An Improved Particle Swarm Optimization Algorithm using Gaussian Mixture Model in Dynamic Environment

Many problems in the real world due to the local and global optimization change over time are a matter of dynamic optimization. Therefore, optimization algorithms despite global optimization and tracking of environments changed over time are needed in these environments. We are faced with two problems in designing the particle swarm optimization algorithm for dynamic environments to find the best solution in a short time and follow the solution after the environmental changes: Outdated memory and loss of population diversity in search area. The problem of loss of population diversity is one of the biggest challenges in dynamic environments because diversifying a convergent population to find dynamic optimization and then turning it into a new optimization greatly reduces the efficiency of the algorithm. Given the challenges presented in this paper, a hybrid particle swarm optimization algorithm based on a Gaussian mixture model is proposed. In the proposed method, the change of each particle is based on the result of the best particles in its cluster. The results of the Moving Peak Benchmark experiments show a better performance of the proposed algorithm than other algorithms.