Research on Optimization Method for Maintenance and Repair Decision-Making of SPTN

Software-Defined Packet Transport Networks (SPTN) are crucial for modern infrastructure like 5G and data centers, offering efficiency and flexibility. However, their inherent complexity and dynamic service demands apply significant challenges for effective Operations and Maintenance (O&M). Specifically, optimizing maintenance decisions in SPTN environments requires balancing multiple objectives under various constraints. This paper addresses the SPTN maintenance decision problem by proposing a novel multi-objective optimization model. Based on health status assessment, this model aims to maximize maintenance utility while minimizing cost, incorporating multi-dimensional constraints such as personnel, resources, and structural dependencies. To efficiently solve this high-dimensional, complex problem, we introduce the Multi-Objective Covariance Matrix Adaptation Evolution Strategy (MO-CMA-ES) algorithm. Experimental results demonstrate that MO-CMA-ES significantly outperforms traditional multi-objective algorithms (e.g., NSGA-II, MOPSO) in terms of hypervolume and overall performance, generating a superior Pareto front. The proposed model, when combined with MO-CMA-ES, improves the utility-cost ratio by approximately 58% compared to traditional strategies, validating its effectiveness and the rationality of our approach.