A Smooth Approximating Method for Solving Single Stage Stochastic Programming

Abstract:

      A smooth approximating method for solving single stage stochastic programs is proposed, and the convergence concerning this method is simultaneously given. The original problem is changed into a determinate nonlinear programs with many constraints by means of some discretizing random variable way, and then the determinate problem is transformed into a nonlinear programs with only simple constraint making use of the maximum entropy principle. Thus difficulties of rapidly growing of the number of constraints due to increasing of the discretizing random variable precision will be overcomed.