欢迎访问中国科学院大学学报,今天是

中国科学院大学学报

• • 上一篇    下一篇

不可分资源的公平和高效分配问题研究*

杨文国, 刘哲, 高随祥   

  1. 中国科学院大学数学科学学院,北京 100190
  • 收稿日期:2025-02-19 修回日期:2025-04-29 发布日期:2025-05-26
  • 通讯作者: E-mail: yangwg@ucas.ac.cn
  • 基金资助:
    *国家自然科学基金项目(12071459)资助

Survey on fair and efficient allocations of indivisible resources

YANG Wenguo, LIU Zhe, GAO Suixiang   

  1. School of Mathematical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2025-02-19 Revised:2025-04-29 Published:2025-05-26

摘要: 资源分配问题是一类基本的组合优化问题,在经济学、计算机科学等领域有着广泛应用;在这些场景中,诸如商品和任务等资源必须在各局中人之间进行分配。本文关注在寻找分配方案时资源不可分割性带来的挑战,分别考虑无嫉妒、成比例、公平的、最大最小收益、帕累托最优和它们的松弛形式等公平和效率准则,全面梳理了相关文献中满足各种公平标准的存在性结果、算法及其近似方法的最新进展,进而讨论了同时实现公平性和追求效率的算法。本文还研究了所列算法的计算复杂度和找到公平且高效的分配的可能性,总结了不可分资源分配问题算法设计技术和研究中的开放性问题。

关键词: 公平性, 效率, 资源分配

Abstract: Resource allocation problem is a basic class of combinatorial optimization problem and has found widespread application across various fields, such as economics and computer science where resources like goods and chores must be allocated among agents. In our survey, we focus on the challenges caused by indivisible resources. We consider fairness and efficiency criteria, including envy-freeness, proportionality, equitability, maximin share, Pareto optimality, and their relaxations. And we survey the recent progress of existential results, algorithms, and approximations that satisfy various fairness criteria in related literature. Additionally, we discuss algorithms that achieve both fairness and efficiency, such as envy free up to one item and Pareto optimality. We also study the computational complexity of these algorithms, and the likelihood of finding fair and efficient allocations. And we summarize the common algorithm design techniques, and open questions for future research.

Key words: fairness, efficiency, resource allocation

中图分类号: