Co-evolutionary Algorithm Research And Its Application In Parameter Optimization Of Torrential Rain Intensity Formula

There are a lot of complex parameter optimization problems in meteorology, especially the problems of models and formulas, in which often exist multiple conflict objectives. Multi-objective evolutionary algorithms (MOEAs), which is proposed as a group search algorithm, is very suitable for solving this kind of optimization problems. Therefore, with regard to the complex meteorological parameter optimization problems, this paper advanced two kinds of multi-objective evolutionary algorithms which are based on co-evolution and to the parameter optimization problem of torrential rain intensity formula, for example, applied them to the meteorological parameter optimization problem, and these works make a useful and necessary exploration for solving the more complex meteorological parameter optimization problems in the future.This paper can be devided into the following three aspects in detail:1. This paper conducted an exploratory development relative to the goal and hot spot of MOEAs research and the existing problems of the classic multi-objective algorithm NSGA-Ⅱ. First, this paper applied the classic multi-objective evolutionary algorithm NSGA-Ⅱ to solve the standard test functions, and found the problem that there are a blindness in the post-search and a lack of diversity in population.Second, based on the current research focus of MOEAs, this paper proposed the solutions relative to these problems.2. This paper carried out a detailed introduction of the two hybrid multi-objective algorithms which is proposed here, and with regard to the research goal and hot spot of MOEAs, using multiple standard test functions, this paper verified the two hybrid algorithms through conparing them with the classic multi-objective algorithm NSGA-II and a relatively new algorithm OMOPSO in multiple indexes., the experimental results show that the effectiveness and the superiority of the two kinds of evolutionary algorithms provide a new concept of the multi-objective evolutionary computation and promote the research of the multi-objective evolutionary optimization.3. With regard to the meteorological parameter optimization problems, this paper applied the two hybrid multi-objective evolutionary algorithms to the parameter optimization of meteorology in case of the torrential rain intensity formula. The experimental results indicate that the two multi-objective evolutionary algorithms proposed here are feasible and most of the obtained results are better than the results given in the literature, and then show that these two algorithms can be taken as a new method for reference to the more complex meteorological parameter optimization in the future.The works proposed above not only enrich the research of MOEAs, but also further expand the application fields. The more important is that this work to the parameter optimization problem of torrential rain intensity formula, for example, provides a method that could be used as a reference of meteorological parameter optimization, provides a paradigm for the combination of meteorology and computer, and helps to promote the more applications of optimization algorithm in meteorological research.