A Chaotic Inertia Weight TLBO Applied to Troubleshooting optimization Problems

摘要

When a failure event occurs in a complex system with multiple interconnected components, identifying the faulty component can be a challenging task. A troubleshooting optimization problem consists in finding the sequence of activities that must be followed to find the faulty component and fix the system with minimum expected cost of repair (ECR). Troubleshooting optimization can be modeled as a combinatorial optimization problem, and different algorithms have been proposed to solve it. This paper proposes a chaotic inertia weight Teaching-Learning Based optimization (TLBO) algorithm for the troubleshooting optimization problem. A chaotic sequence is used to update the inertia weight used in each iteration of TLBO. Numerical experiments using three troubleshooting models and nine chaotic maps are conducted to evaluate the performance of the proposed algorithm. The standard TLBO algorithm is also considered in the experiments to establish a reference baseline. The results showed that the proposed model presented a better performance in terms of average ECR, when compared with the standard TLBO algorithm.