The Optimal Design Of Drainage Pipe Network Design Via Improved Differential Evolution Algorithm

Drainage pipe network is considered as one of the most important entities in the urban infrastructure constructions, its construction investment accounts for about 70% of the total investment. It needs great investment for construction, maintenance and management. The optimization design of drainage pipe network is especially important for decreasing the investment and cost. The drainage pipe network information management system can facilitate the management and maintenance. So it has attached much attention by departments of engineering construction, management and operations.Large-scale drainage pipe network consists of lots of complex, variable information. It is difficult to solve this large-scale optimization problem for the traditional optimization algorithm. The drainage pipe network model and problems are analyzed that exist during the process of optimal design based on the self-adaptive differential evolution algorithm and generalized opposition-based learning method. A new improved differential evolution algorithm is proposed. The standard test functions and three actual drainage pipe networks are used to verify its effectiveness. At last, the drainage pipe network information management system is developed taking advantages of Visual Studio 2010, ArcGIS, MATLAB tools, providing data and theoretical support for urban planning and design personnel. Specific work is conducted as follows:1. The traditional differential evolution algorithm is improved to solve the complicated problem with large-scale pretreatment during the sewage pipe network design. The generalized opposition-based learning method is introduced to refine the initial solution without increasing the number of population. During crossover and mutation, the self-adaptive strategies and associated parameters are adopted to reduce the dependence of the algorithm for initial condition set. The experimental results show that the algorithm has higher precision and faster convergence speed.2. Through analysis of the design parameters, hydraulic calculation of sewage pipeline, a novel optimization model of the drainage pipe network is designed and the optimal design is made with the improved algorithm, then the laying method of three pipe networks is obtained. Based on analysis of the description of the rainfall intensity and calculation process, storm sewer parameters are made clear and the design and calculation of storm sewer model are built. Concluding from the cost of drainage pipe network and flood loss formula, the objective function and constraint conditions satisfying in the process of optimization are determined. Three kinds of drainage pipe network model including low-dimensional single-target model, high-dimensional single-target model and multi-objective model are introduced to make the optimal design using the improved differential evolution algorithm. The results show that this algorithm can solve the complex parameter settings and high solving difficulty that existing in the optimization design of drainage network.3. Aiming at solving the problem that decentralized management mode of the existing management system cannot describe a good characterization of pipeline network and make intelligent decision, the drainage pipe network information management system is developed. Several modules are integrated as a desktop application, which can be easyly installed and uninstalled. Combining with the advantages of C#.net and MATLAB, GUI interface is developed to complete the design of the intelligent decision module of the system and make the optimization results information display visually. ArcGIS Engine controls and database are used to complete the information management module design of system and show drainage pipe network information configuration directly.