Loan Portfolio Decision Model Fuzzy Entropy

In the classic theory of Markowitz variance, the yield is assumed to be normal distribution.Yield variance is used to measure the investment risk, but this assumption is often differentfrom the reality. In China, the loan yields of commercial bank have fuzzy uncertainty, sofuzzy entropy and information entropy is used to measure the degree of uncertainty while theyields are assumed as fuzzy variable.Fuzzy entropy is used to describe the uncertainty of the information. While yield isconsidered as random variable with normal distribution, fuzzy entropy and variance areequivalent in terms of risk measure. But while considering the yield as fuzzy variables and donot obey the normal distribution, fuzzy entropy is more reasonable than the variance, due tothe influence of the different income risk level under different proceeds. Fuzzy entropyimproves the complexities in calculating and the dependence of probability distribution tovariance, and is more practical in loan risk measurement.Due to a complex correlation between different projects, decisions making will focus onone or a few high-yield loan projects if we ignore the solution of correlation building loanportfolio model. Therefore, in order to remedy the small flaw, the constraint condition ofspread risk is added into the model. It could compensate the neglect of the correlationbetween loan projects and the low number of loan portfolio.The hybrid intelligent algorithm with the combination of fuzzy simulation and geneticalgorithm was used to solve problems. This algorithm breaks the routine, obtains the optimalsolution. This paper also verifies the feasibility of the algorithm.