Adaptive Particle Swarm Optimization Algorithm Based on Cloud Model

Adaptive Particle Swarm Optimization based on Cloud Model (CM-APSO) is an intelligent optimization method that integrates cloud model theory. By leveraging the cloud model's uncertainty handling capabilities, it enhances the adaptability and convergence efficiency of traditional Particle Swarm Optimization (PSO).

In conventional PSO, velocity and position updates rely on fixed inertia weights and learning factors, which may cause premature convergence or insufficient search efficiency. CM-APSO utilizes the three numerical characteristics of cloud models—Expectation (Ex), Entropy (En), and Hyper-Entropy (He)—to dynamically adjust algorithm parameters, achieving a better balance between global exploration and local exploitation.

Specifically, Ex guides the swarm's search direction, En determines the dynamic adjustment range of parameters, and He controls the randomness of parameter variations. This mechanism enables adaptive inertia weight adjustment based on population diversity during iterations, preventing premature convergence. Meanwhile, the cloud-generated randomness enhances the algorithm's ability to escape local optima.

Compared to traditional PSO, CM-APSO demonstrates superior performance in optimizing complex multimodal functions and dynamic parameter optimization scenarios. Its core innovation lies in combining cloud models' quantitative uncertainty handling with swarm intelligence's collaborative search, providing a novel improvement approach for intelligent optimization algorithms. Future research could explore its potential in industrial scheduling and hyperparameter optimization for machine learning applications.