Decision tree based energy aware and delay optimized task offloading in IoT networks

The Internet of Things (IoT) has emerged as a key enabling technology in modern digital systems, supporting a wide range of applications such as smart cities, healthcare, and intelligent transportation. Despite its advantages, IoT devices are typically constrained by limited computational capability and energy resources, which makes efficient task execution and quality-of-service maintenance challenging. Task offloading to edge or cloud servers has been widely adopted to address these limitations; however, many existing approaches struggle to jointly optimise energy consumption and processing delay, particularly under dynamic network conditions. This paper presents a lightweight decision tree-based task offloading approach for IoT networks that enables fast and adaptive decision-making. The proposed method evaluates key real-time parameters, including energy consumption, battery level, network delay, and a network load indicator, to determine optimal offloading decisions. Owing to its hierarchical structure and low computational complexity, the approach adapts effectively to changing network and device conditions without relying on computationally intensive learning models. The performance of the proposed approach is evaluated using a MATLAB-based discrete-event simulation environment and compared with state-of-the-art task offloading methods, including rule-based and deep learning-based approaches. Simulation results show that the proposed method reduces total energy consumption by 26.5% compared to rule-based methods and by 15% compared to deep learning-based methods. In addition, network delay is reduced by 28% and 10%, respectively. The proposed approach also achieves a task allocation accuracy of 92% and a task offloading success rate of 96%. These results demonstrate that the proposed decision tree-based model provides an effective, energy-efficient, and delay-aware solution for task offloading in dynamic IoT networks.