Simplex-Particle Swarm Hybrid Algorithm For Parameter Estimation Of Water Quality And Quantity Model

Water quality and quantity parameters are the important data to assess groundwater resources and the important basis to detect water quality changes, the accuracy of the parameters for reasonable exploitation water resources and prediction water pollution have some significance. This thesis studied two kinds of environmental impact assessment water quality model, established optimization algorithms based on existing hydrogeological exploration data, to determine the two models hydrological parameters.Simplex-particle swarm algorithm(SM-PSO) which was combined by particle swarm optimization algorithm and simplex method have been applied to determine the parameters of first type leakage system and the parameters of two-dimension water quality model, aiming at solve the nonlinear and complicated problem of selecting best parameters in groundwater seepage model and water quality model. Basing on the results of the numerical experiment,the advantages of simplex-particle swarm algorithm are analyzed and discussed. The main work of this thesis is as follows:1. Introduced particle swarm optimization algorithm and simplex method principle,analyzed their respective advantages and disadvantages. Combined particle swarm optimization algorithm which has strong global search ability with simplex method which has higher local search ability, the hybrid algorithm structure make up both of their defects.2. The simplex-particle swarm optimization algorithm is applied to analyze the first aquifer parameters of the flow system, through comparing the calculation results with other methods in some references, fitting observation data with actual data, known that hybrid algorithm has good reliability and high efficiency. The size of SM-PSO optimization algorithm population has certain influence on iterations, operation time and converges, the population size should not be selected too small. The estimated parameters of the initial value of the scope has a certain influence on the SM-PSO algorithm, but does not affect the final convergence. Through the analysis of the seepage model, known that the water level drawdown is a decreasing function of the conductivity coefficient and storage coefficient, and the sensitivity of the conductivity coefficient is greater than the sensitivity of the storage coefficient.3. The simplex-particle swarm optimization algorithm is applied to solve the twodimensional water quality parameters, by comparing different calculation results and analysis the iterative curve, known as hybrid algorithm can converge to a more accurate solution in a short time to an more exact solution. Discuss how to value the reflection coefficient,expansion coefficient, contraction coefficient and acceleration factor of SM-PSO algorithm, it is concluded that the appropriate values of reflection coefficient, contraction coefficient and expansion coefficient is 1, 0.5, 1.78, accelerating factor value is 1.72, then the accuracy and performance of the hybrid algorithm are high. By comparing the convergence and time performance index of the hybrid algorithm with particle swarm optimization algorithm, shows that the hybrid algorithm make up the PSO insufficient in solving the parameters of river water quality, improve the performance and efficiency of the algorithm.