OPTIMIZING CLOUD RESOURCE ALLOCATION USING MULTI AGENT DEEP REINFORCEMENT LEARNING-BASED ADAPTIVE HHO ALGORITHM

Abstract

Cloud computing workloads are projected to grow by 23.1% annually, with over 80% of enterprises adopting multi-cloud strategies, creating a pressing need for optimal virtual machine (VM) resource allocation to ensure cost efficiency and performance. However, existing allocation strategies suffer from static optimization limitations and inefficient adaptation to dynamic workloads, leading to frequent resource underutilization and service delays. To address these issues, this work proposes a novel Multi-Agent Deep Reinforcement Learning-Based Adaptive Harris Hawks Optimization (MADRL-AHHO) algorithm for cloud resource allocation using the VM Resource Allocation dataset (VM-0 to VM-5 classes). Initially, the dataset is preprocessed through normalization and feature selection to reduce dimensionality and noise. Feature extraction is enhanced using a Convolutional Neural Network (CNN) integrated with Firefly Optimization (CNN-FFO), which is benchmarked for learning capacity and solution convergence. However, performance limitations in CNN-FFO under dynamic load conditions are overcome by integrating CNN with Adaptive Harris Hawks Optimization (CNN-AHHO), which dynamically adjusts its exploration and exploitation capabilities based on multi-agent reinforcement feedback. The agents interact with the environment to learn optimal VM allocation policies by maximizing resource utilization and minimizing SLA violations. Experimental results demonstrate that CNN-AHHO outperforms CNN-FFO and traditional machine learning methods in terms of allocation accuracy, convergence rate, and computational efficiency, thereby offering a robust and adaptive solution for cloud infrastructure management.