Resource management based on fuzzy classifier in IoT hierarchical environment

Efficient resource management across heterogeneous Internet of Things (IoT) devices is a critical challenge. This paper proposes a hierarchical multi-layer fuzzy inference model that extends conventional Fog–Cloud task classification to a hierarchical IoT architecture, providing a richer decision space for computing layer selection. Building on this model, a Fuzzy Classifier–Deadline Sorted Task-Based Allocation (FC-DSTBA) framework is developed by integrating the fuzzy classifier with a novel Deadline Sorted Task-Based Allocation (DSTBA) for efficient task allocation. The proposed framework is implemented and evaluated using the iFogSim simulator over a 100–1000 heterogeneous task range, latency-sensitive IoT task profiles. The FC-DSTBA framework achieves substantial performance improvements over traditional benchmarks, reducing Makespan by an average of 30.5%, 81%, and 83%, increasing the Guarantee Ratio by 2%, 16%, and 22%, reducing Delay by 80–99%, and lowering Energy Consumption by 72.5–73.7% with respect to cloud-based DSTBA, SJF, and FCFS, respectively. Furthermore, FC-DSTBA consistently achieves a superior, lower Makespan value than the state-of-the-art AEOSSA algorithm across the entire range. Crucially, a rigorous paired t-test analysis confirms the statistical significance of these improvements, proving a significant lower mean Makespan (6.51) for FC-DSTBA compared to DSTBA (9.37, p = 0.0056), SJF (35.05, p = 0.0038), and FCFS (38.14, p = 0.0033).