基于模糊时间序列预测的二目标区间数折衷规划

Binary-Objective Interval Number Compromise Programming Based on Forecasting of the Fuzzy Time-Series

用效用函数代替传统投资规划目标函数中的收益率,采用S型隶属函数描述反映投资收益带给投资者的心理满意度水平,利用模糊时间序列的模糊逻辑关系预测满意度水平,建立目标函数为区间收益和区间风险的二目标区间数规划,在引入优化系数的基础上,并采用折衷规划的方法求解。对上证180成分指数进行实证检验表明:所提出的模型对待定参数依赖性小,具有很强的鲁棒性;在参数在标量范围内,所提出的模型具有很好的投资决策效果。

Using utility function as the objective function instead of the yield rate in the traditional investment programming,adopting the S membership function to describe the level of degree of the satisfaction which the investment yield brings for investors,and using the fuzzy logic relationship of the fuzzy time-series to forecast the level of degree of the satisfaction,we establish the binary-objective interval number programming which the goal function is the interval yield and the interval risk,and resolve the programming using the compromise programming method based on the consideration of the optimization coefficient.Through analysis of the empirical of the SSE 180,the result shows that the new model is less dependent of the system parameter and has very strong robustness;the new model has good effect when the system parameter is in the scalar scope.