Combined optimization strategy for IoT resource allocation with workload prediction

A significant limitation in IoT technology is the challenge of handling the diverse and dynamic nature of IoT workloads, which complicates accurate workload prediction and efficient resource allocation. IoT devices generate vast amounts of heterogeneous data with varying speeds, volumes, and varieties, making traditional methods inadequate for managing this variability and leading to inefficient resource management, suboptimal performance, and increased operational costs. To address these issues, this research proposes a novel hybrid optimization algorithm known as the Lyrebird-Adapted Kookaburra Optimization Algorithm-Improved Analytic Hierarchy Process (LAKO-IAHP) for work load prediction and resource allocation. This approach includes two main phases: the Improved Analytic Hierarchy Process (IAHP) for workload prediction and the LAKO algorithm for resource allocation. The IAHP phase enhances conventional Analytic Hierarchy Process (AHP) techniques by incorporating the Improved k-means clustering (IKMC) process and Euclidean distance calculations to improve the accuracy of workload predictions by considering specific Load Balancing (LB) parameters such as server load and response time. Following this, the LAKO algorithm- an advanced hybrid method combining Kookaburra Optimization Algorithm (KOA) and Lyrebird Optimization Algorithm (LOA)- performs the resource allocation phase, that considers the Quality of Service (QoS) parameters including degree of imbalance, execution time, reliability, and resource utilization. The effectiveness of the LAKO-IAHP approach is demonstrated through various performance metrics and comparisons with existing methods, proving its capability to enhance resource management and maintain high performance and reliability in IoT environments.