On the Application of a Multi-Objective Genetic Algorithm to the LORA-Spares Problem

Abstract

Level of repair analysis (LORA) is often defined as the problem of determining whether a component should be repaired or discarded upon its failure, and the location in the repair network to do such work. A related problem is the determination of the optimal number of spares for a given piece of equipment. Although LORA and spare provisioning are interdependent, they are seldom solved simultaneously due to the complex nature of the relationships between spare levels and system availability. In this paper, we propose to apply a multi-objective genetic algorithm (specifically the Non-dominated Sorting Genetic Algorithm II) with optimization objectives of repair costs (e.g., spare parts, spares transportation, spares storage) and spare parts availability. The approach uses a Monte Carlo simulation to generate scenarios based on a dataset which includes the expected failures of the equipment and their associated probabilities. The objective functions are computed at each genetic algorithm generation based on the generated scenarios. An example that can be used for a trade-off analysis is provided.