基于混合粒子群算法的梯级水电站多目标优化调度

Multi-objective optimal dispatch of cascade hydropower stations based on shuffled particle swarm operation algorithm

提出多目标混合粒子群算法以求解梯级水电站多目标联合优化调度模型。该算法采用混合蛙跳算法的分组-混合循化优化框架以增强算法的全局搜索能力;在族群内通过粒子群算法的飞行调整策略指导个体进化;同时,引入外部精英集,建立了基于自适应小生境的外部精英集维护策略,提高了算法的收敛性和非劣解集的多样性。最后将该算法应用于三峡梯级水电站多目标优化调度工程,计算结果表明,本文算法能够获得计算实时性强、分布均匀、收敛性好的调度方案集,并以此分析明确了调度目标间的耦合关系,可为梯级电站的多目标调度决策提供科学依据。

To solve the multi-objective optimal dispatch problem of cascade hydropower stations,a novel multi-objective shuffled particle swarm optimization algorithm （MOSPSO） is presented.MOSPSO uses the partition and shuffling process of shuffled frog leaping algorithm as basic algorithm framework to get rid of local optimal.The individual in memeplexes evolves with the flying strategy of PSO.An archive maintenance strategy with self-adaptive niche calculation method is established to improve the convergence and diversity of solutions.The proposed algorithm is applied to mid-long term optimal dispatch of the Three Gorges cascade hydropower stations.The results show that the algorithm can generate schemes with good convergence and diversity to guide practical dispatch.