Lou W.-P.,Nanjing University of Information Science and Technology |
Chen H.-Y.,Zhejiang Provincial Meteorological Office |
Qiu X.-F.,Nanjing University of Information Science and Technology |
Yang X.-Z.,Shaoxing City Meteorological Bureau |
Zhao H.-J.,Shaoxing City Meteorological Bureau
Journal of Natural Disasters | Year: 2010
Quantitative evaluation of typhoon disaster is a basis for scientific disaster reduction. Following the relationship between typhoon disaster and disaster-causing factor/disaster-forming environment/disaster-affected body, the BP network method was used for building the evaluation models of casualty and economic loss caused by typhoon disaster. In order to overcome the sensitivity to initial weight and local minima problem of the BP network method, the particle swarm optimization(PSO) algorithms with the global stochastic optimization idea was applied to BP network ' s training in dynamic weighing system. The experimental vesult shows the evaluation models for typhoon disaster based on PSO-BP are feasible.