信贷风险决策问题的一种相互逼近算法

基于最速下降法的基本思想 ,提出了相互逼近算法 ,用以解决信贷风险决策过程中 ,利润曲线和风险曲线寻求公共最优近似解的问题 .该算法表明 ,当利润曲线和风险曲线不存在公共最优近似解时 ,银行追求利润最大化的结果将导致风险上升 ,无法在可接受的风险指数范围内实现其既定的盈利目标 .但当利润曲线和风险曲线存在公共最优近似解时 ,银行根据其所掌握的私有信息以及所观测到企业理性的反应 ,作出相应的决策 .公共最优近似解的存在 ,说明了银行是在风险可接受的前提下按最优性原则给企业发放贷款 .

A Mutual Approximation Algorithm on Credit Risk Decision-making Problem

The paper puts forward a mutual approximation algorithm based on steepest descend method. It is used to solve the problem of looking for common optimal approximating solution on the function of profit and the function of risk in credit risk decision-making. The algorithm shows that, when there exists no any common optimal approximating solution between the function of profit and the function of risk, to maximize profit will increase risk, and hence the bank is unable to achieve their profit goal with accepted risk criterion. However when there exists the common optimal approximating solutions between the function of profit and the function of risk, the bank can make a decision according to the known private information and the rational response of the entrepreneur. The existence of the common optimal approximating solution shows that the bank can provide a loan to the entrepreneur according to the optimal principle that the risk can be accepted as a criterion.