基于CBR的水库洪水调度模式

Study on reservoirs flood dispatching based on Case-Based Reasoning

由于水库洪水调度问题的复杂性和不确定性,单纯利用调度模型进行洪水调度在实际应用中往往存在一些问题.为了使理论研究尽可能与实际应用相结合,将人工智能中的基于事例推理(CBR)技术引入到水库洪水调度中,将以往的洪水调度事例作为历史事例以一定的结构和方式存储在事例库中,对于新的洪水调度问题,从事例库中寻找相似的事例,并根据其调度方案,确定新问题的解决方案.实际算例表明,对于水库洪水调度问题,基于事例推理的水库洪水调度能充分的将以往的调度经验和调度模型结合起来,使得调度结果更合理、可行,具有较强的实用性.

Considering the complexity and uncertainty of reservoir flood dispatching, the flood dispatching which is uniquely basing on dispatching model has some problems in practiee. In order to achieve the integration of theory research and practice application, this paper introduces the ease-based reasoning technology of artificial intelligence into the reservoir flood dispatching, which make the former flood dispatching cases as historical cases store in the case library with certain structure and pattern. To the new problems encotmtered in flood dispatching, it can search the similar eases in the ease library and establish the solutions basing on the dispatching scheme for those new problems. The numerical tests indicated that this method has its better practicality. It can effectively integrate the historical dispatching experience to dispatching model, and make the practical dispatching result more reasonable and feasible.