Fair Division of Indivisible Goods Among Strategic Agents
Siddharth Barman,Ganesh Ghalme,Shweta Jain,Pooja Kulkarni,Shivika Narang
DOI: https://doi.org/10.48550/arXiv.1901.09427
2019-01-27
Computer Science and Game Theory
Abstract:We study fair division of indivisible goods in a single-parameter environment. In particular, we develop truthful social welfare maximizing mechanisms for fairly allocating indivisible goods. Our fairness guarantees are in terms of solution concepts which are tailored to address allocation of indivisible goods and, hence, provide an appropriate framework for fair division of goods. This work specifically considers fairness in terms of envy freeness up to one good (EF1), maximin share guarantee (MMS), and Nash social welfare (NSW). Our first result shows that (in a single-parameter environment) the problem of maximizing welfare, subject to the constraint that the allocation of the indivisible goods is EF1, admits a polynomial-time, 1/2-approximate, truthful auction. We further prove that this problem is NP-Hard and, hence, an approximation is warranted. This hardness result also complements prior works which show that an arbitrary EF1 allocation can be computed efficiently. We also establish a bi-criteria approximation guarantee for the problem of maximizing social welfare under MMS constraints. In particular, we develop a truthful auction which efficiently finds an allocation wherein each agent gets a bundle of value at least $\left(1/2 - \varepsilon \right)$ times her maximin share and the welfare of the computed allocation is at least the optimal, here $\varepsilon >0$ is a fixed constant. We complement this result by showing that maximizing welfare is computationally hard even if one aims to only satisfy the MMS constraint approximately.