Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Credit Fairness: Online Fairness In Shared Resource Pools

Seyed Majid Zahedi, Rupert Freeman

TL;DR

This work addresses fair resource sharing over time when agents have time-varying demands and fixed endowments. It shows that static max-min fairness, though PE and SP, can yield large inequities, motivating a credit-based fairness framework. The authors propose LendRecoup, a credit-fair online mechanism that is Pareto efficient and satisfies online strategyproofness, while providing formal guarantees via a credit system CF1–CF5; they prove that credit fairness strengthens SI with PE but cannot coexist with SP under anonymity. Empirical evaluation on trace-driven workloads demonstrates that LendRecoup achieves competitive welfare while substantially improving equity and avoiding harm from sharing, suggesting strong practical potential for shared-resource systems.

Abstract

We consider a setting in which a group of agents share resources that must be allocated among them in each discrete time period. Agents have time-varying demands and derive constant marginal utility from each unit of resource received up to their demand, with zero utility for any additional resources. In this setting, it is known that independently maximizing the minimum utility in each round satisfies sharing incentives (agents weakly prefer participating in the mechanism to not participating), strategyproofness (agents have no incentive to misreport their demands), and Pareto efficiency (Freeman et al. 2018). However, recent work (Vuppalapati et al. 2023) has shown that this max-min mechanism can lead to large disparities in the total resources received by agents, even when they have the same average demand. In this paper, we introduce credit fairness, a strengthening of sharing incentives that ensures agents who lend resources in early rounds are able to recoup them in later rounds. Credit fairness can be achieved in conjunction with either Pareto efficiency or strategyproofness, but not both. We propose a mechanism that is credit fair and Pareto efficient, and we evaluate its performance in a computational resource-sharing setting.

Credit Fairness: Online Fairness In Shared Resource Pools

TL;DR

This work addresses fair resource sharing over time when agents have time-varying demands and fixed endowments. It shows that static max-min fairness, though PE and SP, can yield large inequities, motivating a credit-based fairness framework. The authors propose LendRecoup, a credit-fair online mechanism that is Pareto efficient and satisfies online strategyproofness, while providing formal guarantees via a credit system CF1–CF5; they prove that credit fairness strengthens SI with PE but cannot coexist with SP under anonymity. Empirical evaluation on trace-driven workloads demonstrates that LendRecoup achieves competitive welfare while substantially improving equity and avoiding harm from sharing, suggesting strong practical potential for shared-resource systems.

Abstract

We consider a setting in which a group of agents share resources that must be allocated among them in each discrete time period. Agents have time-varying demands and derive constant marginal utility from each unit of resource received up to their demand, with zero utility for any additional resources. In this setting, it is known that independently maximizing the minimum utility in each round satisfies sharing incentives (agents weakly prefer participating in the mechanism to not participating), strategyproofness (agents have no incentive to misreport their demands), and Pareto efficiency (Freeman et al. 2018). However, recent work (Vuppalapati et al. 2023) has shown that this max-min mechanism can lead to large disparities in the total resources received by agents, even when they have the same average demand. In this paper, we introduce credit fairness, a strengthening of sharing incentives that ensures agents who lend resources in early rounds are able to recoup them in later rounds. Credit fairness can be achieved in conjunction with either Pareto efficiency or strategyproofness, but not both. We propose a mechanism that is credit fair and Pareto efficient, and we evaluate its performance in a computational resource-sharing setting.
Paper Structure (26 sections, 12 theorems, 14 equations, 2 tables, 1 algorithm)

This paper contains 26 sections, 12 theorems, 14 equations, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Properties of Mechanisms.
  5. Credit Fairness
  6. Formal Definition
  7. Properties
  8. LendRecoup Mechanism
  9. Experiments
  10. Workloads
  11. Evaluation Metrics
  12. Nash Welfare (NW)
  13. Sharing Index (SIx)
  14. Weighted Min--Max (WMM)
  15. Normalized Min--Max (NMM)
  16. ...and 11 more sections

Key Result

Lemma 1

Fix a round $t$ with $\sum_{j} d'_{j,t} > E$. Let $M$ be a mechanism that is credit fair and Pareto efficient. Then, for every agent $i\in[n]$, it holds that $\Delta c_{i,t} \;=\; e_i - a_{i,t}$.

Theorems & Definitions (17)

  • Definition 1
  • Definition 2
  • Definition 3
  • Definition 4
  • Definition 5
  • Lemma 1
  • Theorem 1
  • Corollary 1
  • Proposition 1
  • Theorem 2
  • ...and 7 more