AI-based Online VoI-Aware Healthcare and Medical Monitoring Task Computing

Abstract—In recent years, there has been a significant advancement

in the field of healthcare systems with the introduction

of fifth generation cellular communications and beyond (5GB).

This development has paved the way for the utilization of

telecommunications technologies in healthcare systems with an

level of certainty, reaching up to 99.999 percent. In this paper,

we present a novel task computing framework that can address

the requirements of healthcare systems, such as reliability. In

this regard, we assume that IoT devices that are applied in the

considered healthcare have tasks with uncertain requirements.

On the other hand, we have uncertainty in the computing

resources in the healthcare servers. To address these uncertainties

that we obtain closed-form formulas. Furthermore, we adopt a

partial offloading approach to address the task of IoT devices.

Our goal in the proposed framework is to maximize the total

date rate of the healthcare system. To achieve this, we formulate

an optimization problem that considers a novel constraint that

guarantees the minimum value of information (VoI), minimum

data rate, and computational capacity constraints. To solve the

proposed optimization problem, we adapt a deep reinforcement

learning (DRL) based solution to effectively solve it, compared to

the other baselines. In this regard, we propose a soft actor critic

(SAC)-based algorithm, entitled SAC-based VoI-aware healthcare

networks (SACVAHC), that can address uncertainties exist in the

considered healthcare network. The results obtained show that

the proposed method can improve the total sum rate up to 20%,

compared to the other baselines.