关键词: 威布尔分布；可靠性；参数估计；人工鱼群算法；耐久寿命

Abstract: In the process of component development, reliable estimation of component reliability is crucial. The Weibull distribution is widely used to simulate failure distributions, making parameter estimation of the Weibull distribution essential for component reliability prediction and analysis. To address this, this paper proposes a novel Weibull distribution estimation method based on an improved artificial fish swarm algorithm for vehicle component reliability assessment. Firstly, the paper theoretically constructs a maximum likelihood estimation model for the Weibull distribution and provides a detailed introduction to the relevant model. Subsequently, in light of the low efficiency and suboptimal quality inherent in the original artificial fish swarm algorithm, improvements are made by introducing Levy flight strategy and adaptive adjustment of perceptual distance. Lastly, the improved artificial fish swarm algorithm is employed for Monte Carlo simulation experiments to compare and analyze the fitting effects of Weibull distribution parameters. The results indicate that for Weibull distribution parameter estimation, the proposed improved artificial fish swarm algorithm more accurate parameter estimations compared to other metaheuristic algorithms, thereby providing scientifically effective guidance for automotive component development and reliability assessment.

Key words: Weibull distribution; reliability; parameter estimation; artificial fish swarm algorithm; endurance life