Improved Genetic Algorithm Based On Elitist Strategy And Its Application In Parameter Optimization Of Crop Models

Genetic algorithm (GA) is an adaptive random search method based on the theory of biology evolutionism and genetics, having merits of briefness, strong robust. However, GA still has many issues need to be improved such as poor local searching ability, premature convergence and slow convergence speed. Elitist strategy is to use the information of the elite individual in population to improve the search performance, the existing improved algorithm has effectively solved the problem of slow convergence speed, but there is still a problem of poor local searching ability in anaphase and premature convergence when solving ill test function.Crop growth model is a mathematical model of theoretical generalization and quantitative analysis of the relationship between crop physiological processes and the environmental, technical based on the inherent discipline of crop growth and development. The variety parameters of the model need to be re-optimized in different environmental conditions or for different varieties, which belongs to the multivariate and complex nonlinear optimization problem. Genetic algorithm as a heuristic global optimization method is very suitable for solving the optimization problem, but the simple genetic algorithm has the problem of premature convergence in the application process. Improved genetic algorithm would obtain more accurate model parameters. However, there are many problems such as a complex structure, many parameters and low solution efficiency.Aiming at these problems, a newly improved genetic algorithm, named individual advantage genetic algorithm (IAGA), is proposed and applied to the field of the parameters evaluation of the rice development stages model in this thesis, providing a new method to solve the crop model parameter optimization problem. Then we develop crop development stages model parameter optimization system based on individual advantages genetic algorithm (CDSMPOS-IAGA). The main contributions of this paper are as follows:(1) Propose individual advantage genetic algorithm to enhance the search ability of convergence speed and convergence rate.Based on the elitist strategy, an elitist sub-population was introduced in evolution population. While maintaining global search convergence, it enhances the ability of local search in the optimal solution area. Firstly, we introduced the semi-particle swarm mutation operator (SPSMO) into the genetic algorithm to improve the speed of reaching the neighborhood of the optimal solution in prophase. Then, the individual advantage operator (IAO) was proposed to improve the diversity of advantageous individual. Theoretical analysis proves that the algorithm converges to the global optimization solution. Tests on14benchmark functions show that the algorithm has fast convergence rate and better search capability. Compared with existing similar algorithms, IAGA has balanced the contradiction between convergence speed and global convergence and further improved the speed of convergence and precision.(2) Propose a new method of parameter optimization for the rice growth stage model based on IAGA and realize cultivar parameter automatic estimation.While determining the application object is RiceGrow and ORYZA2000rice development stage model, we design the coupling frame between IAGA algorithm and rice growth stage model, the objective function and the fitness function. Experiments on parameter optimization of Shanyou63, Yanjing2, Lusizhan, Xuehuanian and Liangyoupeijiu show that:1)The experimental verification results which cover RMSE<3.05d, R2>0.9877, indicated that the accuracy of the model parameters obtained by IAGA was pretty high.2) The amount of data used for the parameters estimation had little effect on the results. The maximum NRMSE of the fitting results increased from2.58%to3.08%when we changed the amount of data used for the parameters estimation from three years to six years.3)More accurate model parameters were obtained when we select the data of every other year, including the maximum and minimum value of the whole growth period.4) Compared with the shuffled complex evolution algorithm, genetic simulated annealing algorithm and standard particle swarm algorithm, IAGA could obtain more accurate model parameters.(3) Develop the crop development stage model parameter optimization system based on IAGA (CDSMPOS-IAGA) and provide automatic model parameter adjustment software tools.Based on the method of IAGA-based crop development stage model parameter optimization, we developed CDSMPOS-IAGA system using component-based development method. Application case on model parameter optimization for RiceGrow and ORYZA2000shows that CDSMPOS-IAGA can quickly and accurately estimate the model parameters of rice development stage.