Scholarlink

<p>The interbank lending business connects banks into a network, providing a channel for risk contagion. To control the spread of this risk, the government needs to provide timely bailout. The current bailout algorithms face several challenges: (1) most research on bailout strategies only considers static environments after risk convergence; (2) the large number of banks in real-world banking systems results in an enormous decision space. So, to address the first issue, this paper transforms the dynamic bailout problem into a static multi-objective optimization model and develops a dynamic government bailout framework. To efficiently handle high-dimensional decision variables, an improved genetic algorithm, the Large-Scale Non-dominated Sorting Genetic Algorithm, is proposed. This method analyzes the relationships between decision variables and objectives, groups related variables, and decomposes the large decision space into smaller subspaces. It also introduces two local search strategies to enhance exploration and improve solution quality. Experiments on both benchmark tests and real-world cases confirm that the proposed method outperforms existing algorithms in terms of solution quality and computational efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A multi-objective optimization algorithm for optimal bailout allocation in large-scale banking systems under risk contagion

  • 期刊论文

  • 发表日期：

    2025年7月16日

查看全文
  • Yijun Chen,
  • Huanlan Yan,
  • Meiqin Pan,
  • Qingyuan Wang,
  • Zhijun Ding

摘要

The interbank lending business connects banks into a network, providing a channel for risk contagion. To control the spread of this risk, the government needs to provide timely bailout. The current bailout algorithms face several challenges: (1) most research on bailout strategies only considers static environments after risk convergence; (2) the large number of banks in real-world banking systems results in an enormous decision space. So, to address the first issue, this paper transforms the dynamic bailout problem into a static multi-objective optimization model and develops a dynamic government bailout framework. To efficiently handle high-dimensional decision variables, an improved genetic algorithm, the Large-Scale Non-dominated Sorting Genetic Algorithm, is proposed. This method analyzes the relationships between decision variables and objectives, groups related variables, and decomposes the large decision space into smaller subspaces. It also introduces two local search strategies to enhance exploration and improve solution quality. Experiments on both benchmark tests and real-world cases confirm that the proposed method outperforms existing algorithms in terms of solution quality and computational efficiency.