Research Of Objective Optimization Algorithm And Applied In Estimate Of Typhoon Disaster

With the rapid development of economy and society,a large number of multi-objective optimization problems appear in practical engineering which are unable to solve by the current level of mathematics;the dynamic multi-objective optimization which is latest proposed is more difficult in optimization field and unable to optimize primely by using new algoritnms.Therefore, it is one of the key problems how to find a new strategy to improve the accuracy in optimization theory according to the essential characteristics and optimization difficulties of dynamic problems.The typhoon disaster risk assessment is one of the important topic of research in the field of meteorological disasters.Because the disaster is a complex natural phenomena,therefore we can’t assess the disaster quantificationally.Based on the particularity and contingency,how to use effective methods to assess disaster and provide the accurate basis for decision-making, it has become the focus of attention in the meteorological field;in additon,due to the lack of accurate data,we can not evaluate its performance even if the evaluation method is proposed for better performance; finally,event if the evaluation method is effective, how to put it into practice in disaster prevention and reduction, and how to rapidly release products of the real-time meteorological evaluation, they are also important parts of meteorological service.According to the research questions put forward above, we research the following three areas and obtain the corresponding research results:1. Based on the research of dynamic constraints multi-objective optimization, we propose a hybrid crossover dynamic constrained multi-objective evolutionary algorithm based on selfadaptive start strategy. Firstly, we analyze the optimizaton principle of dynamic optimization of existing algorithms and explore the problems which can’t converge fastly. Secondly, according to the dynamic optimization results distribution, we find the internal relevance in the distributed Pareto optimal frontier of continuous problems. Combined with the mainstream optimization strategy, we improve the effiective algorithm and verify it by simulation experiments. The new algorithm provides a new way for dynamic multi-objective optimization and promotes the development of dynamic multi-objective optimization technology.2. Combined with the research of projection pursuit, we propose a projective pursuit evaluation model based on optimization theory. Firstly, through the analysis of the projection pursuit, we find that the index selection and projection direction optimization are its core problems. Secondly, according to the problems, we improve the proposed algorithm to solve the actual problem in projection pursuit. Finally, we propose the projective pursuit evaluation model based on optimization theory, and use it to evaluate the sensitivity, vulnerability and prevention of typhoon disaster successfully in Shanghai Jiading.3. Based on the evaluation model, we design and development a disaster risk assessment system of typhoon depending on geographical information technology. We introduce the demand analysis, technical route, function modele and database design, and analysis the application effect of projective pursuit evaluation model. This work promoted the development of typhoon disaster assessment.The above work not only enriches the theory of multi-objective optimization, but also further extends its application field. Optimization theory combined with the typhoon disaster risk assessment, it provides a new method for hazard assessment and promotes the application of optimization in the meteorological research.