Design, Containerization and Performance of Distributed Evolutionary Computation

The development of parallel and distributed algorithms has long been an important branch of evolutionary computation research. Whether to balance workloads across multiple computational nodes for enhanced runtime performance, or to enable innovative algorithmic strategies like coevolutionary approaches that optimize convergence behavior, the importance of parallel and distributed computation is undeniable. Although several software projects support parallel and distributed evolutionary computation, there remains potential to enhance the usability, flexibility, and runtime performance in this domain. This paper presents an open-source software system that incorporates an algorithm design language and containerization technology to address these challenges. We present the software’s architecture including its graphical user frontend, microservice backend, and the container runtime environment. Further on, we describe application scenarios and results from various tests regarding computational speed and data exchange, which demonstrate the system’s applicability and its ability to facilitate future research directions.