Modelling Approaches to Optimise Spares in Multi-echelon Systems

Spare parts optimisation requires an integrated approach that starts with the removal of factors that create noise: bad coding, resulting in parts proliferation; lack of classification and thus of focus; poor network practices (uncoupled warehouses, poor relation with key suppliers); and poor data integrity (non-centralised, non-real time data). Emphasis is also placed on improving forecasts and the application of multi-echelon inventory models for repairable items. The METRIC model is then applied to a hypothetical and two real cases, an underground system and a mobile telephone company in Venezuela.