基于图论的抚河流域水资源多目标空间优化配置研究

水资源空间优化配置研究对于保障区域水资源可持续与合理高效利用具有重要意义。针对江西省抚河流域降雨时空不均和开发利用过度所造成的水资源短缺现状,利用多目标空间优化的建模思想研究该流域水资源优化配置问题。通过对抚河流域系统进行概化,构建流域网络模型,考虑经济、社会和环境影响从而设置三个相应的目标函数,并根据流域特点及节点设置列出模型的约束条件,建立基于图论的抚河流域水资源多目标空间优化配置模型。在流域现状年和规划年供需平衡分析的基础上,确定以流域较为缺水的95%来水频率下的供需水量作为优化配置基准,进行折衷规划算法(CP模型)求解,得到抚河流域各县(区)水资源优化配置结果,基本满足各县(区)需水条件,验证了模型的可行性与实用性。

Graph theory-based study on multi-objective spatial optimal allocation of water resources in Fuhe River Watershed

Research of spatial optimal allocation of water resources is of great significance to sustainable，reasonable and effective utilization of regional water resources． Aiming at the status of water resources shortage caused by uneven spatial-temporal distribution of rainfall and the excessive development and utilization of water resources in Fuhe River Watershed of Jiangxi Province，the optimal allocation of the water resources in the watershed is studied herein by means of the modeling thought of multi-objective spatial optimization． Through generalizing the Fuhe River Watershed system，establishing the watershed network model and considering the economic，social and environmental impacts，three corresponding objective functions are set，and then the constraints of the model are listed out and a graph theory-based multi-objective spatial optimal model for the water resources in Fuhe River Watershed is established in accordance with the characteristics and the node settings of the watershed． On the basis of the analyses of supply and demand balances in both current-year and planning-year，the water supply and demand under the incoming-water frequency of 95%，which is relatively short of water in the watershed，is determined as the optimal allocation benchmark and then is solved with the compromise programming algorithm ( CP model) ，thus the optimal allocation results of all the counties ( districts) within Fuhe River Watershed are obtained，from which not only the water demand requirements of all the counties ( districts) can be satisfied，but the feasibility and practicability of the model can also be verified．