A Big Data based Approach to Chance Constrained Problems Using Weighted Stratified Sampling and Differential Evolution

Abstract

This paper proposes a new usage of big data for solving optimization problems under uncertainties. Chance Constrained Problem (CCP) is formulated by using big data. In order to evaluate probabilistic constraints in CCP from big data, a new stratified sampling technique called Weighted Stratified Sampling (WSS) is proposed. Then a group-based adaptive differential evolution called JADE2G is combined with WSS and applied to CCP. The proposed optimization method is demonstrated through two types of CCPs, namely Joint CCP (JCCP) and Separate CCP (SCCP).