Reliability Optimization Allocation Method Based on Improved Dynamic Particle Swarm Optimization

Abstract

Aiming at the reliability optimization allocation, the basic particle swarm optimization algorithm has the disadvantages of slow convergence and easy to fall into local extremum. This paper proposes an improved dynamic particle swarm optimization algorithm, which uses the extrapolation technique in mathematics to guide the evolution direction of particles. At the same time, the idea of transforming the multimodal function and dynamically adjusting the group size is introduced, and the elite set is used to preserve the optimal individual of each generation. Finally, this paper uses a certain type of sonar as an example to establish a reliability mathematical model, which uses the basic particle swarm optimization algorithm and the improved particle swarm optimization algorithm proposed in this paper. The results show that the proposed method has a stronger search ability, higher accuracy and better stability.