Research On Evolutionary Strategy Algorithm And Its Application In Meteorological Optimization Problem

Most of engineering practice and scientific research programs can conversion into optimization problems.The optimization problems have applied in range of areas widely. Meteorological inverse problem and data assimilation are one kind of optimization problems.One side,researching on meteorological inverse problem is benefit for improving accuracy of forecast model.An other side, researching on data assimilation is benefit for accuracy of initial conditions.Both side is benefit for accuracy of numerical weather prediction.So,it is of Science sense and worth.At the same time,it is useful to preparieding before disaster.By researching the evolution strategy, the dissertation improve the ranking process.what is more, this paper also researches meteorological inverse problem and data assimilation with evolution strategy. In a word, the works discussed in the dissertation will be gived as follows:1. About evolution strategy. Using evolution strategy solving the nonlinear system of equations,we find that the investigation may converg to the local optimum. By analyzing the ranking process, a new ranking process is given.2. About the model parameter identification problems.Usually,the parameters is hard to determine.In the dissertation, a new meteorological inverse problems solution based on evolutionary strategy is presented.Two ideal numerical experiments are implemented by using the simple pediction models of one-dimensional diffusion equation and Lorenz-96.The result shows that the algorithm is effective.3. About initial conditions. Data assimilation is just the method for initial conditions. When researching data assimilation,we use evolution strategy to sovle the object function.Ideal experiments are given on Lorenz-63 and Lorenz-96.What is more, this dissertation do the numberical experiments on barotropic primitive equations,the resule shows that evolution strategy is effective.In the dissertation,two questions are discussed. By research nonlinear system of equations, an improved ranking method is proposed. When research meteorological optimization problem, evolution strategy is desigend. The works in this dissertation are very important to the research of the evolutionary strategy and the optimization applications in the real world.