Research On The Optimization Design Of Pile-anchor Supporting For Deep Excavation Based On Improving Particle Swarm Optimization

Deep excavation engineering is a complex and comprehensive project, and always designed as a provisional establishment. It is required to harmonize the relationship between the cost and reliability of engineering. Therefore, to optimize the design of retaining structure for deep excavation is particularly important.The optimization design of supporting structure usually comes down to a large number of variables and combination variables, including continuous variables and discrete variables. In addition, the relationship between optimization target and design variables is relatively complex. It is difficult to construct the explicit expression of the rational and making it hard to solve the complex deep excavation optimization problems by using traditional optimization methods. Thus, to establish the mathematical model of optimal design and search for a rational and feasible modern optimization algorithm is the key. Based on the pile anchor supporting structure as an example, the design variables which have great influence on the optimization results are selected, and analyzing the main constraint condition of the system and deal with it by using the penalty function method. Finally, the optimization model is established.This thesis contrasts and analyses the different calculation-method, and improves the Particle Swarm Optimization(PSO) in order to be suitable for the optimization design of supporting structure of deep excavation. The strategies are mainly in the following two aspects:1. Aimed at Particle Swarm Optimization Algorithm is prone to premature convergence phenomenon, this paper introduces genetic algorithm selection, crossover, and mutation operator that poor adaptive value judging part particles by a close relative, shall be carried out in accordance with the dynamic mutation probability to initialize the variable. So it can be improved the diversity of the particle population;2. New searching way is proposed to improve the local search ability in the late stage of the Particle Swarm Optimization Algorithm, and the convergence performance of the algorithm is guaranteed, which is made up of the Hooke-Jeeves method and Particle Swarm Optimization when it partially calculates in the later time.The improving Particle Swarm Optimization Algorithm for deep excavation, carrying on MATLAB language, is used in the optimization design for the project, “the retaining structure for deep excavation of Nanhai district of Foshan city”. The design results are comparative analyzed between the optimized scheme and the original design scheme as well as the PSO and Breed PSO optimization results. It is proved that the improved particle swarm algorithm has a good effect on the late search ability, the calculation precision and the convergence stability, and is suitable for the optimization design of the supporting structure of deep excavation.