Polyrun: A Java library for sampling from the bounded convex polytopes

Abstract

Polyrun is a Java library that provides methods for exploiting the bounded convex polytopes. Such polytopes define a space of feasible problem parameters with a set of linear constraints. The software makes available an implementation of the Hit-and-Run algorithm, which is the Markov Chain Monte Carlo method for an efficient uniform sampling from the convex polytopes. Moreover, it implements other procedures, such as Ball-Walk, Sphere-Walk, or Grid-Walk, for making random steps within a polytope. The software provides a Java Application Programming Interface (API) along with an intuitive Command Line Interface (CLI) for Hit-and-Run. It has been used to support real-world decision making in various application areas, including logistics, land use planning, nanotechnology, and energy. The software is free and open-source.

Keywords

• Monte Carlo simulation
• Hit-and-Run
• Convex polytope
• Uniform sampling
• Linear constraints
• Java