Flexible and efficient agent-based metaheuristic computing.

Agent-based metaheuristics computing paradigm (EMAS) has been proposed over 20 years ago by Cetnarowicz. Since then, many efforts were made in order to evaluate, formally analyze and further develop this paradigm towards creating new algorithms as EMAS hybrids, or EMAS-inspired techniques. However, at the same time a significant work has been done in order to build efficient software frameworks supporting this (and similar) computing paradigms. These frameworks were based not only on classic object-oriented programming, but also on functional approach and recently also utilizing heterogeneous infrastructure. This paper presents an overview of the most important findings in this area, including novel ways of processing the agents and component orientation, which allow for both high flexibility and high efficiency of provided solutions. The discussed concepts are illustrated with a case study of a system solving hard computational problem leveraging GPGPU.