An Additive Approximation Scheme for the Nash Social Welfare Maximization with Identical Additive Valuations

TL;DR

This paper introduces a novel additive PTAS for maximizing Nash social welfare in allocating identical additive goods, improving approximation performance over previous PTAS methods.

Contribution

It presents the first additive PTAS for Nash social welfare maximization with identical additive valuations, enhancing approximation accuracy.

Findings

The algorithm achieves an additive error of ε v_max.

It outperforms existing PTAS in approximation quality.

The method combines preprocessing with load balancing techniques.

Abstract

We study the problem of efficiently and fairly allocating a set of indivisible goods among agents with identical and additive valuations for the goods. The objective is to maximize the Nash social welfare, which is the geometric mean of the agents' valuations. While maximizing the Nash social welfare is NP-hard, a PTAS for this problem is presented by Nguyen and Rothe. The main contribution of this paper is to design a first additive PTAS for this problem, that is, we give a polynomial-time algorithm that maximizes the Nash social welfare within an additive error $\varepsilon v_{\rm max}$, where $\varepsilon$ is an arbitrary positive number and $v_{\rm max}$ is the maximum utility of a good. The approximation performance of our algorithm is better than that of a PTAS. The idea of our algorithm is simple; we apply a preprocessing and then utilize an additive PTAS for the target load balancing problem given recently by Buchem et al. However, a nontrivial amount of work is required to evaluate the additive error of the output.