Research On Hybrid Algorithm Of Genetic And Simulated Annealing For Optimal Design Of Drainpipe

Network of drainpipe is an important constituent in city municipal constructive, its ratio among investments occupy about 70% in whole drainage system invested. The investment cost as high as two hundred million RMB upwards of the one medium scale urban network of drains. So, it has gotten all-important economic significance of proceeding optimum design toward of drainpipe and depressing it engineering investment cost.For a city, the optimal design of drainpipe contains two aspects, one is the route optimization of drainpipe, and another is parameter optimization of drainpipe. According to the hydraulic calculation formula of drainpipe, the parameter contains quantity of flow, diameter, laying depth, velocity, depth ratio and hydraulic gradient. Once designers select the right diameter and laying depth, other parameters could define by hydraulic calculation formula. So to a certain pipe, when quantity of flow has been defined, there are many combinations of diameter and hydraulic gradient. And this thesis chooses to optimize the combination parameter of diameter and hydraulic gradient.In order to search the most optimum design method of drainpipe, a large number of scholars at home and abroad had done lots of research. With the development of systems analysis methods and computer technology, the genetic algorithm has been a new global optimization search algorithm. Because it simply used for many optimization problems and could be easier to give a satisfactory solution for parallel processing such as the characteristics of no special requirements, the genetic algorithm has shown unique charm in the field of engineering optimization. But the algorithm itself has defect of"easy to fall into local minimum", affecting the quality of optimization. To overcome the drawbacks of genetic algorithms to improve the optimization of the quality and efficiency of design, this article will combine the simulated annealing algorithm which has more robust and less sensitive to the initial value, and use the hybrid algorithm of genetic and simulated annealing to optimize the laid drainpipe system design.Based on the hybrid algorithm of genetic and simulated annealing and the calculation methods of drainpipe network, the article did the following works for optimizing drainpipe providing laid drainpipe network: ①By analyzing the theory of genetic algorithm and simulated annealing algorithm, discussing the advantages and disadvantages of genetic algorithm and simulated annealing algorithm, proposed a new hybrid algorithm that mixed the genetic algorithm with simulated annealing algorithm.②By studying the optimal design of drainpipe system, proposed that optimized the parameter of drainpipe by hybrid algorithm of genetic algorithm and simulated annealing algorithm. That was the choice of diameter and depth of combinatorial optimization to achieve the optimization of relevant parameters of drainpipe.③By using the hybrid genetic algorithm and simulated annealing algorithm to design the drainpipe of a certain subdistrict in Chongqing, constructed an optimal design model of drainpipe by using the drainpipe cost as the objective function, and combined genetic algorithm and simulated annealing algorithm for the diameter and depth for combinatorial optimization. Compared with traditional algorithm for drainpipe's parameters optimization, this thesis got following conclusions: satisfied the general constraints in drainpipe system design and took advantage of the new hybrid algorithm which got the optimal combination of diameter and depth, the whole drainpipe system had reduced investment by 10%.It can get more appropriate and reasonable pipe diameter set by choosing little design velocity and larger design sufficient fullness and limiting different pipe diameter under the different velocity, and it also can get optimization results of drainpipe network with high efficiency by using hybrid algorithm of genetic and simulated annealing, and establishes a path for using hybrid algorithm of genetic and simulated annealing in the drainpipe network field.