Mixed reliability importance-based solving algorithm design for the cost-constrained reliability optimization model

• Published in: Reliability engineering & system safety 2023-09, Vol.237, p.109363, Article 109363
• Main Authors: Liu, Mingli, Wang, Dan, Si, Shubin
• Format: Article
• Language: English
• Subjects: Joint reliability importance; Mixed reliability importance; Optimization algorithm; Reliability optimization model
• ISSN: 0951-8320; 1879-0836
• DOI: 10.1016/j.ress.2023.109363

•A novel cost-constrained reliability importance is proposed.•Mechanisms of importance to guide the design of optimization rules are explored.•A solving algorithm integrating the advantages of importance and GA is developed. In the field of reliability engineering, importance measures are widely used to prioritize components within a system and facilitate the improvement of system performance. However, current multi-component importance measures, such as joint reliability importances (JRIs) and their extensions, do not fully account for the potential impact of limited resource constraints, which can significantly impede efforts to improve system reliability. To address this issue, this paper proposes a novel JRI of two components for the cost-constrained reliability optimization model (ROM), which incorporates constraint factors into the JRI calculation. This new JRI can be used to evaluate the interaction effect of two components on system reliability under cost constraints. Subsequently, a cost-constrained, ROM-based, mixed reliability importance (CRMRI) is introduced by integrating the features of single-component importance measures with the newly devised JRI. Given equivalent costs for improving each component, the CRMRI approach can identify the two components whose simultaneous improvement contributes the most to enhancing system reliability. Lastly, we develop a CRMRI-based genetic algorithm (CRMGA) to solve the cost-constrained ROM. Experimental results on systems of various scales demonstrate that CRMGA can produce superior solutions with faster convergence speed, enhanced robustness, and higher efficiency compared to other optimization algorithms.