Part of Advances in Neural Information Processing Systems 35 (NeurIPS 2022) Main Conference Track
Francisco Criado, David Martinez-Rubio, Sebastian Pokutta
The proportional fair resource allocation problem is a major problem studied in flow control of networks, operations research, and economic theory, where it has found numerous applications. This problem, defined as the constrained maximization of $\sum_i \log x_i$, is known as the packing proportional fairness problem when the feasible set is defined by positive linear constraints and $x \in \mathbb{R}_{\geq 0}^n$. In this work, we present a distributed accelerated first-order method for this problem which improves upon previous approaches. We also design an algorithm for the optimization of its dual problem. Both algorithms are width-independent.