摘要
粒子群优化(particle swarm optimization,PSO)算法模拟鸟群或鱼群中生物体的运动行为,是一类优秀的元启发式算法。PSO算法的研究现状是进行自适应多策略的探索。所谓多策略是指采用多种策略分别实现保持多样性、逃脱停滞/局部极值、加速收敛和局部搜索等目的,而自适应是指根据种群/粒子的演化状态动态地更新各策略中用到的关键参数以及恰当地进行策略的调用、转换和设置。通过对文献中各种自适应多策略PSO算法进行综述,分析得出PSO算法的发展趋势是结合维和更小尺度的搜索经验知识进行自适应多策略的研究。
Particle swarm optimization( PSO) is a powerful class of meta-heuristics. PSO simulates the movements of organisms in a bird flock or fish school. The research status quo of PSO is the investigation of adaptive multi-strategy. Multi-strategy refers to the use of multiple strategies in order to realize preserving diversity,escaping from stagnation / local optimum,accelerating convergence and local search. Adaptive means dynamically updating key parameters involved in each strategy and appropriately invoking /switching / setting the strategies. Based on a survey of various adaptive multi-strategies PSO algorithms proposed in literature,this paper comes to the conclusion that the future research trend of PSO is to study adaptive multi-strategy through incorporating search experience knowledge at the dimension and smaller scales.
出处
《南昌工程学院学报》
CAS
2016年第3期71-75,共5页
Journal of Nanchang Institute of Technology
基金
国家自然科学基金资助项目(61261039
51209008)
江西省教育厅科学技术研究项目(GJJ151099)
关键词
粒子群优化
自适应
多策略
综述
particle swarm optimization
adaptive
multi-strategy
review