Production and Rework: Separated or Combined?, Analytical and Simulation Models in the Frame of Reverse Logistics-Edición Única

Abstract

Reverse Logistics has been stretching out worldwide, involving all the layers of supply chains in various industry sectors. While some actors in the chain have been forced to take products back, others have proactively done so, attracted by the value of used products. When customer demand is given and we have no other choice than accomplish it, we have to look for the most economical way to meet that demand, keeping in mind both quality and delivery within the customer expected time. In a framework of Reverse Logistics, under a scheme of production and rework, one key factor to deal with is to decide whether to carry out both production and rework as separated or combined. This thesis report examines behavior of operational cost over a mixture of production line’s configurations mainly differentiated by carrying out rework using an off-line repair station or doing it at on-line way, and also by sharing tooling or work force, and having a restriction where the average time an item spends in the system should be less than certain amount of time. Parameters such as percentage of good items, production rate (operators; cross-functional workers), and rework/repair rate (repair men, cross-functional workers) were varied to evaluate system performance metrics. These parameters showed to be the most sensible ones to compare behavior of production/repair schemes carried out in a separated (off-line) or in a combined (on-line) configuration. Report presents five scenarios where three of them were solved using an analytical method based on a markovian open network with birth and death queues with infinite capacity and using an M/M/s queueing system to evaluate key performance metrics of the workstation, where the main difference among scenarios is how rework is carried out (off-line vs on-line). Other two remain scenarios were solved via simulation using Promodel where one scenario distinguishes from each other by what they share: tools or workforce. No matter what kind of method is used, system performance metrics to deal with are operational cost and the average time spent by a customer (or part) in the whole system, as main restriction.