城市水资源管理系统动态交互规划研究

A dynamic interactive programming for urban water resources management system

针对不确定性条件下区域水资源管理系统的优化控制,建立了基于模糊可信度约束规划的动态交互规划模型。该模型以四种水资源系统安全状态(相当安全、基本安全、不安全和极不安全)作为约束条件,引入可信度以反应系统的模糊性。将此模型应用于北京市水资源管理系统,并提出交互式模糊满意度算法求解该模型,以全局最优满意度权衡经济和环境目标之间的冲突性。结果表明:当水资源系统安全等级从相当安全降低为极不安全时,规划期内的优化配水量降低了57.95×10^(8) m^(3),系统的经济效益也降低了164.9亿元,污染物排放量显著上升22.05%;可信度水平越高,供水不足和污染物排放过多的风险就越低,经济利益和污染物排放也就越低;可信度越低,经济效益和污染物排放越高,但同时系统风险也在增加。可信度水平能够被视为一个评价指标以评估最终解决方案的可靠水平。相比于传统水资源配置方法,该模型更加真实模拟了多层目标和多决策者的动态交互过程。

Establish a dynamic interactive programming for planning the regional water resources management system under multiple uncertainties.Four security states of water resources system are treated as constraints integrated into the modeling framework,including quite safe,basically safe,unsafe,and extremely far from being safe.Credibility level is in⁃troduced to express the system's fuzzy events.The developed model is then applied in the water resources management system of Beijing.An interactive fuzzy satisfaction algorithm is proposed to solve the developed model,and the conflict between economic and environmental objectives is balanced by the global optimal satisfaction.Results show that when the water resources system security changes from quite safe to extremely far from being safe,the optimal amount of water distribution during the planning period decreases by 57.95×10^(8) m^(3);the system's economic benefit reduces by 16.49 bil⁃lion yuan,and the amount of pollutant emissions significantly increases by 22.05%.A high credibility level corresponds to a low risk of insufficient water supply and excessive pollutant emissions,leading to low economic benefits and pollut⁃ant emissions.In contrast,a low credibility level results in high economic benefits and pollutant emissions,but the sys⁃tem risk increases.Credibility level thus can be considered as an evaluation index to evaluate the reliability level of the final decisions.Compared with the traditional methods of water resources allocation,the developed model can truly simu⁃late the dynamic interactions of multi-level objectives and multi decision makers.