Probabilistic Assignments of Identical Indivisible Objects and Uniform Probabilistic Rules.

We consider a probabilistic approach to the problem of assigning k indivisible identical objects to a set of agents with single-peaked preferences. Using the ordinal extension of preferences, we characterize the class of uniform probabilistic rules by Pareto efficiency, strategy-proofness, and no-envy. We also show that in this characterization no-envy cannot be replaced by anonymity. When agents are strictly risk averse von-Neumann-Morgenstern utility maximizers, then we reduce the problem of assigning k identical objects to a problem of allocating the amount k of an infinitely divisible commodity.

Allocation via Deferred-Acceptance under Responsive Priorities

In many economic environments - such as college admissions, student placements at public schools, and university housing allocation - indivisible objects with capacity constraints are assigned to a set of agents when each agent receives at most one object and monetary compensations are not allowed. In these important applications the agent-proposing deferred-acceptance algorithm with responsive priorities (called responsive DA-rule) performs well and economists have successfully implemented responsive DA-rules or slight variants thereof. First, for house allocation problems we characterize the class of responsive DA-rules by a set of basic and intuitive properties, namely, unavailable type invariance, individual rationality, weak non-wastefulness, resource-monotonicity, truncation invariance, and strategy-proofness. We extend this characterization to the full class of allocation problems with capacity constraints by replacing resource- monotonicity with two-agent consistent con ict resolution. An alternative characterization of responsive DA-rules is obtained using unassigned objects invariance, individual rationality, weak non-wastefulness, weak consistency, and strategy-proofness. Various characterizations of the class of "acyclic" responsive DA-rules are obtained by using the properties efficiency, group strategy-proofness, and consistency.

House allocation via deferred-acceptance

We study the simple model of assigning indivisible and heterogenous objects (e.g., houses, jobs, offi ces, etc.) to agents. Each agent receives at most one object and monetary compensations are not possible. For this model, known as the house allocation model, we characterize the class of rules satisfying unavailable object invariance, individual rationality, weak non-wastefulness, resource-monotonicity, truncation invariance, and strategy-proofness: any rule with these properties must allocate objects based on (implicitly induced) objects' priorities over agents and the agent-proposing deferred-acceptance-algorithm.

Strategy-proofness makes the difference: deferred-acceptance with responsive priorities

In college admissions and student placements at public schools, the admission decision can be thought of as assigning indivisible objects with capacity constraints to a set of students such that each student receives at most one object and monetary compensations are not allowed. In these important market design problems, the agent-proposing deferred-acceptance (DA-)mechanism with responsive strict priorities performs well and economists have successfully implemented DA-mechanisms or slight variants thereof. We show that almost all real-life mechanisms used in such environments - including the large classes of priority mechanisms and linear programming mechanisms - satisfy a set of simple and intuitive properties. Once we add strategy-proofness to these properties, DA-mechanisms are the only ones surviving. In market design problems that are based on weak priorities (like school choice), generally multiple tie-breaking (MTB)procedures are used and then a mechanism is implemented with the obtained strict priorities. By adding stability with respect to the weak priorities, we establish the first normative foundation for MTB-DA-mechanisms that are used in NYC.

Resource-Monotonicity for House Allocation Problems

We study a simple model of assigning indivisible objects (e.g., houses, jobs, offices, etc.) to agents. Each agent receives at most one object and monetary compensations are not possible. We completely describe all rules satisfying efficiency and resource-monotonicity. The characterized rules assign the objects in a sequence of steps such that at each step there is either a dictator or two agents who “trade” objects from their hierarchically specified “endowments.”

Consistent House Allocation

In practice we often face the problem of assigning indivisible objects (e.g., schools, housing, jobs, offices) to agents (e.g., students, homeless, workers, professors) when monetary compensations are not possible. We show that a rule that satisfies consistency, strategy-proofness, and efficiency must be an efficient generalized priority rule; i.e. it must adapt to an acyclic priority structure, except -maybe- for up to three agents in each object's priority ordering.

Centralized allocation in multiple markets

We study the problem of centralized allocation of indivisible objects in multiple markets. We show that the set of allocation rules that are group strategy-proof and Pareto-efficient are sequential dictatorships. Therefore, the solution of the joint al-location in multiple markets is significantly narrower than in the single-market case. Our result also applies to dynamic allocation problems. Finally, we provide conditions under which the solution of the single-market allocation coincides with the multiple-market case, and we apply this result to the study of the school choice problem with sibling priorities.

Budget-Balance, Fairness and Minimal Manipulability

A common real-life problem is to fairly allocate a number of indivisible objects and a fixed amount of money among a group of agents. Fairness requires that each agent weakly prefers his consumption bundle to any other agent’s bundle. Under fairness, efficiency is equivalent to budget-balance (all the available money is allocated among the agents). Budget-balance and fairness in general are incompatible with non-manipulability (Green and Laffont, 1979). We propose a new notion of the degree of manipulability which can be used to compare the ease of manipulation in allocation mechanisms. Our measure counts for each problem the number of agents who can manipulate the rule. Given this notion, the main result demonstrates that maximally linked fair allocation rules are the minimally manipulable rules among all budget-balanced and fair allocation mechanisms. Such rules link any agent to the bundle of a pre-selected agent through indifferences (which can be viewed as indirect egalitarian equivalence).

(Minimally) 'epsilon'-incentive compatible competitive equilibria in economies with indivisibilities

We consider competitive and budget-balanced allocation rules for problems where a number of indivisible objects and a fixed amount of money is allocated among a group of agents. In 'small' economies, we identify under classical preferences each agent's maximal gain from manipulation. Using this result we find the competitive and budget-balanced allocation rules which are minimally manipulable for each preference profile in terms of any agent's maximal gain. If preferences are quasi-linear, then we can find a competitive and budget-balanced allocation rule such that for any problem, the maximal utility gain from manipulation is equalized among all agents.

Efficient Priority Rules

We study the assignment of indivisible objects with quotas (houses, jobs, or offices) to a set of agents (students, job applicants, or professors). Each agent receives at most one object and monetary compensations are not possible. We characterize efficient priority rules by efficiency, strategy-proofness, and reallocation-consistency. Such a rule respects an acyclical priority structure and the allocations can be determined using the deferred acceptance algorithm.

Top trading with fixed tie-breaking in markets with indivisible goods

We study markets with indivisible goods where monetary compensations are not possible. Each individual is endowed with an object and a preference relation over all objects. When preferences are strict, Gale's top trading cycle algorithm finds the unique core allocation. When preferences are not necessarily strict, we use an exogenous profile of tie-breakers to resolve any ties in individuals' preferences and apply Gale's top trading cycle algorithm for the resulting profile of strict preferences. We provide a foundation of these simple extensions of Gale's top trading cycle algorithm from strict preferences to weak preferences. We show that Gale's top trading cycle algorithm with fixed tie-breaking is characterized by individual rationality, strategy-proofness, weak efficiency, non-bossiness, and consistency. Our result supports the common practice in applications to break ties in weak preferences using some fixed exogenous criteria and then to use a 'good and simple' rule for the resulting strict preferences. This reinforces the market-based approach even in the presence of indifferences because always competitive allocations are chosen.

Object allocation via deferred-acceptance: strategy-proofness and comparative statics

We study the problem of assigning indivisible and heterogenous objects (e.g., houses, jobs, offices, school or university admissions etc.) to agents. Each agent receives at most one object and monetary compensations are not possible. We consider mechanisms satisfying a set of basic properties (unavailable-type-invariance, individual-rationality, weak non-wastefulness, or truncation-invariance). In the house allocation problem, where at most one copy of each object is available, deferred-acceptance (DA)-mechanisms allocate objects based on exogenously fixed objects' priorities over agents and the agent-proposing deferred-acceptance-algorithm. For house allocation we show that DA-mechanisms are characterized by our basic properties and (i) strategy-proofness and population-monotonicity or (ii) strategy-proofness and resource-monotonicity. Once we allow for multiple identical copies of objects, on the one hand the first characterization breaks down and there are unstable mechanisms satisfying our basic properties and (i) strategy-proofness and population-monotonicity. On the other hand, our basic properties and (ii) strategy-proofness and resource-monotonicity characterize (the most general) class of DA-mechanisms based on objects' fixed choice functions that are acceptant, monotonic, substitutable, and consistent. These choice functions are used by objects to reject agents in the agent-proposing deferred-acceptance-algorithm. Therefore, in the general model resource-monotonicity is the «stronger» comparative statics requirement because it characterizes (together with our basic requirements and strategy-proofness) choice-based DA-mechanisms whereas population-monotonicity (together with our basic properties and strategy-proofness) does not.

A Convex Optimization Framework for Almost Budget Balanced Allocation of a Divisible Good

We address the problem of allocating a single divisible good to a number of agents. The agents have concave valuation functions parameterized by a scalar type. The agents report only the type. The goal is to find allocatively efficient, strategy proof, nearly budget balanced mechanisms within the Groves class. Near budget balance is attained by returning as much of the received payments as rebates to agents. Two performance criteria are of interest: the maximum ratio of budget surplus to efficient surplus, and the expected budget surplus, within the class of linear rebate functions. The goal is to minimize them. Assuming that the valuation functions are known, we show that both problems reduce to convex optimization problems, where the convex constraint sets are characterized by a continuum of half-plane constraints parameterized by the vector of reported types. We then propose a randomized relaxation of these problems by sampling constraints. The relaxed problem is a linear programming problem (LP). We then identify the number of samples needed for ``near-feasibility'' of the relaxed constraint set. Under some conditions on the valuation function, we show that value of the approximate LP is close to the optimal value. Simulation results show significant improvements of our proposed method over the Vickrey-Clarke-Groves (VCG) mechanism without rebates. In the special case of indivisible goods, the mechanisms in this paper fall back to those proposed by Moulin, by Guo and Conitzer, and by Gujar and Narahari, without any need for randomization. Extension of the proposed mechanisms to situations when the valuation functions are not known to the central planner are also discussed. Note to Practitioners-Our results will be useful in all resource allocation problems that involve gathering of information privately held by strategic users, where the utilities are any concave function of the allocations, and where the resource planner is not interested in maximizing revenue, but in efficient sharing of the resource. Such situations arise quite often in fair sharing of internet resources, fair sharing of funds across departments within the same parent organization, auctioning of public goods, etc. We study methods to achieve near budget balance by first collecting payments according to the celebrated VCG mechanism, and then returning as much of the collected money as rebates. Our focus on linear rebate functions allows for easy implementation. The resulting convex optimization problem is solved via relaxation to a randomized linear programming problem, for which several efficient solvers exist. This relaxation is enabled by constraint sampling. Keeping practitioners in mind, we identify the number of samples that assures a desired level of ``near-feasibility'' with the desired confidence level. Our methodology will occasionally require subsidy from outside the system. We however demonstrate via simulation that, if the mechanism is repeated several times over independent instances, then past surplus can support the subsidy requirements. We also extend our results to situations where the strategic users' utility functions are not known to the allocating entity, a common situation in the context of internet users and other problems.

Sparsely Faceted Arrays: A Mechanism Supporting Parallel Allocation, Communication, and Garbage Collection

Conventional parallel computer architectures do not provide support for non-uniformly distributed objects. In this thesis, I introduce sparsely faceted arrays (SFAs), a new low-level mechanism for naming regions of memory, or facets, on different processors in a distributed, shared memory parallel processing system. Sparsely faceted arrays address the disconnect between the global distributed arrays provided by conventional architectures (e.g. the Cray T3 series), and the requirements of high-level parallel programming methods that wish to use objects that are distributed over only a subset of processing elements. A sparsely faceted array names a virtual globally-distributed array, but actual facets are lazily allocated. By providing simple semantics and making efficient use of memory, SFAs enable efficient implementation of a variety of non-uniformly distributed data structures and related algorithms. I present example applications which use SFAs, and describe and evaluate simple hardware mechanisms for implementing SFAs. Keeping track of which nodes have allocated facets for a particular SFA is an important task that suggests the need for automatic memory management, including garbage collection. To address this need, I first argue that conventional tracing techniques such as mark/sweep and copying GC are inherently unscalable in parallel systems. I then present a parallel memory-management strategy, based on reference-counting, that is capable of garbage collecting sparsely faceted arrays. I also discuss opportunities for hardware support of this garbage collection strategy. I have implemented a high-level hardware/OS simulator featuring hardware support for sparsely faceted arrays and automatic garbage collection. I describe the simulator and outline a few of the numerous details associated with a "real" implementation of SFAs and SFA-aware garbage collection. Simulation results are used throughout this thesis in the evaluation of hardware support mechanisms.

Asymmetrically fair rules for an indivisible good problem with a budget constraint

We study a particular restitution problem where there is an indivisible good (land or property) over which two agents have rights: the dispossessed agent and the owner. A third party, possibly the government, seeks to resolve the situation by assigning rights to one and compensate the other. There is also a maximum amount of money available for the compensation. We characterize a family of asymmetrically fair rules that are immune to strategic behavior, guarantee minimal welfare levels for the agents, and satisfy the budget constraint.