A variant of group search optimizer for global optimization

作者：Chen, Jia-Jia; Ji, Tianyao; Wu, Peter; Li, Mengshi^*

来源：Journal of Computational Methods in Sciences and Engineering, 2016, 16(2): 219-230.

DOI：10.3233/JCM-160614

• Adaptive learning
• Benchmark functions
• Comparison result
• Exploration and exploitation
• Group search optimizer (GSO)
• Multimodal benchmark
• Optimization problems
• Population-based optimization

Group search optimizer (GSO) is a stochastic, population-based optimization technique that has shown better performance as for global searching when optimizing multimodal benchmarks. However, it suffers from poor convergence because of its producer-scrounger model, which makes it unable to get a reliable estimiator for evolution path. In order to enhance the local search ability without impairing the global seach ability of GSO, this paper proposes a variant of group search optimizer (VGSO), which develops a producer-organizer model whereby group members are assumed to search for opportunities of either 'finding'(producing) or 'learning'(organization). This model enables VGSO to achieve a good trade-off between its exploration and exploitation abilities. A comprehensive experimental study is conducted on a set of benchmark functions and a practical optimization problem. Comparison results show that VGSO obtains a promising performance on these optimization problems.

• 出版日期2016-6-6
• 单位华南理工大学; 山东大学