A Survey on Distributed Evolutionary Computation

The rapid development of parallel and distributed computing paradigms has brought great revolution in computing. Thanks to the intrinsic parallelism of evolutionary computation (EC), it is natural to implement EC on these systems. On the one hand, the high computing power can significantly improve the efficiency and scalability of EC. On the other hand, distributed data bring new challenges and development to EC. In this paper, a systematic review on distributed EC (DEC) is given. First, this paper provides an overview of DEC from a broader perspective, including not only the work of improving the efficiency of EC in parallel and distributed platforms, but also the work of applying DEC to solve distributed optimization problems derived from modern distributed environments like the Internet of Things and multi-agent systems. A new taxonomy to review DEC is proposed according to problems, implementation environments, communication models, and algorithms. Based on this taxonomy, problems solved by DEC are categorized into centralized optimization problems and distributed optimization problems according to the information owner. Correspondingly, two major purposes, improving efficiency through parallel processing for centralized optimization and cooperating distributed individuals/sub-populations with partial information for distributed optimization can be distilled. Noting that the latter is an emerging and attractive trend, a systematic definition of distributed optimization is given, including dimension distributed-, data distributed-, and objective distributed-optimization problems. Various DEC studies for these problems are surveyed. Furthermore, challenges, and potential research directions are discussed to enlighten the design of DEC and pave the way to future developments.