基于鲁棒规划方法的农业水资源多目标优化配置模型

在干旱半干旱地区,调整种植结构可以促进农业水资源的高效利用。农业水资源配置需要在多个目标间权衡博弈,对各目标的偏好和赋权直接影响着优化模型的输出和决策方案的制定,但以往研究往往忽略了权重确定过程中因主观等因素的影响而普遍存在的不确定性。针对农业水资源多目标规划中存在的权重不确定性难题,建立了基于鲁棒优化方法的农业水资源多目标优化配置模型方法(MRPWU)。该方法可以把权重中蕴含的复杂不确定性信息纳入建模过程,产生可靠的模型结果;并能提供效益值及风险值均定量化的方案集,便于决策者在权衡效益与风险后确定最优方案。模型以作物种植经济收益和碳吸收量最大化为目标、以水土资源供需平衡等为约束条件,并应用于农业水资源供需矛盾突出的甘肃省民勤县。优化结果表明,随着保护度水平的提高,生态效益上升,经济效益和综合效益下降,系统面临的风险也随之下降。相比于权重为确定参数的模型,MRPWU模型可以在综合效益下降3.7%的同时,较大地提高系统应对权重不确定性以及风险的能力。与2017年的实际情况相比,MRPWU模型可以减少种植面积1.6%、节省灌溉用水3.9%,同时提高生态效益1.6%。

A multi-objective robust programming model for agricultural water management with uncertain weights

It is of great significance to adjust crop planting patterns for efficient allocation of water resources in arid and semi-arid areas. Agricultural water management is a complex problem that needs to deal with the tradeoffs among multiple objectives,and objectives weights will greatly influence the decision-making process. Due to the subjective opinions from different decision makers, weights are uncertain in nature. However, weights were often simplified as fixed values in previous studies, which neglected the uncertain feature of weights. Thus, a multi-objective robust programming model for agricultural water management with uncertain weights (MRPWU) was developed in this study. This proposed method could deal with the uncertainties existing in weights, manipulate the violation-risks of the system, and provide robust solutions for agricultural water management. It could also provide solutions with quantified benefits and risks,helping decision makers to determine the optimal solutions. The objectives were to maximize the economic benefits of crop planting and the ecological benefits of carbon absorption. A set of constraints including water resources availabilities and production demands were considered. Minqin County in northwestern China was selected as the study area, where the contradiction between water supply and demand was very prominent. The results show that, with increased protection level, ecological benefit would increase, while economic and comprehensive benefits,as well as system risks would decrease. Compared with results obtained from a multi-objective programming model with deterministic weights, MRPWU could improve the system''s ability against weight uncertainty and violation risk, with a 3.7% reduction of synergy benefits. Compared with the benchmark practices in the base year of 2017,the optimal solutions of MRPWU model could reduce the total planting area by 1.6%,save the irrigation water by 3.9% and increase ecological benefit by 1.6%.