Credit portfolio optimization: A multi-objective genetic algorithm approach

Abstract

The algorithm for optimization of a credit portfolio has not been fully demonstrated. This paper fills the gap in the literature by presenting a general approach for optimizing a credit portfolio by minimizing the default risk of the entire portfolio. Default risk is measured with quadratic weighting and a matrix containing information about the default intensity of two stocks and the correlation in default between them. The default correlation and the default intensity are represented with a novel bivariate intensity model. A multi-objective genetic algorithm is introduced to optimize a credit portfolio with the purpose of overcoming limitations in the analytical method and improving the efficiency of optimization. The algorithm can be applied to a portfolio's credit risk management, which is particularly crucial for investors and regulars in emerging markets.