摘要
为了克服标准微粒群算法在求解资源受限项目调度问题上存在的早熟现象,提出一种改进的文化微粒群算法。该算法框架基于微粒群算法的主群体空间和文化算法的知识空间,两种空间具有各自的群体并可独立并行演化,形成双演化双促进机制,提高了算法的全局搜索能力和运行效率。同时为了避免文化算法知识空间自我演化限制,引入遗传算法的演化机制来改进知识空间的演化操作。通过具体的算例比较,验证了提出的改进文化微粒群算法在求解资源受限项目问题时的有效性。
In order to overcome the premature phenomen of standard particle swarm optimization（PSO） for solving RCPSP,this paper proposed an improved cultural particle swarm optimization（ICPSO）algorithm.The framework of the proposed algorithm was based on the main population space of the PSO and the knowledge space of the cultural algorithm（CA）,where two spaces having respective groups as well as evolving independently,and formed the mechanism of ＂double evolution,double promotion＂.At the same time,in order to avoid the restriction of self-evolving by the knowledge space of the CA,it introduced the evolutinary mechanism of the genetic algorithm（GA）into the knowledge space to improve its evolutionary operations.A specific comparing example verifies the validity of the ICPSO for solving the RCPSP.
出处
《计算机应用研究》
CSCD
北大核心
2013年第1期90-93,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(70871088)
关键词
文化微粒群算法
资源受限项目调度问题
知识空间
主群体空间
cultural particle swarm optimization（CPSO） algorithm
resource-constrained project scheduling problem（RCPSP）
knowledge space
main population space
