Abstract

A contemporary business concept that has become popular is cloud computing (CC), which gives consumers unlimited access to virtual resources. This paradigm provides facilities for virtual machines (VMs) and data centers. Infrastructure as a service (IaaS) is one of the several service models. The secret to improving data center efficiency and lowering energy usage is the effective use and administration of the resources offered by these services. Because of the variety and computational complexity of this activity, effective task scheduling—which moves cloud jobs onto virtual machines (VMs)—is essential. To enhance this process, meta-heuristic techniques are frequently applied. This research offers a new adaptive elitism-based genetic algorithm (AGA-E), which integrates elitism with conditional parameter adjustment to increase convergence speed and solution quality. The conditional parameter tuning technique constantly modifies the algorithm's parameters in response to population diversity and fitness levels, whereas the elitist method maintains the best-performing solution over successive generations. The results of the proposed algorithm were compared with traditional approaches, including Min-Min, Max-Min, Adaptive Incremental GA (AIGA), and Standard GA (SGA). Experimental evaluations conducted on Amazon EC2 demonstrated that the proposed approach outperforms these existing methods regarding task completion time, resource utilization, and convergence performance.  These results suggest that a method combining elitism and adaptive algorithms is effective in building a scalable and stable solution for work schedules in high-demand cloud environments.