Generalized Linear Modeling for Assessment of a Performance-Based Logistics Strategy (PBL) on Demand/Cost Reductions in an Aging Complex System

• Activity Number: 573
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Mark Carpenter*+ and Wesley Randall
• Companies: Auburn University and Auburn University
• Keywords: discrete valued time series ; non-homogeneous Poisson ; mixture modeling ; Bayesian analysis ; performance based logistics

Abstract

This paper provides a methodology to investigate a hypothesized link between spare part consumption, reliability improvement, cost reductions, increases in mean time between failures (MTBF), and a Performance Based Logistics strategy. More specifically we investigate the assertion that a 5 year firm fixed price (FFP) PBL contract created an incentive for prime contractor to invest in repair process and reliability improvements, which resulted in specific cost reductions through efficiency gains and cost reduction in maintenance, repair, and overhaul (MRO) type activities. Assuming that the observed demands for each part or collection of parts over time are from a non-homogeneous Poisson process, we investigate short- and long-run trends in demand over time, using generalized Poisson regression, modeling time series of counts (discrete valued time series), Bayesian models, hierarchical and discrete mixture modeling.