Research On Loan Portfolio Optimization Model For Commercial Banks And Its Intelligent Algorithm

Nowadays, the commercial bank's loans portfolio optimization decision problems is one of the issues that the banking common concern, whether or not the decision-making loans directly related to the survival of the bank. Therefore, the loan portfolio is selected scientifically and reasonably, which has important research value.It can effectively avoid the risks of commercial banks, and increases their income. In the paper, with actual operation of commercial banks in China, we make the following several loan portfolio optimization models for commercial banks, and give their intelligent optimization algorithms:(1) A decision-making model of the commercial bank's loans portfolio optimization is established by using complex risk weight and the unit risk-income biggest principles and the enterprise's credit graduation situation.The model is essentially a nonlinear 0-1fraction integer programming problem with the upper and lower bounds, the model is difficult to solve.We give two different intelligent optimization algorithms to solve it.One is a hybrid genetic algorithm with greedy transformation; the other is a adaptive particle swarm optimization algorithm. The numerical experiment indicates that the model corresponds to the actual requirements and the two algorithms for solving this model are reasonable.(2) For the accumulate loans problem, we establish two different models: First, a portfolio optimization decision model of the accumulate loans and incremental loans based on the unit risk–income biggest principle is established. It is a nonlinear 0-1 programming problem with the upper and lower bounds, and we solve it by a hybrid genetic algorithm with greedy which is referred in (1); Second, A double objective decision-making model of the commercial bank's loans portfolio optimization is established, its objective functions are income of loans portfolio and Conditional Value-at-Risk (CVaR).We present an adaptive particle swarm optimization algorithm to solve the double objective decision-making model. The numerical experiment indicates that the two models and the given solution methods are effective and it can play a certain role for commercial bank making loans'decision.(3) Considering the problems of the credit risk and loan cycle, we first establish a multi-period dynamic loans portfolio optimization model for banks based on the adjusted credit risk, in which the adjusted credit risk is considered.According to the model's feature, we give a solving method, which consists of the dynamic algorithm based on the Monte Carlo simulation and the multi-period differential evolutionary algorithm. The former solves the various types loans'expected income rate, the latter solves the various loans'the investment optimization proportion in every period. Second, considering the bank's asset-liability management, we establish a multi-period dynamic loans portfolio optimization model for banks consider the asset-liability management. Compared with the previous model, the model simply add more risk-free assets and the corresponding limitations. So he previous algorithm can solve the model. But in order to expand the solution ways for such models, we present a new way– a multi-period adaptive particle swarm optimization algorithm to solve the various loans'the investment optimization proportion in every period. The numerical experiment indicates that the two models for is reasonable, the given algorithms are effective and feasibe.