A Particle Swarm Optimization Algorithm with Differential Evolution

• Published in: 2007 International Conference on Machine Learning and Cybernetics Vol. 2; pp. 1031 - 1035
• Main Authors: Zhi-Feng Hao, Guang-Han Guo, Han Huang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2007
• Subjects: Benchmark testing; Chromium; Computer science; Cybernetics; Data mining; Differential evolution; Evolutionary computation; Explosions; Global minimization problem; Machine learning; Machine learning algorithms; Particle swarm optimization
• ISBN: 1424409721; 9781424409723
• ISSN: 2160-133X
• DOI: 10.1109/ICMLC.2007.4370294

Differential evolution (DE) is a simple evolutionary algorithm that has shown superior performance in the global continuous optimization. It mainly utilizes the differential information to guide its further search. But the differential information also results in instability of performance. Particle swarm optimization (PSO) has been developing rapidly and has been applied widely since it is introduced, as it can converge quickly. But PSO easily got stuck in local optima because it easily loses the diversity of swarm. This paper proposes a combination of DE and PSO (termed DEPSO) that makes up their disadvantages. DEPSO combines the differential information obtained by DE with the memory information extracted by PSO to create the promising solutions. Finally, DEPSO is tested to solve several benchmark optimization problems. The experimental results show the effectiveness of DEPSO algorithm for the multimodal function, and also verify that DEPSO can perform better than other algorithms (DE, CPSO) in solving the benchmark problems.