arXiv subjects

Eklavya Sharma

Publications and source records attributed to Eklavya Sharma.

17 recordsLinked to original sources

arXiv2026

Equilibrium Pricing in Oligopolistic Data Markets

We study equilibrium pricing in oligopolistic data markets with budget-constrained buyers (e.g., machine learning companies purchasing data to improve model accuracy) and strategic data sellers. Sellers compete by setting prices for their datasets, giving rise to a pricing game whose pure Nash equilibria correspond to equilibrium prices. While equilibrium prices are guaranteed for rivalrous goods via competitive equilibrium, we show that the non-rivalry of data fundamentally alters this picture: an exact Nash equilibrium (NE) need not exist, and in fact, 1.363-approximate NE may also not exist under uniform pricing. We therefore investigate relaxed equilibrium notions. Allowing sellers to use beyond-uniform pricing---specifically, piecewise-linear convex pricing functions---guarantees approximate stability within a constant factor: there exists a pricing profile in which no seller can improve revenue by a factor of two by deviating to any uniform price (a 2-approximate NE against uniform deviations). Finally, our simulations demonstrate fast convergence and empirical approximation guarantees that outperform the worst-case bound of 2.

cs.GT↗

arXiv2026

Revenue-Optimal Pricing for Budget-Constrained Buyers in Data Markets

We study revenue-optimal pricing in data markets with rational, budget-constrained buyers. Such a market offers multiple datasets for sale, and buyers aim to improve the accuracy of their prediction tasks by acquiring data bundles. The market's objective is to price datasets to maximize total revenue, considering that buyers with quasi-linear utilities choose their bundles optimally under budget constraints. We allow the buyers to purchase fractions of datasets, and the amount they pay is proportional to the fraction they receive. Although competitive equilibrium gives revenue-optimal pricing in rivalrous markets with quasi-linear buyers, we show that revenue maximization in data markets is APX-hard. Despite the hardness, we design a 2-approximation algorithm when datasets arrive online, and a $(1-1/e)^{-1}$-approximation algorithm for the offline setting.

cs.GT↗

arXiv2025

Proportional and Pareto-Optimal Allocation of Chores with Subsidy

We consider the problem of allocating $m$ indivisible chores among $n$ agents with possibly different weights, aiming for a solution that is both fair and efficient. Specifically, we focus on the classic fairness notion of proportionality and efficiency notion of Pareto-optimality. Since proportional allocations may not always exist in this setting, we allow the use of subsidies (monetary compensation to agents) to ensure agents are proportionally-satisfied, and aim to minimize the total subsidy required. Wu and Zhou (WINE 2024) showed that when each chore has disutility at most 1, a total subsidy of at most $n/3 - 1/6$ is sufficient to guarantee proportionality. However, their approach is based on a complex technique, which does not guarantee economic efficiency - a key desideratum in fair division. In this work, we give a polynomial-time algorithm that achieves the same subsidy bound while also ensuring Pareto-optimality. Moreover, both our algorithm and its analysis are significantly simpler than those of Wu and Zhou (WINE 2024). Our approach first computes a proportionally-fair competitive equilibrium, and then applies a rounding procedure guided by minimum-pain-per-buck edges.

cs.GT↗

arXiv2025

Automating the Search for Small Hard Examples to Approximation Algorithms

Given an approximation algorithm $A$, we want to find the input with the worst approximation ratio, i.e., the input for which $A$'s output's objective value is the worst possible compared to the optimal solution's objective value. Such hard examples shed light on the approximation algorithm's weaknesses, and could help us design better approximation algorithms. When the inputs are discrete (e.g., unweighted graphs), one can find hard examples for small input sizes using brute-force enumeration. However, it's not obvious how to do this when the input space is continuous, as in makespan minimization or bin packing. We develop a technique for finding small hard examples for a large class of approximation algorithms. Our algorithm works by constructing a decision tree representation of the approximation algorithm and then running a linear program for each leaf node of the decision tree. We implement our technique in Python, and demonstrate it on the longest-processing-time (LPT) heuristic for makespan minimization.

cs.DS↗

arXiv2025

Exploring Relations among Fairness Notions in Discrete Fair Division

Fair allocation of indivisible items among agents is a fundamental and extensively studied problem. However, fairness does not have a single universally accepted definition, leading to many competing fairness notions. Some of these notions are considered stronger or more desirable, but they are also more difficult to guarantee. In this work, we examine 22 different fairness notions and organize them into a hierarchy. Formally, we say that a notion $F_1$ implies another notion $F_2$ if every $F_1$-fair allocation is also $F_2$-fair. We give a near-complete picture of implications among fairness notions: for almost every pair of notions, we either prove an implication or give a counterexample demonstrating that the implication does not hold. Although some of these results are already known, many are new. We examine multiple settings, including the allocation of goods, chores, and mixed manna, and different valuation classes like additive, submodular, and subadditive. We believe this work clarifies the relative strengths and applicability of these notions, providing a foundation for future research in fair division. Moreover, we develop an inference engine to automate part of our work. It is available as a user-friendly web application and may have broader applications beyond fair division.

cs.GT↗

arXiv2024

EF1 for Mixed Manna with Unequal Entitlements

We study fair division of indivisible mixed manna when agents have unequal entitlements, with weighted envy-freeness up to one item (WEF1) as our primary notion of fairness. We identify several shortcomings of existing techniques to achieve WEF1. Hence, we relax WEF1 to weighted envy-freeness up to 1 transfer (WEF1T), and give a polynomial-time algorithm for achieving it. We also generalize Fisher markets to the mixed manna setting, and use them to get a polynomial-time algorithm for two agents that outputs a WEF1 allocation.

cs.GT↗

arXiv2024

Best-of-Both-Worlds Fairness of the Envy-Cycle-Elimination Algorithm

We consider the problem of fairly dividing indivisible goods among agents with additive valuations. It is known that an Epistemic EFX and $2/3$-MMS allocation can be obtained using the Envy-Cycle-Elimination (ECE) algorithm. In this work, we explore whether this algorithm can be randomized to also ensure ex-ante proportionality. For two agents, we show that a randomized variant of ECE can compute an ex-post EFX and ex-ante envy-free allocation in near-linear time. However, for three agents, we show that several natural randomization methods for ECE fail to achieve ex-ante proportionality.

cs.GT↗

arXiv2023

Nash Equilibria of Two-Player Matrix Games Repeated Until Collision

We introduce and initiate the study of a natural class of repeated two-player matrix games, called Repeated-Until-Collision (RUC) games. In each round, both players simultaneously pick an action from a common action set $\{1, 2, \dots, n\}$. Depending on their chosen actions, they derive payoffs given by $n \times n$ matrices $A$ and $B$, respectively. If their actions collide (i.e., they pick the same action), the game ends, otherwise, it proceeds to the next round. Both players want to maximize their total payoff until the game ends. RUC games can be interpreted as pursuit-evasion games or repeated hide-and-seek games. They also generalize hand cricket, a popular game among children in India. We show that under mild assumptions on the payoff matrices, every RUC game admits a Nash equilibrium (NE). Moreover, we show the existence of a stationary NE, where each player chooses their action according to a probability distribution over the action set that does not change across rounds. Remarkably, we show that all NE are effectively the same as the stationary NE, thus showing that RUC games admit an almost unique NE. Lastly, we also show how to compute (approximate) NE for RUC games.

cs.GT↗

arXiv2023

Improving Approximation Guarantees for Maximin Share

We consider fair division of a set of indivisible goods among $n$ agents with additive valuations using the fairness notion of maximin share (MMS). MMS is the most popular share-based notion, in which an agent finds an allocation fair to her if she receives goods worth at least her ($1$-out-of-$n$) MMS value. An allocation is called MMS if all agents receive their MMS values. However, since MMS allocations do not always exist, the focus shifted to investigating its ordinal and multiplicative approximations. In the ordinal approximation, the goal is to show the existence of $1$-out-of-$d$ MMS allocations (for the smallest possible $d>n$). A series of works led to the state-of-the-art factor of $d=\lfloor3n/2\rfloor$ [Hosseini et al.'21]. We show that $1$-out-of-$4\lceil n/3\rceil$ MMS allocations always exist, thereby improving the state-of-the-art of ordinal approximation. In the multiplicative approximation, the goal is to show the existence of $\alpha$-MMS allocations (for the largest possible $\alpha < 1$), which guarantees each agent at least $\alpha$ times her MMS value. We introduce a general framework of "approximate MMS with agent priority ranking". An allocation is said to be $T$-MMS, for a non-increasing sequence $T = (\tau_1, \ldots, \tau_n)$ of numbers, if the agent at rank $i$ in the order gets a bundle of value at least $\tau_i$ times her MMS value. This framework captures both ordinal approximation and multiplicative approximation as special cases. We show the existence of $T$-MMS allocations where $\tau_i \ge \max(\frac{3}{4} + \frac{1}{12n}, \frac{2n}{2n+i-1})$ for all $i$. Furthermore, we can get allocations that are $(\frac{3}{4} + \frac{1}{12n})$-MMS ex-post and $(0.8253 + \frac{1}{36n})$-MMS ex-ante. We also prove that our algorithm does not give better than $(0.8631 + \frac{1}{2n})$-MMS ex-ante.

cs.GT↗

arXiv2023

Simplification and Improvement of MMS Approximation

We consider the problem of fairly allocating a set of indivisible goods among $n$ agents with additive valuations, using the popular fairness notion of maximin share (MMS). Since MMS allocations do not always exist, a series of works provided existence and algorithms for approximate MMS allocations. The Garg-Taki algorithm gives the current best approximation factor of $(\frac{3}{4} + \frac{1}{12n})$. Most of these results are based on complicated analyses, especially those providing better than $2/3$ factor. Moreover, since no tight example is known of the Garg-Taki algorithm, it is unclear if this is the best factor of this approach. In this paper, we significantly simplify the analysis of this algorithm and also improve the existence guarantee to a factor of $(\frac{3}{4} + \min(\frac{1}{36}, \frac{3}{16n-4}))$. For small $n$, this provides a noticeable improvement. Furthermore, we present a tight example of this algorithm, showing that this may be the best factor one can hope for with the current techniques.

cs.GT↗

arXiv2022

New Fairness Concepts for Allocating Indivisible Items

For the fundamental problem of fairly dividing a set of indivisible items among agents, envy-freeness up to any item (EFX) and maximin fairness (MMS) are arguably the most compelling fairness concepts proposed until now. Unfortunately, despite significant efforts over the past few years, whether EFX allocations always exist is still an enigmatic open problem, let alone their efficient computation. Furthermore, today we know that MMS allocations are not always guaranteed to exist. These facts weaken the usefulness of both EFX and MMS, albeit their appealing conceptual characteristics. We propose two alternative fairness concepts, called epistemic EFX (EEFX) and minimum EFX share fairness (MXS), inspired by EFX and MMS. For both, we explore their relationships to well-studied fairness notions and, more importantly, prove that EEFX and MXS allocations always exist and can be computed efficiently for additive valuations. Our results justify that the new fairness concepts can be excellent alternatives to EFX and MMS.

cs.GT↗

arXiv2021

Harmonic Algorithms for Packing d-dimensional Cuboids Into Bins

We explore approximation algorithms for the $d$-dimensional geometric bin packing problem ($d$BP). Caprara (MOR 2008) gave a harmonic-based algorithm for $d$BP having an asymptotic approximation ratio (AAR) of $T_{\infty}^{d-1}$ (where $T_{\infty} \approx 1.691$). However, their algorithm doesn't allow items to be rotated. This is in contrast to some common applications of $d$BP, like packing boxes into shipping containers. We give approximation algorithms for $d$BP when items can be orthogonally rotated about all or a subset of axes. We first give a fast and simple harmonic-based algorithm having AAR $T_{\infty}^{d}$. We next give a more sophisticated harmonic-based algorithm, which we call $\mathtt{HGaP}_k$, having AAR $T_{\infty}^{d-1}(1+ε)$. This gives an AAR of roughly $2.860 + ε$ for 3BP with rotations, which improves upon the best-known AAR of $4.5$. In addition, we study the multiple-choice bin packing problem that generalizes the rotational case. Here we are given $n$ sets of $d$-dimensional cuboidal items and we have to choose exactly one item from each set and then pack the chosen items. Our algorithms also work for the multiple-choice bin packing problem. We also give fast and simple approximation algorithms for the multiple-choice versions of $d$D strip packing and $d$D geometric knapsack.

cs.CG↗

arXiv2021

Geometry Meets Vectors: Approximation Algorithms for Multidimensional Packing

We study the generalized multidimensional bin packing problem (GVBP) that generalizes both geometric packing and vector packing. Here, we are given $n$ rectangular items where the $i^{\textrm{th}}$ item has width $w(i)$, height $h(i)$, and $d$ nonnegative weights $v_1(i), v_2(i), \ldots, v_{d}(i)$. Our goal is to get an axis-parallel non-overlapping packing of the items into square bins so that for all $j \in [d]$, the sum of the $j^{\textrm{th}}$ weight of items in each bin is at most 1. This is a natural problem arising in logistics, resource allocation, and scheduling. Despite being well studied in practice, surprisingly, approximation algorithms for this problem have rarely been explored. We first obtain two simple algorithms for GVBP having asymptotic approximation ratios $6(d+1)$ and $3(1 + \ln(d+1) + \varepsilon)$. We then extend the Round-and-Approx (R&A) framework [Bansal-Khan, SODA'14] to wider classes of algorithms, and show how it can be adapted to GVBP. Using more sophisticated techniques, we obtain better approximation algorithms for GVBP, and we get further improvement by combining them with the R&A framework. This gives us an asymptotic approximation ratio of $2(1+\ln((d+4)/2))+\varepsilon$ for GVBP, which improves to $2.919+\varepsilon$ for the special case of $d=1$. We obtain further improvement when the items are allowed to be rotated. We also present algorithms for a generalization of GVBP where the items are high dimensional cuboids.

cs.DS↗

arXiv2021

Tight Approximation Algorithms for Geometric Bin Packing with Skewed Items

In the Two-dimensional Bin Packing (2BP) problem, we are given a set of rectangles of height and width at most one and our goal is to find an axis-aligned nonoverlapping packing of these rectangles into the minimum number of unit square bins. The problem admits no APTAS and the current best approximation ratio is $1.406$ by Bansal and Khan [SODA'14]. A well-studied variant of the problem is Guillotine Two-dimensional Bin Packing (G2BP), where all rectangles must be packed in such a way that every rectangle in the packing can be obtained by recursively applying a sequence of end-to-end axis-parallel cuts, also called guillotine cuts. Bansal, Lodi, and Sviridenko [FOCS'05] obtained an APTAS for this problem. Let $λ$ be the smallest constant such that for every set $I$ of items, the number of bins in the optimal solution to G2BP for $I$ is upper bounded by $λ\operatorname{opt}(I) + c$, where $\operatorname{opt}(I)$ is the number of bins in the optimal solution to 2BP for $I$ and $c$ is a constant. It is known that $4/3 \le λ\le 1.692$. Bansal and Khan [SODA'14] conjectured that $λ= 4/3$. The conjecture, if true, will imply a $(4/3+\varepsilon)$-approximation algorithm for 2BP. According to convention, for a given constant $δ>0$, a rectangle is large if both its height and width are at least $δ$, and otherwise it is called skewed. We make progress towards the conjecture by showing $λ= 4/3$ for skewed instance, i.e., when all input rectangles are skewed. Even for this case, the previous best upper bound on $λ$ was roughly 1.692. We also give an APTAS for 2BP for skewed instance, though general 2BP does not admit an APTAS.

cs.CG↗

arXiv2021

Approximation Algorithms for Generalized Multidimensional Knapsack

We study a generalization of the knapsack problem with geometric and vector constraints. The input is a set of rectangular items, each with an associated profit and $d$ nonnegative weights ($d$-dimensional vector), and a square knapsack. The goal is to find a non-overlapping axis-parallel packing of a subset of items into the given knapsack such that the vector constraints are not violated, i.e., the sum of weights of all the packed items in any of the $d$ dimensions does not exceed one. We consider two variants of the problem: $(i)$ the items are not allowed to be rotated, $(ii)$ items can be rotated by 90 degrees. We give a $(2+ε)$-approximation algorithm for this problem (both versions). In the process, we also study a variant of the maximum generalized assignment problem (Max-GAP), called Vector-Max-GAP, and design a PTAS for it.

cs.DS↗

arXiv2020

An Approximation Algorithm for Covering Linear Programs and its Application to Bin-Packing

We give an $α(1+ε)$-approximation algorithm for solving covering LPs, assuming the presence of a $(1/α)$-approximation algorithm for a certain optimization problem. Our algorithm is based on a simple modification of the Plotkin-Shmoys-Tardos algorithm (MOR 1995). We then apply our algorithm to $α(1+ε)$-approximately solve the configuration LP for a large class of bin-packing problems, assuming the presence of a $(1/α)$-approximate algorithm for the corresponding knapsack problem (KS). Previous results give us a PTAS for the configuration LP using a PTAS for KS. Those results don't extend to the case where KS is poorly approximated. Our algorithm, however, works even for polynomially-large $α$.

cs.DS↗

arXiv2020

Analysis of the Harmonic Function Used in Bin-Packing

The harmonic function was first introduced by Lee and Lee (JACM 1985) for analyzing their online bin-packing algorithm. Subsequently, it has been used to obtain approximation algorithms for many different packing problems. Here we slightly generalize the harmonic function and give alternative proofs of its important properties.

cs.DS↗