Reinforcement Learning-based Task Offloading in the Internet of Wearable Things

TL;DR

This work addresses the challenge of executing latency-sensitive tasks on energy-constrained wearables by offloading to nearby edge resources, specifically a smartphone acting as an edge server. It formulates the task offloading problem as a model-free Markov Decision Process and applies a lightweight Q-learning policy to jointly optimize task completion time and energy consumption, including idle energy during edge execution. The approach is evaluated in ns-3 across multiple applications, demonstrating that Q-learning can substantially reduce task completion time for heavy workloads while managing energy use, with performance tunable via the time-energy tradeoff weights. The study highlights practical considerations for IoWT deployments and suggests extensions to multi-layer edge architectures, real-device validation, and more flexible task models, underscoring the method's potential to enable responsive, energy-efficient wearable computing.

Abstract

Over the years, significant contributions have been made by the research and industrial sectors to improve wearable devices towards the Internet of Wearable Things (IoWT) paradigm. However, wearables are still facing several challenges. Many stem from the limited battery power and insufficient computation resources available on wearable devices. On the other hand, with the popularity of smart wearables, there is a consistent increase in the development of new computationally intensive and latency-critical applications. In such a context, task offloading allows wearables to leverage the resources available on nearby edge devices to enhance the overall user experience. This paper proposes a framework for Reinforcement Learning (RL)-based task offloading in the IoWT. We formulate the task offloading process considering the tradeoff between energy consumption and task accomplishment time. Moreover, we model the task offloading problem as a Markov Decision Process (MDP) and utilize the Q-learning technique to enable the wearable device to make optimal task offloading decisions without prior knowledge. We evaluate the performance of the proposed framework through extensive simulations for various applications and system configurations conducted in the ns-3 network simulator. We also show how varying the main system parameters of the Q-learning algorithm affects the overall performance in terms of average task accomplishment time, average energy consumption, and percentage of tasks offloaded.

Figures (12)

• Figure 1: Overview of the IoWT and MEC concepts
• Figure 2: System architecture and the Q-learning-based task offloading process
• Figure 3: Average task accomplishment time for different applications for 3 task execution scenarios ($\beta_{E}=\beta_{T}=0.5$).
• Figure 4: Average energy consumption for different applications for 3 task execution scenarios ($\beta_{E}=\beta_{T}=0.5$).
• Figure 5: Average energy consumption breakdown for different applications for Local execution vs. Offloading scenarios.