The parallel processing design of particle swarm optimization based on TMS320C6678

Abstract

In view of the real-time requirement of PSO in practical applications, the parallelism of the algorithm is analyzed. According to the architecture characteristics of TMS320C6678 multi-core processor, this paper designs a local parallel and global serial model, and maps the application program to multi-core processor efficiently. The experimental data shows that the design gives full play to the performance advantage of C6678 and improves the real-time processing capability of the system effectively. The design promotes the application of PSO algorithm in practice effectively and is of important significance to other kinds of swarm intelligent algorithm.

Keywords

• particle swarm optimization
• TMS320C6678
• parallel processing
• speedup rate