Replacement Policy of Systems with Dependent Components via Integration of Dynamic Programming and Simulated Annealing

TL;DR

This paper proposes a novel replacement policy for dependent multi-component systems by integrating dynamic programming and simulated annealing, improving replacement decisions based on past data and deterioration rates.

Contribution

It introduces a new method combining dynamic programming and simulated annealing to optimize replacement policies in systems with dependent components.

Findings

The proposed method outperforms traditional limit replacement approaches.

Simulation results demonstrate the effectiveness of the integrated approach.

The method accurately estimates deterioration rates from past replacement data.

Abstract

In a dependent multi-component system, increasing the deterioration of a part leads to the increased deterioration rate of other parts as well. In these systems, a deterioration limit is usually pre-determined for each part and the considered part is replaced while reaching this limit. In this paper, replacement conditions of these parts were examined according to the replacement times in the past. Using dynamic programming, for every deterioration rate of part 1, there is a deterioration limit for part 2, after which either part 2 or both parts should be replaced. The only available system data are the replacement time of the parts in the past according to the replacement policy at the time of reaching deterioration limit. Therefore, simulated annealing optimization method was used for estimating deterioration rates. Finally, two examples were presented for comparing the proposed method with the special limit replacement method, which showed the significance superiority of the former.