Hybrid Particle Swarm Optimization Based on Thermodynamic Mechanism

Authors:
Yu Wu;Yuanxiang Li;Xing Xu;Sheng Peng
Affiliations:
State Key Lab. of Software Engineering, Wuhan University, Wuhan, China 430072;State Key Lab. of Software Engineering, Wuhan University, Wuhan, China 430072;State Key Lab. of Software Engineering, Wuhan University, Wuhan, China 430072;State Key Lab. of Software Engineering, Wuhan University, Wuhan, China 430072
Venue:
SEAL '08 Proceedings of the 7th International Conference on Simulated Evolution and Learning
Year:
2008

Citing 2
Cited 0

Comparison between Genetic Algorithms and Particle Swarm Optimization

EP '98 Proceedings of the 7th International Conference on Evolutionary Programming VII
The particle swarm - explosion, stability, and convergence in amultidimensional complex space

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a thermodynamic particle swarm optimizer (TDPSO) based on the simple evolutionary equations. Inspired by the minimum free energy principle of the thermodynamic theoretics, a rating-based entropy (RE) and a component thermodynamic replacement (CTR) rule are implemented in the novel algorithm TDPSO. The concept of RE is utilized to systemically measure the fitness dispersal of the swarm with low computational cost. And the fitness range of all particles is divided into several ranks. Furthermore, the rule CTR is applied to control the optimal process with steeply fast convergence speed. It has the potential to maintain population diversity. Compared with the other improved PSO techniques, experimental results on some typical minimization problems show that the proposed technique outperforms other algorithms in terms of convergence speed and stability.