A New Method for Solving the Task Offloading Problem in Smart Mobile Internet of Things Using an Intelligent Adaptive Learning-Based Optimization Algorithm
HaoJi Xia, Mehdi Darbandi · IETE Journal of Research · 2025
Recently, Internet of Things (IoT) has unlocked opportunities across different industries, utilizing big data analysis and real-time management to enable intelligent manufacturing. There are, however, some constraints when it comes to the growth of IoT and, in particular, the use of mobile devices, which are too limited in terms of processing power, battery power and memory resources; hence, the rise of the concept of Mobile Internet of Things (MIoT). MIoT aims to improve the performance of mobile devices by providing better computing and storage capabilities, which enhances the ability to support different applications. However, the existing methods of offloading tasks in the MioT environment are not effective. These approaches often face challenges such as low fault tolerance, program offloading that is only for one use, and shorter execution times. In order to address these drawbacks, this paper offers an intelligent Multiple Adaptive Learning-Based Penguin Search Optimization Algorithm (MLPeSOA) whose goal is to enhance the robustness of the task management mechanism designed for MIoT. This method improves on the previous population division method by also eliminating the inertia problem in the weight of the particle using a logistic chaotic function. As a result, the MLPeSOA allows penguins to explore the solution space with flexible search steps, reducing the chances of getting stuck in local optima or converging too early. Performance evaluations show that this proposed method achieves up to 10% improvements in execution time, cost, and energy consumption for running applications compared to recent alternatives in this field.