Emergency Decision-making Based On The Bayesian Method

With the rapid development of economy and society, the emergencies occur frequently in various areas around the world. These emergencies not only cause incalculable economic losses but also claim the lives of hundreds of millions, leading to increasing global impact and harm. When an emergency occurs, policymakers must make an efficient and reasonable emergency decision to minimize the harm. Therefore Emergency decision-making is the core of emergency management.When an emergency occurs, policymakers choose and implement a satisfactory plan and correct mistakes in the decision-making process. This dynamic process is called emergency decision-making. In many cases, emergency decision-making needs to select and execute an optimal from multiple alternatives. Bayesian decision method exactly can solve the problem of optimal selection, which has been widely used in many fields.Bayesian decision method is a common method of optimal selection. The basic principle of Bayesian decision method is: under incomplete information, applying Bayesian formula to amend prior probability, and then calculate utility expectations of various programs based on the amended probability. The program with maximum utility expectation will be optimal. Traditional Bayesian method has some shortcomings in determining of utility functions and state variables. For example the complexity of emergencies’ structures is not taken into account and lack of secondary events during decision-making. Introducing secondary events, interval numbers and probability weights to Bayesian decision method can solve these problems. Secondary events refer to the related events the native events have triggered the development process. Actual emergency-decision should consider the disposal of the native events and secondary events. The Execution results of a program may vary in different environments of events, but utility factors of traditional Bayesian methods typically adopt exact value. So adopting interval numbers is more reasonable. It also needs to distinguish the importance of various decision goals in making decisions, which can be described by the probability weights. It is more suitable for multi-objective decision-making situations. Moreover it enriches original application background and builds an emergency decision-making model based on improved Bayesian decision method. At last, in order to verify the validity and rationality of the new decision model, the paper applies this method to earthquake disaster emergency decision problem.