Remanufacturing cost analysis under uncertain core quality and return conditions: extreme and non-extreme scenarios

Abstract

Uncertainties in core quality condition, return quantity and timing can propagate and accumulate in process cost and complicate cost assessments. However, regardless of cost assessment complexities, accurate cost models are required for successful remanufacturing operation management. In this paper, joint effects of core quality condition, return quantity, and timing on remanufacturing cost under normal and extreme return conditions is analyzed. To conduct this analysis, a novel multivariate stochastic model called Stochastic Cost of Remanufacturing Model (SCoRM) is developed. In building SCoRM, a Hybrid Pareto Distribution (HPD), Bernoulli process, and a polynomial cost function are employed. It is discussed that core return process can be characterized as a Discrete Time Markov Chain (DTMC). In a case study, SCoRM is applied to assess remanufacturing costs of steam traps of a chemical complex. Its accuracy analyzed and variations of SCoRM in predictive tasks assessed by bootstrapping technique. Through this variation analysis the best and worst cost scenarios determined. Finally, to generate comparative insights regarding predictive performance of SCoRM, the model is compared to artificial neural network, support vector machine, generalized additive model, and random forest algorithms. Results indicate that SCoRM can be efficiently utilized to analyze remanufacturing cost. Keywords: Remanufacturing, extreme value theory, hybrid Pareto distribution, stochastic model.