PMF: A multicore-enabled framework for the construction of metaheuristics for single and multiobjective optimization

Abstract

This paper describes the design and implementation of the Parallel Metaheuristics Framework (PMF), a C++ framework for the construction of single and multiobjective metaheuristics utilizing Intel's Threading Building Blocks library to allow easy parallelization of computationally intensive algorithms. The framework demonstrates a generic approach to the construction of metaheuristics, striving to provide a general representation of core operators and concepts, thus allowing users to more easily tailor the system for novel problems. The paper describes the overall implementation of the framework, demonstrates a case study for implementing a simple metaheuristic within the system, and discusses a range of possible enhancements. © 2010 Springer-Verlag.