Optimization is an important branch of mathematics study,it involves various fields.Intelligent optimization algorithm to solve these problems provides a new thought and method.At the same time,it got very good results in these application.In recent years,a new metaheuristic intelligent algorithms-flower pollination algorithm was presented.This paper has carried on the research and analysis.And applied flower pollination algorithm to analyze anisotropic aquifer under the condition of experiment data then determine the anisotropy aquifer parameters,and got a good result.Due to flower pollination algorithm exists some disadvantages,with the use of chaotic sequence to initialize pollens,and join the simplex algorithm.At the same time,improved flower pollination algorithm was used to analyze straight water supply boundary aquifer pumping test data to determine aquifer parameters and soil water characteristic curve VG model parameter optimization problem.In this paper,the research work includes the following aspects:1.Firstly,we introduce the production of the flower pollination algorithm and flower pollination algorithm is introduced in the domestic and foreign research situation.Detailedly introduce the basic principle and its implementation steps.Use some commonly used test function to verify,analyze and summary the advantages and disadvantages of the algorithm.2.Flower pollination algorithm was applied to determine the anisotropy aquifer parameters,by the numerical experiment,the results of flower pollination algorithm,comparing the results of calculation with other algorithms,and results of flower pollination algorithm and test the observed value for curing fitting,show that flower pollination algorithm can be successfully used in determining the anisotropy aquifer parameters.And the flower pollination algorithm's result is verified to be reliable.The algorithm's population size has certain effect in running time and number of iterations.By analyzing the parameters of the initial scope of algorithm,the effect of flower pollination algorithm has good convergence is proved.3.To improve the flower pollination algorithm.Firstly,introduce some common chaotic sequence.Detailedly introduce the basic principle and implementation steps of the simplex algorithm.Aiming at the shortcomings of the flower pollination algorithm,into the chaoticsequence and simplex algorithm to improve,and establish a better improved flower pollination algorithm.4.Improved flower pollination algorithm was applied to analyze straight water supply boundary aquifer pumping test data to determine aquifer parameters.By comparing calculation result of the different algorithm shows that improved flower pollination algorithm than other algorithms has better reliability and higher accuracy.The bigger the population size,the higher precision of algorithm.By the analysis of influence of the ranges of initial values of parameters,the improved flower pollination algorithm's convergence is superior to the basic flower pollination algorithm.At the same time,through analyzing the sensitivity of the parameters in the aquifer,storage coefficient sensitivity is relatively high.Improved flower pollination algorithm was applied to the soil water characteristic curve VG model parameter optimization,its calculation results with the observed value is very close,and curve fitting effect is very good.Further show that improved flower pollination algorithm can apply in calculating soil water characteristic curve VG model parameter and expand the application field of the improved flower pollination algorithm. |