摘要
根据梯级水电站优化调度特点,建立并行遗传算法求解多阶段最优化问题数学模型,重点研究了粗粒度并行遗传算法。在基于单向环结构交流局部最优解的并行遗传算法(PGA)研究的基础上,提出了基于解的多样性自适应地调节信息交流周期策略的PGA。应用开发的分布式水库群优化调度并行计算系统,将上述两种策略的PGA和串行PGA分别应用于雅砻江梯级水库群优化调度中。通过对其优化结果的分析表明,PGA有利于提高运算速度和求解精度以及改善算法的收敛性能。
A mathematical model is established to solve multi-stage optimization problem by using parallel genetic algorithm（PGA） for an optimal operation of cascade hydropower stations,focusing on the coarse grain algorithm in this work.This model is based on our previous work of PGA that seeks a local optimal solution with a one direction loop.The PGA strategy proposed in this paper adaptively regulates the information exchange cycle based on the diversity of solution.To test the performance of these two strategies,PGA models and serial GA model are applied to the operation of Yalong cascade reservoirs.A comparative analysis of the parallel /serial GA shows the advantages of PGA in speeding up of calculation,better accuracy and improved convergence.
出处
《水力发电学报》
EI
CSCD
北大核心
2010年第6期66-70,共5页
Journal of Hydroelectric Engineering
基金
国家自然科学基金项目(51009029)
广西自然科学基金(2010GXNSFE013005)
广西大学科研基金资助项目(X081098)
关键词
水库调度
并行计算
遗传算法
梯级水电站
reservoir operation
parallel computation
genetic algorithm
cascade hydropower station
