随机决策方法在气象服务中的应用

The Applications of Stochastic Decision-making to Meteorological Service

以案例分析为主要形式,阐述了随机决策方法结合气象服务在企业决策经营中的应用。案例1的分析结果表明,当未来高影响天气出现的概率未知时,其为不确定型气象服务决策问题,可根据服务对象对待风险的偏好,相应采取乐观法、保守法或最小后悔值法来选择合适的服务方案。案例2表明,当未来高影响天气出现的概率已知时,其为确定型气象服务决策问题,可采用决策树模型的最大期望估算进行辅助决策。在确定型决策问题的灵敏度分析中,可将其转化为求临界概率的数学问题,并用图示法进行求解。

The applications of stochastic decision-making to meteorological service are introduced primarily in two cases. The first case study shows that the meaning of stochastic decision-making is uncertain when the probability of high impact weather event is uncertain. In this case, the optimistic, pessimistic, or minimum repentant value （MRV） decision-making can be chosen respectively according to the risk preferences of service. The second case shows that it is a certain meaning of stochastic decision-making when the probability of high impact weather event is known. In this case, the best scheme could be setup up by the maximal expectation value （ME） to provide meteorological service. Sensitivity analysis of stochastic decision-making could be transformed as a mathematic question about calculating the critical probability and figured out directly by using graphical method.