Cost-efficient quantum cloud task offloading with quantum-inspired particle swarm optimization

Abstract

Quantum cloud computing (QCC) empowers application users (AUs) to manage computationally intensive and resource-demanding applications, particularly those involving intractable and complex problems. This research focuses on quantum task offloading (QTO) within the QCC environment. Successful QTO decisions require careful consideration of energy consumption, execution delay, service cost, and load balancing. Incorporating task urgency, the quantum task offloading problem (QTOP) is mathematically formulated to prioritize the execution of urgent tasks while satisfying budget and deadline constraints. It is shown that QTOP is a non-deterministic polynomial-time (NP-complete) problem. To address this challenge, a quantum-inspired particle swarm optimization (QPSO) algorithm is proposed. A novel quantum particle (QP) encoding scheme is introduced and decoded using a linear hashing approach to generate valid task offloading solutions. An effective fitness function is designed by integrating two penalty variables to eliminate infeasible solutions that violate resource and budget constraints. Extensive simulations are conducted to evaluate the performance of QPSO against several baseline algorithms, where QPSO consistently outperforms the others. Furthermore, the proposed cost model is benchmarked against existing models, demonstrating superior efficiency. Statistical analysis, as well as exploration and exploitation behavior analysis, further validate the robustness of the proposed method.