Advanced hybrid optimization algorithms for energy efficient cloud resource allocation

Abstract

Cloud computing is highly sought after for its dynamic resource allocation capabilities and pay-per-use model. However, previous research has identified several challenges, such as lower coverage, high integration rates, longer computation times, and complex operators, all of which are associated with NP-hard problems. These issues negatively impact the efficiency of resource allocation and scheduling, leading to slower processes, inefficiencies in multi-objective optimization, lower throughput, and higher power consumption. To address these challenges, we propose a unique Hybridized Optimization Algorithm that integrates Crow Swarm Optimization (CSO), Cuckoo Search Optimization (CSO), and Cat Hunting Optimization (CHO). Initially, Particle Swarm Optimization (PSO) and the Crow Search Algorithm (CSA) handle the exploitation and exploration phases to balance task loads. Subsequently, Cuckoo Search Optimization (CSO) enhances resource utilization and addresses NP-hard issues, while Cat Hunting Optimization (CHO) refines the search from global to local optimal spaces to achieve the best values. The results demonstrate that the proposed hybrid technique effectively reduces user request waiting times, lowers energy consumption, and decreases execution times on cloud servers compared to baseline approaches, thereby significantly improving overall system performance.