Benders计算机程序库的开发及其在环境规划中的应用

混合整变量线性规划模型在国外许多行业和我国环境规划领域得到广泛应用。Benders分解算法是求解大规模混合整变量线性规划问题的一种先进技术,该技术在国外求解运输问题和网络问题范畴的实际问题得到了成功。我们应用Benders技术研制开发了大型计算机软伴;这是在我国系统工程领域和国际环境规划工作中的首次应用。在程序库开发工作中,引进和采用了当今国际上先进的运筹学方法和计算机技术。程序库在求解实际的环境规划问题中获得了初步的成功。我们应用所开发的程序库和数学模型求解了一个具有635个约束条件,1087个变量(其中0—1整变量为126个)的实际城市综合整冶规划问题。这是至今国内所求解的最大规模的混合整变量线性规划问题,也是国际上用Benders方法求解的最大的环境规划问题。

Development of Benders Computer Program Library and Its Application to Environmental Planning

Mixed Integer Linear Programming (MILP) has been widely applied to the diverse fields in the world and to environmental planning in China. Benders partitioning algorithm is an advanced technique for solving Iarge MILP problems.Applications of Benders technique to real problems classified as transportation or network problems have been successfully achieved. A large computer program library has been developed using Benders technique in this work, which is the first application to the field of the systematic engineering in China and of environmental planning in the world. In development of the program library, the modern operations research methods and computer echniques were introduced and used.The program library made it originally successful to solve real environmental planning problems. A real urban integrated environ-mental planning problem which had 635 constraints and 1087 variables including 126 binary integers was solved by the program library and associated mathematical model. As we know so far,this is the event to have solved the largest MILP problem in China and the largest environmental planning problem in the world by Benders technique.