基于模态空间及空间压缩理论的粒子群算法在输电网络规划中的应用

The Particle Swarm Optimization Based on Model Space and Space Compression Theory and Its Application in Power Network Expansion Planning

提出了模态空间及空间压缩理论,将每次搜索到的一个满足N-1原则的网络定义为一个模态,并定义模态空间.将该理论应用于粒子群算法(PSO),通过防止重复搜索已搜索的模态空间对整个可行空间进行压缩,提高了算法的全局搜索效率,并提出了模态结构变异法,为算法提供了跳出局部最优的途径.通过算例与普通PSO算法的分析比较,证明了模态理论的正确性和应用模态理论的PSO算法具有全局、快速的搜索性能.

A model space and space compression theory was presented. The theory defines an N-1 security network as a model, and then defines a model space according to the given model. This theory can prevent particle from searching in the found areas and improve its search efficiency. A model structure mutation method is also put forward, which can make particle jump out of the local optimum. The comparison of the performance between basic particle swarm optimization （PSO） and model based PSO （MPSO） demonstrate these methods are correct and efficient.