具有模糊收益率的贷款组合优化决策  

Decision-making of loan portfolio optimization with fuzzy return rates

作　　者：严维真[1] 宁玉富[1,2] 郭长友[2] 

机构地区：[1]天津大学系统工程研究所,天津300072 [2]德州学院计算机系,山东德州253023

出　　处：《系统工程学报》2008年第2期168-173,共6页Journal of Systems Engineering

基　　金：中国博士后科学基金资助项目(20060400704)

摘　　要：通过把贷款的收益率刻画为模糊变量,提出了贷款组合的优化决策模型,即均值-方差模型,模型中的贷款收益率可以是任意的模糊变量.对于贷款收益率是特殊的三角模糊变量的情况,给出了模型的清晰等价类,这些等价类模型可以用传统的方法进行求解.对于贷款收益率的隶属函数比较复杂的情况,设计了基于模糊模拟的混合优化算法求解模型.该算法集成了模糊模拟、神经网络、遗传算法和同步扰动随机逼近算法,既具有较强的全局搜索能力,又具有高效的局部搜索能力.经数值仿真,验证了算法的可行性.This paper proposes two loan portfolio optimization decision-making models, i. e. , mean value-variance models, by characterizing loan return rates as fuzzy variables, where the loan return rates can be any fuzzy variables. For the case where the loan return rates are special triangular fuzzy variables, the crisp equivalents of the models are given. The traditional methods can be used to solve the equivalent models. To solve the models where the membership functions of loan return rates are complex, a hybrid optimization algorithm based on fuzzy simulation is designed. The algorithm integrates fuzzy simulation, neural network, genetic algorithm and simultaneous perturbation stochas- tic approximation algorithm, which possesses both the strong global search ability and efficient local search ability. The feasibility of the algorithm is verified by an example.

关 键 词：贷款组合 模糊变量 模糊模拟 遗传算法 同步扰动随机逼近 

分 类 号：F830.9[经济管理—金融学]