摘要
大系统的递阶优化方法是解决许多实际问题的重要方法之一,它已广泛成功地应用于解决如工业过程控制、经济规划、城市供水系统的调度、水资源系统的管理及城市交通控制等许多领域的重大问题.从数学规划的观点看,递阶优化方法主要可分为两类,一类是基于数学规划中的拉格朗日函数法.
In this paper,a new method for constructing the augmented Lagrangian function is pro-posed on the basis of the principle of multiplier methods.Under this construction,not onlyis a nonconvex problem convexified,but also the separability necessary for application of thehierarchical algorithm is preserved.The equivalency between the convexified problem andthe former problem is proven.Furthermore,with the aid of primal-dual methods,a hierar-chical optimization algorithm is proposed,and the rate of its convergence is analysed.Themethod has been successfully used to solve the optimal distribution problem of an urban watersupply system,and satisfactory results are obtained.
出处
《系统科学与数学》
CSCD
北大核心
1993年第1期10-19,共10页
Journal of Systems Science and Mathematical Sciences
