Abstract:

In order to further improve the accuracy of precipitation forecast,aiming at the problems of the diversity of forecast products and differences of forecast results,we propose an integrated precipitation forecast method based on hybrid algorithms of the particle swarm optimization (PSO) and the genetic algorithm (GA).The method combines the advantages of the two,and the integrated processing of multi-model precipitation data is realized by screening and processing various real precipitation data provided by the Tianjin Municipal Meteorological Observatory.We compare our method with traditional integration methods,single PSO and single GA.Experimental results show that the proposed method is better than traditional methods,single PSO,single GA and any model member.

Key words: hybrid algorithm, multi-model integration, genetic algorithm, PSO, precipitation forecast