摘要
针对传统的优化算法求解多目标动态环境经济调度(MODEED)模型时极难获得高质量的可行解,且收敛速度慢等问题,根据MODEED模型约束特征,设计了一种约束修补策略;然后将该策略嵌入非支配排序算法(NSGAⅡ),进而提出一种修补策略的约束多目标优化算法(CMEA/R);接着借助模糊决策理论给出了多目标问题的最优决策向量;最后,以经典的10机系统为例,验证了CMEA/R的求解能力,并比较了不同群体规模下CMEA/R与NSGAⅡ的性能。仿真结果表明,在不同群体规模下,与NSGAⅡ相比,CMEA/R的污染排放平均减少了480 lb(217.7 kg),燃料成本平均减少了7 800美元,执行时间平均减少了0.021 s;覆盖率(HR)性能优于NSGAⅡ,且收敛速度较NSGAⅡ快。
The classical multiobjectve optimization algorithm is difficult to achieve high quality feasible solutions on Multiobjectve Dynamic c Dispatch （MODEED） model, and shows a slower convergence speed. Firstly, a new constraint repairing strategy based on the constraint characteristic of MODEED was developed. Secondly, the proposed repairing approach was inserted into the Nondominated Sorting Genetic Algorithm-II （NSGA II ）, and a Constrained Muhiobjective Evolutionary Algorithm based on repairing Strategy （CMEA/R） was proposed. Thirdly, fuzzy decision theory was applied to determine the best compromise solution of the MODEED. Finally, to validate the optimization ability of the CMEA/R, it was applied to solve the MODEED problem of standard IEEE 30-bus 10-generator system, and a comparative analysis with NSGA I1 was presented under various population size. The simulation results revealed that the pollutant emission and fuel cost obtained by CMEA/R were reduced by 480 lb （217.7 kg） and 7 800 dollar, respectively, the average implication time was reduced by 0. 021 second. Furthermore, CMEA/R shows a superior performance in terms of Hypervolume Rate （HR） indicator and convergence ability.
出处
《计算机应用》
CSCD
北大核心
2015年第8期2249-2255,共7页
journal of Computer Applications
基金
国家自然科学基金资助项目(61304146)
贵州省教育厅优秀科技创新人才奖励计划项目(黔教合KY字[2014]255)
贵州省科学技术基金资助项目(20152002)
关键词
电力系统
动态环境经济调度
多目标优化
修补策略
收敛性
power system
dynamic environment/economic dispatch
muhiobjective optimization
repairing strategy
convergence
