摘要
针对传统多目标优化算法的不足,提出了基于拥挤距离的多目标粒子群优化算法(MOPSO-CD),该算法用拥挤距离来维持精英策略和选取全局极值,同时引入动态非均匀变异算子,用以维持粒子多样性、减缓算法收敛速度、避免早熟收敛。以漳河水库为例,建立了以灌溉缺水量最小和发电量最大为目标函数的两目标优化调度模型,将MOPSO-CD应用于模型的求解中,得到了足够多且较均匀的非劣(Pareto)解前端。
In view of the disadvantages of the traditional multi-objective optimization algorithm, multi-objective particle swarm optimization based on crowding distance （MOPSO-CD） was proposed. The crowding distance was used to keep the elite strategy and select the global extremum in this algorithm. Meanwhile, the dynamic non-uniform mutation operator was introduced to maintain the diversity of particles, slow the speed of convergence and avoid premature convergence. The two-objective optimal operation model of Zhanghe reservoir was established by taking both minimum irrigation water shortage and maximum electric energy production as objective functions. The MOPSO-CD algorithm was applied to solve this model. Example calculations show that the algorithm could get enough non-inferior solutions and much more uniform fronts.
出处
《水电能源科学》
北大核心
2013年第4期42-45,共4页
Water Resources and Power
基金
国家自然科学基金资助项目(50909073
51179130)
关键词
多目标优化
拥挤距离
基于拥挤距离的多目标粒子群优化算法
动态非均匀变异
漳河水库
multi-objective optimization
crowding distance
multi-objective particle swarm optimization based on crowding distance
dynamic non-uniform mutation
Zhanghe reservoir
