A Multi-Objective Archimedes Optimizer for VM Selection in Healthcare Hybrid Cloud-IoT Systems
Healthcare applications deployed over hybrid Cloud-IoT infrastructures must satisfy strict latency and reliability requirements while remaining energy- and cost-efficient. In such environments, virtual machine (VM) selection becomes a challenging multi-objective optimization problem involving conflicting criteria. This paper formulates healthcare-aware VM selection as a discrete multi-objective task-to-VM assignment problem and proposes MO-AOA, a multi-objective extension of the Archimedes Optimization Algorithm. The proposed method combines buoyancy-inspired search dynamics with Pareto dominance ranking, crowding-distance diversity preservation, an external archive of non-dominated solutions, and a repair mechanism for feasible discrete allocations. MO-AOA is evaluated using both CloudSim Plus simulations and real workload traces derived from the Google Borg dataset, and is compared with NSGA-II, PSO, and ACO. Experimental results show that MO-AOA achieves up to 18% latency reduction, 22% energy savings, and 25% higher SLA compliance, while also improving Pareto-front quality. These results demonstrate that MO-AOA is an effective optimization framework for healthcare-aware resource orchestration in hybrid Cloud-IoT systems.