The Application And Research Of Neural Network Algorithms On The Optimal Design Of Water Supply Networks

The water supplying pipes are an important part that forms the water supplying system of the city, And the expenditure covers about 70% of all investments, involving the great costs of energy consumption and operation management. So the designing plan that minimizes the costs of the system is more important significance to improve economical and social benefits.The optimal design of water supply networks has been broadly and deeply studied by a lot of domestic and foreign scholars just because of its important status in water supply engineering. These scholars put forward a few kinds of optimal methods, such as linear programming method, nonlinear programming method, dynamic programming method and simple genetic algorithms (SGA). However, these methods all have somelimitations on practice.According to the current achievement, directing to the existing deficiency of the designing method of the water supplying pipes, this dissertation applies the Neural Network theory to formulate the optimizing Computational Energy Function, researching the several problems of the pipe optimization especially in the way of the classifying optimization of the ringed pipes, also it improves on the optimizing arithmetic in the course of optimization, the result shortens the running time. The author describes the application in the course of optimization and demonstrates it feasible by comparing the achievement with the traditional result.