中文摘要：

提出基于上下限的二叉树情景生成法、基于矩匹配的二叉树情景生成法和基于聚类分析的二叉树情景生成法,提出改进的情景分层法。进一步以我国经济环境为依托,考虑未来的不确定性,对多阶段金融投资决策中的情景生成问题进行研究。将3种情景生成方法预测值同实际值进行比较,得出基于上下限的二叉树情景生成法更优的结论。

英文摘要：

Three methods of binary tree scenario generation are put forward. These methods are based on upper and lower limits, moment match and cluster analysis respectively so that the binary tree scenarios for rate of return are constructed and stratified random sampling is improved. In the empirical research, considering the future uncertainty, scenario generation for multistage investment decision is studied in the background of China＇s economic environment. The results are compared with the true value and we come to the conclusion that binary tree scenario generation based on upper and lower limits is better than the other two methods.