An Online Algorithm for Computation Offloading in Non-Stationary Environments

TL;DR

This paper introduces an online algorithm for task offloading that minimizes latency in dynamic wireless environments by modeling server selection as a multi-armed bandit problem, outperforming existing methods.

Contribution

The paper proposes a novel online learning algorithm based on optimism in the face of uncertainty tailored for non-stationary environments, emphasizing the importance of discounting past rewards.

Findings

Outperforms state-of-the-art algorithms by up to ~1 second in latency reduction.

Heavily discounting past rewards improves adaptability in dynamic environments.

Model effectively captures the temporal variability of wireless links and server availability.

Abstract

We consider the latency minimization problem in a task-offloading scenario, where multiple servers are available to the user equipment for outsourcing computational tasks. To account for the temporally dynamic nature of the wireless links and the availability of the computing resources, we model the server selection as a multi-armed bandit (MAB) problem. In the considered MAB framework, rewards are characterized in terms of the end-to-end latency. We propose a novel online learning algorithm based on the principle of optimism in the face of uncertainty, which outperforms the state-of-the-art algorithms by up to ~1s. Our results highlight the significance of heavily discounting the past rewards in dynamic environments.