摘要
将电力系统中机组组合这一复杂的多约束混合整数规划问题分解为具有整型变量和连续变量的两个优化子问题,提出采用改进离散二进制粒子群算法和标准粒子群算法相结合的双层嵌套方法,分别对外层机组的启、停状态变量和内层功率经济分配进行交替迭代优化求解。同时在算法中引入基于机组优先顺序的变异技术和修补策略,能有效地处理机组最短启、停时间约束,并提高算法的全局寻优能力和计算效率。通过对10机系统的算例计算,并同其他算法的结果进行比较分析,仿真结果表明新方法求解精度高、收敛速度快,从而验证了新方法的可行性和有效性。
his paper integrates an improved discrete binary particle swarm optimization(BPSO) with the standard PSO method for solving unit commitment (UC) problem with complicated constraints mixed-integer programming. The UC problem is decomposed into two embedded optimization sub-problems: a unit on/off status schedule problem with integer variables that can be solved by the BPSO method and an economic dispatch problem with continuous that can be solved by the standard PSO method. At the same time the swap mutation operator based on the priority-ranked and repair strategy are introduced in the proposed method, which can be effectively dealt with the minimum up/down time constraints and enhance the algorithm' s global optimal performance and computational efficiency. The feasibility and validity of the new method is demonstrated for 10 -unit system, and the test results are compared with those previously reported methods. Simulation results show that the proposed method performs better in terms of solution's precision and convergence property.
出处
《电力系统自动化》
EI
CSCD
北大核心
2005年第1期34-38,共5页
Automation of Electric Power Systems
基金
国家自然科学基金资助项目(50409010
50309013)中国博士后科学基金资助项目(2003033464)。
关键词
粒子群优化
机组组合
经济调度
群体智能
particle swarm optimization (PSO)
unit commitment
economic dispatch
swarm intelligence