摘要
实际工程中存在许多大规模、非线性多约束的序贯决策问题,传统算法解决起来较为困难。蚁群系统(ACS)是一个用来解决大规模多约束组合优化问题的现代启发式算法,根据序贯决策的特点设计了多层结构的蚁群系统,给出了算法的组成结构;为了节约计算内存和优化时间,详细阐述了淘汰劣质解机制的精英策略;并通过梯级水电站短期优化调度这一实际工程序贯决策问题,来验证所构造的算法,给出了优化调度的数学模型及算法的求解思路。最后,采用我国西南地区某梯级流域中三个水电站的相关数据建立了调度仿真模型,仿真结果证实了所采用算法的有效性和可行性。
In practical project, there are many large-scale, nonlinear and multi-constraints sequential decision-making problems （SDP）, which are very difficult to be solved by the traditional approaches. The Ant Colony System （ACS） is a modem heuristic algorithm used to well solve the combination optimization problem of large-scale and multi-constraints. The multi-layer ant colony system （ACS-ML） was constmced based on the SDE and the framework of this algorithm was constmcted. Furthermore, the elite strategy that the worse solutions were eliminated was adopted to saving the memory of the computing and the optimization time. In the end, the practical sequential decision-making problem of the short-term operation scheduling among the cascaded hydroelectric plant was used to validate the algorithm constructed, to this, the mathematics model of the operation scheduling and the solving ideal were proposed. Simultaneity, the scheduling model was built and simulated by applying the relative data of three hydroelectric plants of some cascaded streams, and the simulated results demonstrate the feasibility and effectiveness of the proposed approach.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第6期1444-1447,共4页
Journal of System Simulation
基金
国家自然科学基金资助(60673057)
关键词
序贯决策
蚁群算法
精英策略
梯级水电站
短期优化
sequential decision-making
ACS
elite strategy
cascaded hydroelectric plants
short-term operation
