A two-level particle swarm optimization: profiling and software/hardware implementation

Particle swarm optimization (PSO), a population-based stochastic method, was developed to address difficult optimization issues like the flexible job shop scheduling problem. In this work, the aim is to analyze the Two-Level PSO algorithm and identify the most time-consuming function for hardware implementation while minimizing a particular objective function. For designers who want to include metaheuristics in general and PSO in particular, this work is particularly helpful. The implementation of the selected functions as hardware accelerators, also known as Intellectual Propriety, during the Design Space Exploration phase causes the users of these algorithms’ tremendous anxiety. This study serves as a dashboard for programmers who want to optimize the way their apps use metaheuristics.