Stochastic optimization programming for multi-reach river system using GA combined with stochastic simulation

When multi-reach river systems are programmed to meet a certain water quality standard, it's difficult for traditional optimization methods to search the optimal solutions if objective or constraint functions comprise stochastic variables. However, Genetic algorithm (GA) can effectively solve this problem because of its distinct advantages. GA method combined with stochastic simulation technique is used for solving stochastic optimization programming for a factual section of river divided into seven reaches. From the perspective of the effectiveness of the proposed method, the optimized solutions have been stable after evolution of about 150 generations. More importance lies in that this method is easy to handle stochastic optimization problems without special solutions to probabilistic density, thus the complex differential and integral operations can be omitted.