Multi-objective balanced maintenance decision optimization of aviation equipment based on improved particle swarm optimization algorithm

Purpose This study aims to address the challenge of balancing reliability and cost during the wear phase of aviation equipment, overcoming the limitations of static age-reduction models and the potential for algorithms to converge to local optima. Design/methodology/approach Focusing specifically on aircraft landing gear hydraulic pumps, it proposes a maintenance optimization framework based on a reliability threshold. A dynamic dual-factor rollback degradation model is developed alongside a dual-objective optimization system aimed at simultaneously minimizing cost rate and maximizing average reliability. To enhance optimization efficiency and diversity, an improved particle swarm optimization (PSO) algorithm (IPSO) incorporating dynamic weights, hybrid mutation and a sequential ranking fitness function is employed. Findings Further analysis of the maintenance cycle and associated costs revealed that the reliability-oriented strategy, by intensifying early preventive measures and reducing the initial repair cycle to 139. 01 h, advanced maintenance timing and effectively slowed the increase in failure rates. In contrast, the economic strategy, with a longer initial repair cycle of 167. 18 h, faced greater challenges in managing failures during later stages. The reliability strategy reduces C₁ from 16. 6% to 13. 3% through enhanced preventive maintenance, effectively limiting failure rate growth and lowering long-term failure costs. While the economic strategy allocates 21. 9% to C₁ and C₄, the reliability strategy shifts cost allocation to C₂ and C₄, totaling 38. 1%, thereby emphasizing “pre-failure prevention” over post-failure handling. Overall, the reliability strategy aligns better with the actual equipment conditions discussed in this study. Research limitations/implications The case study focuses on the hydraulic pump of a typical aircraft landing gear. The developed model, optimization strategies, and conclusions are primarily applicable to aerospace components exhibiting similar “wear-type” failure characteristics. For other types of equipment, such as those dominated by random failures or for complex systems in different industries, the direct applicability of this method may be limited and would require targeted adjustments and validation. Additionally, the study demonstrates that the proposed approach can significantly reduce total lifecycle costs while ensuring higher average reliability. This suggests that maintenance of such aerospace components should shift from a fixed-interval approach to a predictive, dynamic maintenance model based on reliability thresholds. Practical implications The results verify the superiority and provide a dynamic maintenance method for aviation equipment maintenance decision-making. Social implications The critical aviation components with strict safety standards, like landing gear, adopt a “reliability-first” approach to enhance both safety and operational efficiency. This model offers theoretical guidance for the maintenance of aviation equipment belonging to the same series. Originality/value The improved IPSO algorithm effectively facilitates multi-objective optimization of the models and strategies presented in this paper, providing a stable and efficient tool for balancing maintenance reliability and cost in aerospace equipment. The identified optimal transition intervals for the economic dynamic strategy and for the reliability dynamic strategy offer practical guidance for establishing phase transition points in segmented maintenance planning under varying objectives. The dynamic dual-factor decremental model introduced herein transitions from a static maintenance strategy based on reliability thresholds to a dynamic approach incorporating phase transition intervals. This method outperforms traditional fixed-interval maintenance schemes employing static age-decrement models and enables self-iterative updates in maintenance decision-making.