Mobility‐Aware Dynamic Task Offloading in Fog Computing: A Multi‐Objective Discrete PSO Approach

ABSTRACT Fog computing extends the Cloud to the network Edge, reducing latency for Internet of Things (IoT) applications. The dynamic Fog computing environment and user mobility present significant challenges in optimizing resource allocation and maintaining service quality. Task offloading decisions involve balancing multiple objectives, such as minimizing execution delay while constraining network usage and energy consumption. To address these challenges, a modified discrete Particle Swarm Optimization (PSO) based Mobility‐aware Dynamic Task Offloading (MDTO) algorithm is presented in this paper. MDTO introduces a modified velocity update mechanism tailored for the discrete decision space of task‐to‐device mapping and employs a weighted fitness function that simultaneously optimizes task completion rates and physical offloading distance. We implement the MDTO algorithm utilizing the iFogSim2 simulator and conduct a comprehensive performance evaluation using the Remote Patient Monitoring (RPM) healthcare application case study. The performance of MDTO is compared against three baseline strategies: No offloading, a cluster‐based heuristic approach (FCTO), and a Genetic Algorithm‐based Task Offloading (GATO) strategy. Experimental results show that MDTO significantly outperforms these baseline strategies. Specifically, MDTO reduces latency, achieving an average loop delay of 153.17 ms, and minimizes the response time for offloaded tasks by approximately 58% compared to FCTO and GATO. Additionally, MDTO results in the lowest average offloading distance of 62.46 meters, in contrast to over 150 meters for FCTO and GATO. These findings validate MDTO as a promising solution for time‐sensitive, mobility‐aware Fog computing applications.