Equipment replacement decisions based on life-cycle cost analysis

The Texas Department of Transportation (TxDOT) maintains an active fleet of approximately 17,000 units, annually replacing about ten percent. It also maintains a comprehensive data base that includes data relevant to life-cycle cost analysis. Ideally, equipment units whose life cycle costs have been increasing longer and/or at a faster rate should have higher replacement priority than those still amortizing. However, fleet managers could not routinely use life-cycle costs as replacement criteria because there was no reliable method to automatically prioritize thousands of life-cycle cost histories. This paper summarizes three innovative concepts in fleet management developed and implemented by the authors: (1) Bayesian analysis of life-cycle costs. This project used a SAS algorithm to adjust a trend component to each life-cycle cost time series; (2) Trendscore, a concept that enables a computer to mimic replacement decisions made by a person visually inspecting life-cycle cost Bayesian trends; and (3) Multi-attribute priority ranking, an algorithm that ranks replacement priorities based on the condition of the units relative to the rest of the fleet within its class or group, instead of the current practice of slating a unit for replacement when it reaches some threshold value. The replacement budget can be matched to the units with top replacement priority, making budget shortfalls clearly visible. These concepts were implemented through a menu-driven decision aid software that calculates life cycle cost profiles and trends, and generates replacement priority lists based on the concepts described above.