Exponentially Weighted Algorithm for Online Network Resource Allocation with Long-Term Constraints

TL;DR

This paper introduces a randomized exponentially weighted algorithm for online network resource allocation that effectively manages long-term constraints, minimizes costs, and outperforms reinforcement learning in numerical tests.

Contribution

The paper proposes a novel exponentially weighted algorithm that handles long-term constraints in online network resource allocation, with proven regret bounds and superior empirical performance.

Findings

The algorithm achieves a tight upper bound on regret.

It maintains constraints with bounded cumulative violations.

Numerical experiments show it outperforms reinforcement learning methods.

Abstract

This paper studies an online optimal resource reservation problem in communication networks with job transfers where the goal is to minimize the reservation cost while maintaining the blocking cost under a certain budget limit. To tackle this problem, we propose a novel algorithm based on a randomized exponentially weighted method that encompasses long-term constraints. We then analyze the performance of our algorithm by establishing an upper bound for the associated regret and the cumulative constraint violations. Finally, we present numerical experiments where we compare the performance of our algorithm with those of reinforcement learning where we show that our algorithm surpasses it.