Balancing workloads and optimizing QoS in cloud manufacturing through enhanced metaheuristics

Cloud Manufacturing (CMfg) enables flexible and customized manufacturing services through dynamic service composition. However, achieving optimal service composition remains challenging due to the need to meet complex Quality of Service (QoS) requirements, including cost, time, quality, and resource workload balance. Notably, previous studies on service composition models have rarely considered workload balancing as part of their QoS criteria, which is critical for maintaining efficient and sustainable resource use. This study addresses this gap by presenting an advanced service composition model that integrates workload balance as an essential QoS metric alongside traditional factors like composite service quality, time, and cost. To further support optimization, the Simulated Annealing (SA) and Tabu Search (TS) algorithms are enhanced with a novel shaking mechanism designed to expand the search space and mitigate premature convergence risks common in metaheuristics. Experimental evaluations conducted on an OR-Library dataset confirm that the enhanced SA algorithm achieves up to a 25% improvement in the fitness function and a 7% reduction in computational time, while the improved TS algorithm achieves a 2% reduction in the fitness function and a 21% decrease in computational time. These findings highlight the model's potential to enhance CMfg service composition efficiency, offering substantial performance benefits over traditional methods. The core contributions of this study include the development of a workload-integrated service composition model and enhancements to SA and TS algorithms for effective problem-solving within this framework.