Study On Urban Water Demand Forecast And Multiple Water Sources Optimization Allocation Under Uncertainty

Water is the important resource for urban survival and development.Urban water supply system is the most fundamental way to ensure the water demand of these users.In reality,there exist uncertainties in both demand and supply subsystems of water supply system,which pose a threat to production and living of a city and cause losses by affecting the normal operation of the system.Water demand prediction can help ensure the balance between water supply and water consumption by adjusting the supply of water,so as to avoid waste of resources and future water crisis.Furthermore,making reasonable planning for limited water resources is important for guiding urban planning and construction of future water supply.At present,there are some deficiencies in theory and practice of water resources management,which still need further study.Based on uncertain statistics and uncertain set,this dissertation builds a regression model,and the regression model is utilized to water demand prediction.In view of the characteristics of water supply system,an uncertain variable-based multi-objective chance-constrained programming model is established to obtain the optimal scheme of water supply.This model can provide decision makers with a trade-off between system cost-effectiveness and default risk under uncertainty.Then,the built models are introduced to water resources management in Handan City.The main contents are as follows.Firstly,a regression model is built based on uncertain set,and the linear regression models whose coefficients are symmetrical triangular,symmetrical trapezoidal and normal uncertain sets are discussed respectively.Then a solution for parametric estimation problem is introduced via linear programming and nonlinear programming technique.Next,by analyzing the factors which have influences on water demand,the regression model between water demand and its influencing factors is established.And then we make demand predictions of domestic,industrial,agricultural and ecological water in different planning level years.Secondly,taking the impacts of uncertain factors into consideration,the decision may not meet the constraints when emergency occurs,which may lead to conservative optimization plan.And combined with the features of multiple water sources,water-users and objectives for water supply system,the uncertain chance-constrained programming model is established.The parameters with uncertainties are denoted by uncertain variables,and uncertain measure is used to measure the possibility that an event will occur.Moreover,this model sets the economic,social and ecological environmental benefits as the objective functions during water resources allocation.It attempts to achieve the maximization of overall benefits in the best possible to meet the supply and demand balance constraints.The final optimization strategy is calculated by model transformation.After that,based on inverse uncertainty distribution,the chance-constrained programming model is transformed to an equivalent determined model,and the optimal scheduling scheme is worked out by Lingo software.Thirdly,taking Handan City for an example,the established uncertain regression model is used to predict the water demand in 2020 of different water users.Beyond that,multi-objective chance-constrained programming model is applied to water supply in urban area of Handan City on the basis of water prediction results.Thus the desired water-allocation scheme is obtained.The research will enrich the water forecast methodology and broaden the application of uncertain statistics and uncertain programming.The results will be conducive to water demand forecast and the process during water supply planning and construction.