FAULT DIAGNOSIS BASED ON NATURE-INSPIRED FEATURE SELECTION METHODS

Fault detection problems in industrial and computer fields have a major impact on the operating stability of any system. The number of inputs influences the analysis time and requires a considerable amount of memory. The use of meta-heuristics avoids exhaustive research in the solution space and represents an ideal solution to reduce the number of inputs in the fault detection process. In this article, we have proposed a set of improvements on algorithms inspired from the natural world, adapted to the problem of feature selection. Our major contribution lies in the proposal of a new algorithm based on the behavior of real ants and its application to accelerate the process of fault detection in three different systems. The obtained results demonstrate the effectiveness of the proposed algorithm