Abstract
Assume that two players have strict rankings over an even number of indivisible items. We propose two algorithms to find balanced allocations of these items that are maximin—maximize the minimum rank of the items that the players receive—and are envy-free and Pareto-optimal, if such allocations exist. To determine whether an envy-free allocation exists, we introduce a simple condition on preference profiles; in fact, our condition guarantees the existence of a maximin, envy-free, and Pareto-optimal allocation. Although not strategy-proof, our algorithms would be difficult to manipulate unless a player has complete information about its opponent’s ranking. We assess the applicability of the algorithms to real-world problems, such as allocating marital property in a divorce or assigning people to committees or projects.
Similar content being viewed by others
Notes
Procedures, based on cardinal utility, to divide an odd number of items are of course available, but the properties that we focus on would have to be redefined to allow cardinal comparisons of the players’ subsets.
Note that we consider only balanced allocations, i.e., the two players’ subsets have equal cardinality.
For a detailed study of how to rank sets of items, see Barbera et al. (2004).
See Bouveret et al. (2016) for a discussion.
Note that several distinct injective mappings may guarantee EF. Moreover, the injective mappings \(f_A\) and \(f_B\) may or may not be inverses. If EF is guaranteed by some bijection, then it is guaranteed by a canonical bijection defined as follows: Order A’s assigned items according to A’s preference and then order B’s items to A’s preference. The bijection maps the \(k{\text {th}}\) item on the first list to the \(k{\text {th}}\) item on the second list.
It could as well be called a minimax allocation, because minimizing the maximum depth, and maximizing the minimum depth, are equivalent processes, though in game theory maximin and minimax outcomes may not be the same.
Note that the maximin depth cannot be less than m, even for unbalanced allocations.
Step 4 is taken from the AL algorithm of Brams et al. (2014).
Repetitions of step 1 ensure the envy-freeness of the items it allocates, because a player is assigned its own singles, which it ranks at or above the current maximin rank, m, to the opponent’s singles, which it ranks below rank m. In other words, each player item-wise prefers its singles to the other player’s singles in each repetition.
However, if the least-preferred unassigned items are identical after the first assignment of singles, the procedure followed under ISD is exactly the same as under SD.
An allocation is PO if and only if it is the product of a sequence of sincere item choices by the players, i.e., choices that are consistent with the players’ preference rankings (Brams and King 2005). Thus in Example 1, if A and B sincerely choose items in the alternating order ABABABAB, they obtain the allocation (1234, 8765); if the alternating order is BABABABA, they obtain the allocation (1235, 8764). The first allocation is PO but neither EF nor MX, whereas the second is PO and EF but not MX.
If an SD allocation were not PO, there would be a preferred allocation of singles or doubles to each player. But at every stage at which singles are allocated, neither player would prefer any of its opponent’s singles to its own, so neither player can—by trading any of its singles for the other player’s singles—improve its allocation at that stage. As for the doubles, they are allocated according to AL, which guarantees that at least one allocation will be PO (Brams et al. 2014).
In fact, any “bottom-up” procedure involving two players who submit rankings is vulnerable to misrepresentation (Brams and Taylor 1999, pp. 21–24). Under such a procedure, one player chooses first—by giving its worst item to the other player—and the players anticipate a sequence of later choices (in which, for example, their choices alternate). Optimal choices in such a game constitute a Nash equilibrium (Kohler and Chandrasekaran 1971; Brams and Straffin 1979; Levine and Stange 2012), but they do not necessarily yield an EF allocation when one exists.
References
Aziz H (2015) A note on the undercut procedure. Soc Choice Welf 45(4):723–728
Aziz H, Gaspers S, Mackenzie S, Walsh T (2014) Fair assignment of indivisible objects under ordinal preferences. In: Proceedings of the 13th international conference on autonomous agents and multiagent systems (AAMAS’14)
Barbera S, Bossert W, Pattanaik P (2004) Ranking sets of objects. In: Barbera S, Hammond PJ, Seidl C (eds) Handbook of utility theory, vol 2. Kluwer Academic Publishers, Boston
Bouveret S, Chevaleyre Y, Maudet N (2016) Fair allocation of indivisible goods. In: Brandt F, Conitzer V, Endriss U, Lang J, Procaccia AD (eds) Handbook of computational social choice. Cambridge University Press, Cambridge
Bouveret S, Lang J (2008) Efficiency and envy-freeness in fair division of indivisible goods: logical representation and complexity. J Artif Intell Res 32:525–564
Bouveret S, Endriss U, Lang J (2010) Fair division under ordinal preferences: computing envy-free allocations of indivisible goods. In: Proceedings of the 19th European conference on artificial intelligence
Brams SJ, Edelman PH, Fishburn PC (2003) Fair division of indivisible goods. Theory Decis 55(2):147–180
Brams SJ, Fishburn PC (2000) Fair division of indivisible items between two people with identical preferences: envy-freeness, pareto-optimality, and equity. Soc Choice Wel 17(2):247–267
Brams SJ, Kilgour DM (2001) Fallback bargaining. Group Decis Negot 10(4):287–316
Brams SJ, Kilgour DM, Klamler C (2015) How to divide things fairly. Math Mag 88:5
Brams SJ, Kilgour DM, Klamler C (2014) Two-person fair division of indivisible items: an efficient, envy-free algorithm. Not AMS 61(2):130–141
Brams SJ, Kilgour DM, Klamler C (2012) The undercut procedure: an algorithm for the envy-free division of indivisible items. Soc Choice Welf 39:615–631
Brams SJ, King DL (2005) Efficient fair division: help the worst off or avoid envy? Ration Soc 17(4):387–421
Brams SJ, Straffin PD Jr (1979) Prisoners’ dilemma and professional sports drafts. Am Math Mon 86(2):80–88
Brams SJ, Taylor AD (1996) Fair division: from cake-cutting to dispute resolution. Cambridge University Press, New York
Brams SJ, Taylor AD (1999) The win-win solution: guaranteeing fair shares to everybody. Norton, New York
Budish E (2011) The combinatorial assigment problem: approximate competitive equilibrium from equal incomes. J Polit Econ 119(6):1061–1103
Edelman PH, Fishburn PC (2001) Fair division of indivisible items among people with similar preferences. Math Soc Sci 41(3):327–347
Herreiner D, Puppe C (2002) A simple procedure for finding equitable allocations of indivisbile items. Soc Choice Welf 19(2):415–430
Kohler DA, Chandrasekaran R (1971) A class of sequential games. Oper Res 19(2):270–277
Levine L, Stange KE (2012) How to make the most of a shared meal: plan the last bite first. Am Math Mon 119(7):550–565
Procaccia AD, Wang J (2014) Fair enough: guaranteeing approximate maximin shares. In: Proceedings of the 15th ACM conference on economics and computation
Pruhs K, Woeginger GJ (2012) Divorcing made easy. In: Kranakis E, Krizanc D, Luccio F (eds.) FUN 2012, LNCS 7288, Springer, Berlin, pp 305–314
Author information
Authors and Affiliations
Corresponding author
Additional information
We acknowledge the valuable comments of the audience of the Dagstuhl Seminar 15241 “Computational Social Choice: Theory and Applications”. Klamler was supported by the University of Graz project, “Voting-Selection-Decision”.
Rights and permissions
About this article
Cite this article
Brams, S.J., Kilgour, D.M. & Klamler, C. Maximin Envy-Free Division of Indivisible Items. Group Decis Negot 26, 115–131 (2017). https://doi.org/10.1007/s10726-016-9502-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10726-016-9502-x