Application of genetic algorithm in decision-making optimization of underground gas pipeline risk model

This paper denotes the gas pipeline´s risk degree by expectation wealth loss. We propose to classify the events that result in pipelines accidents according to independency principle and partition this kind of events into exclusive small events. The aim of risk evaluation is to carry through risk control. Aim to decision indexes´ diversification, this paper in the first instance transforms the risk control problem to a decision optimization problem. Considering the complex of this problem, we adopt genetic algorithm to deal with it. We also make many experiments by partial pipelines. The result indicates that this algorithm can find the solution in little time. So we can debase the expectation loss by a little loss.