线性微分递减自适应粒子群算法在水电站优化调度中的应用  

Application of Particle Swarm Optimization with Linear Differential Decline Adaptive in Operation Optimization of Hydropower Station

作　　者：万忠海[1] 叶生进 郑姣[1] 

机构地区：[1]四川省都江堰东风渠管理处,四川成都610072

出　　处：《水力发电》2017年第9期85-88,共4页Water Power

摘　　要：针对求解水电站优化调度粒子群算法的改进,分析了粒子群算法在求解水电站优化调度问题时对关键参数惯性权重调整的需要,提出了线性微分递减的自适应粒子群算法。通过前期减小缓慢的惯性权重,增加算法的探索能力跳出局部最优解;通过后期减小较快的惯性权重,提升算法的开发能力加快算法收敛。以葛洲坝水电站优化调度为例,对比了改进算法和传统算法。优化调度实例表明:线性微分递减自适应策略增强了算法的寻优能力和稳定性。改进算法能够有效改善由于水电站优化调度目标函数非凸性带来的粒子群求解易早熟问题,为水电站优化调度粒子群算法惯性权重的改进提供了新思路。For improving the Partic!e Swarm Optimization in solving hydropower station optimal scheduling, the Particle Swarm Optimization with Linear Differential Decline Adaptive is proposed by analyzing the adjustment need of key parameters＇ inertia weight. The improved method can increase the exploration ability to kip local optimal solution by slowly reducing inertia weight in earlier stage, and can improve the development ability to accelerate algorithm convergence through quickly reducing inertia weight in later stage. Taking the optimal operation of Gezhouba Hydropower Station as an example, the improved method and traditional algorithm are compared. The results show that the Linear Differential Decline Adaptive strategy can enhance the search capability and stability of algorithm and the proposed method can improve the premature problem of Particle Swarm Optimization caused by the non-convex objection function of optimal scheduling. The results provide a new idea for improving the inertia weight of Particle Swarm Optimization in hydropower station optimal scheduling.

关 键 词：水电站优化调度 粒子群算法 惯性权重 线性微分递减 

分 类 号：TV697.1[水利工程—水利水电工程]