摘要
为了进一步增强微粒群算法的优化性能,提出了一种改进微粒群算法,并将其用于求解梯级水库群的优化调度。该算法引进了类似遗传算法的交叉和变异算子来提高搜索效率,其中交叉是微粒在解空间中的位置以一定的概率随机进行算术交叉,变异是微粒以一定的概率随机使速度矢量的某一维分量变为0。为了加速收敛,初始微粒群生成时采用了有条件的随机自动生成方式,并利用惩罚函数法来处理边界条件和其它非等式约束。实例计算结果表明,改进微粒群算法具有比常规动态规划法和常规微粒群算法更快的计算速度,且优化调度结果比较满意。
An improved PSO algorithm is proposed and applied to the optimal operation of cascade reservoirs.To increase the searching efficiency,the improved PSO introduces a crossover operator and a mutation operator,both similar to those of generic algorithm.The crossover operator makes the particles arithmetically crossover their positions in solution space and the mutation operator changes one velocity component of the particles into zero,both in a random manner at a certain probability.To accelerate convergence,the initial particle swarm is generated by a stochastic method under certain conditions,and a penalty function is used to implement the boundary conditions and inequality constraints.A study case of cascade reservoirs shows faster convergence and satisfactory optimal operation results of the improved PSO in comparison with the traditional dynamic programming algorithm or the traditional PSO.
出处
《水力发电学报》
EI
CSCD
北大核心
2009年第4期49-55,共7页
Journal of Hydroelectric Engineering
基金
国家自然科学基金委员会
二滩水电开发有限公司雅砻江水电联合研究基金项目(50579095)
关键词
水电工程
梯级水库群
优化调度
改进微粒群算法
hydropower engineering
cascade reservoirs
optimal operation
improved PSO