraint programming solver. Users have the flexibility to choose among several options for formulating their optimization models: direct interfaces to the solver libraries are available for several programming languages such as C, Java, Julia, .NET, Python, R. Alternatively, the user can opt to work with an algebraic modelling language, most prominently FICO Xpress Mosel, a free software that forms part of the Xpress suite and that provides access to all solvers of the Xpress suite.

Model development is perhaps the most intellectually challenging part of practical mathematical optimization. But having once developed an algebraic model of the business problem at hand, OR practitioners are faced with deciding how to implement the model inside some larger IT system. Modelling APIs aim at high speed to market by easing tasks such as verifying model correctness, model maintenance and modification, and they typically also provide support for data access and manipulation. Algorithmic considerations (problem types, support for more advanced tasks such as infeasibility handling or callbacks) also are of importance. Not least, model execution speed might equally be considered among the evaluation criteria.

This presentation discusses a non-comprehensive list of open-source third party modelling APIs for Xpress, including Pyomo, PuLP, Google OR-Tools, Linopy, PyOptInterface, Jump, AMPL as well as Xpress' Python API on top of which some of these APIs are built and Xpress Mosel.