Research On The Algorithm And Application For A Class Of Stochastic Chance-constrained Programming

The solution and application of a class of stochastic chance-constrained programming (CCP) were discussed and researched in this paper. Firstly, some deterministic equivalents of the chance-constrained condition were discussed as the random vector was with normal, uniform and exponential probability distribution, and some theorems were proved. Secondly, when the Monte Carlo (MC) based method was used to solving CCP, the sample size was usually chosen by subjective experiment or other estimation, and no more efficient solutions were employed. A dynamic search method based on MC simulation named retrospective approximation (RA) was presented in this paper, and a numerical example was used to describe the feature. Finally, a hybrid intelligent algorithm including genetic algorithms, random simulation and neural network was discussed to solve CCP, and a numerical example was employed to demonstrate the efficiency of the algorithm.