Adaptive Punishment Genetic Algorithm To Optimal Design Of Water

Urban water network is an important part of the warter supply systerm,it accounts for 50~80 percent of the payment of total water supply systerm.With the enlargement of city scale , criterion of users to the quality and quantity of water is stricter than before, so the optimal design of water pipe network poses a key task to the water and wastewater engineers. Based on the achievement of current study of optimal design of water network and GA's rationale and practice method, this paper puts forword a improved GA—self-adapative penalty GA, to carry out water network optimal design.The self-adapative penalty function apply the penalty factor which can adaptivly change according to solution of every generation based on the common penalty function, and can make the penalty function adaptivly change. The self-adapative penalty GA sets up the model on optimal design of water pipe network by means of taking some effective measures on coded system, selection operator, crossover operator, mutation operator and using the self-adaptive penalty function to dispose the boundary constraint. It makes the improved GA can not only save the feasible solution but also use the feasible part of non-feasible solution sufficiently, and the algorithms can get the feasible optimal solution effectively to avoid the partial optimal solution. It uses the node hydraulic pressure method as the subprogram of hydraulic calculation in the GA of water network optimal design, and uses the node pressure to determine the individual sufficiency can get better calculation efficiency.The study shows that the self-adaptive penalty GA have the better ability to search the global optimum solution ,higher computational efficiency and better convergence.It has very important significance to the optimal design of water nerwork and great value of reference in the engineering.