Identifier-Driven Resource Orchestration With Quantum Computing for Differentiated Services in IoT-MMEC Networks

The prevalence of Artificial Intelligence and Multi-access Mobile Edge Computing (MMEC) technologies has laid a solid foundation for next-generation Internet of Things (IoT) applications, e.g., industrial automation and smart healthcare fields. However, the explosive data and ubiquitous services significantly exacerbate the consumption and unreliability of constrained edge resources. In this paper, we investigate a joint resource orchestration problem for IoT-MMEC networks with different service performances. We first develop an identifier space mapping model to represent the matching relationship between access attributes and space resources, which respectively denote the computing task description and the set of allocated resources. To obtain an optimal resource partition policy for dependable and low-budget auxiliary calculation, we formulate a mixed-integer nonlinear programming problem. Then, we devise an identifier-driven resource orchestration scheme, which decouples the problem into computation offloading and resource allocation subproblems. Based on the expected utility function theory and access attributes, we apply a mixed deep neural network inference model to infer the offloading location, for realizing the resource supply-demand balance. To derive the optimal resource allocation scheme, we exploit the quantum genetic algorithm and multi-path offloading factor, which can explore a large search space to find potential solutions while exploiting the best solutions. Finally, the experimental simulations validate our theoretical analysis, and the results indicate that the proposed scheme can achieve lower consumption and enhance offloading reliability.