基于事例推理技术的水库洪水调度研究与应用  

Research and application of reservoir flood dispatching based on CBR techniques

作　　者：罗军刚[1] 解建仓[1] 陈田庆[1] 汪妮[1] 

机构地区：[1]西安理工大学西北水资源与环境生态教育部重点实验室,陕西西安710048

出　　处：《水科学进展》2009年第1期32-39,共8页Advances in Water Science

基　　金：国家高技术发展计划(863)资助项目(2006AA01A126);国家自然科学基金资助项目(50279041);西安理工大学优秀博士学位论文研究基金资助项目(207-210008)~~

摘　　要：将基于事例推理(CBR)技术应用到水库洪水调度中,提出了一种基于CBR的水库洪水调度模式和方法。利用关系数据库实现了洪水调度事例的表示、组织、索引和存储,并建立了水库洪水调度事例库。将遗传禁忌算法与最近相邻法结合起来,构建了基于遗传禁忌算法和最近相邻法相结合的混合检索算法,提高了事例检索的速度和质量。利用遗传算法对事例特征属性权重进行优化,采用多目标决策方法进行事例的优选,并给出了事例调整和学习的策略。最后,应用该方法开发了基于CBR的水库洪水调度系统,并给出了一个应用实例。基于CBR的水库洪水调度为水库洪水调度提供了一种简化知识获取、提高调度效率和质量、进行知识积累和重用的新的思路和方法。This paper proposes a reservoir flood dispatching mode and method based on the case-based reasoning（CBR） techniqnes with the artificial intelligence （AI）. By adopting relational database to resolve the representation, organization, index and storage of the flood dispatching cases, a database related to reservoir flood dispatching case is built. Based on the combination of the genetic algorithm and tabu search and nearest neighbor algorithm, a hybrid approach is constructed, improring the speed and quality of the case retrieval. The strategy of the case modification and study is presented by using the genetic algorithm to optimize case characteristics weight and adopting the multi-objective decision method to optimally select the case. Finally, the reservoir flood dispatching system based on CBR is developed. The research shows that for the reservoir flood dis： patching the case-based reasoning method provides a new thought and method to simplify the knowledge acquisition, improve the efficiency and quality of dispatching, accumulate and reuse the knowledge concerned.

关 键 词：基于事例推理 洪水调度 事例检索 遗传算法 

分 类 号：TV697.1[水利工程—水利水电工程] TP18[自动化与计算机技术—控制理论与控制工程]