信用资产组合优化的“条件在险值-补偿”型随机规划模型

Stochastic Programming Model with CVaR-Recourse Criterion For Credit Portfolio Optimization

本文对于信用资产组合的优化问题给出了一个稳健的模型,所建模型涉及了条件在险值(CVaR)风险度量以及具有补偿限制的随机线性规划框架,其思想是在CVaR与信用资产组合的重构费用之间进行权衡,并降低解对于随机参数的实现的敏感性．为求解相应的非线性规划,本文将基本模型转化为一系列的线性规划的求解问题．

This paper proposes a robust model for credit portfolio optimization. The model is based on the Conditional Value-at-Risk （CVaR） risk measure and Stochastic Linear Programming with restricted recourse. The idea is to consider the trade-off between CVaR and the restructure cost of the credit portfolio , and reduce the sensitivity of the solution to the realization of the random parameters. In order to find a satisfactory solution to the nonlinear programming, this paper transform the foundation model into the solution problem of the sequential linear programming.