Quantum-Inspired Scheduling Algorithms for Hybrid Cloud Workflows

Abstract

The rapid proliferation of hybrid cloud deployments—comprising on-premises private clouds integrated with one or more public cloud platforms—has revolutionized how organizations handle compute-intensive and data-driven workloads. In such environments, efficient scheduling of interdependent tasks is critical to achieving low latency, high resource utilization, and minimized operational costs. Conventional scheduling algorithms (e.g., First-Come-First-Serve, Round Robin, heuristic genetic algorithms) often struggle with the combinatorial explosion of possibilities and the stochastic fluctuations in resource availability inherent to hybrid clouds. Quantum computing, with its ability to explore vast solution spaces via phenomena like superposition and tunneling, presents a promising new paradigm for optimization. However, practical quantum hardware remains in its infancy. Quantum-inspired algorithms (QIAs) emulate key quantum behaviors on classical systems, offering a tractable approach to harnessing quantum advantages today. This study develops a Quantum-Inspired Genetic Algorithm (QIGA) tailored for scheduling Directed Acyclic Graph (DAG) workflows across heterogeneous hybrid cloud resources. We detail the encoding of scheduling decisions as “qubit” probability amplitudes, the application of quantum-rotation gates for guided search-space traversal, and a multiobjective fitness function balancing makespan, cost, and utilization. Through extensive simulations on both synthetic benchmarks and real-world workloads (scientific workflows, transactional pipelines, media rendering), we compare QIGA against FCFS, classical GA, and Particle Swarm Optimization (PSO). Statistical analyses over 100 workflow instances demonstrate that QIGA reduces average makespan by 20–30%, improves overall resource utilization by 8–12%, and achieves up to 15% energy savings relative to the best classical heuristic.