172. The Value of Reliability in Modular Production Network Design
Invited abstract in session FA-10: Network Design, stream Supply Chain Management and Production.
Friday, 8:45-10:15Room: H16
Authors (first author is the speaker)
| 1. | Lea Raubuch
|
| Junior Professorship of Management Science, TU Dresden | |
| 2. | Stefan Lier
|
| FH Südwestfalen | |
| 3. | Tristan Becker
|
| TU Dresden |
Abstract
Technical systems are subject to disruptions, and production units may consequently experience failures over time. In this context, a key advantage of modular production systems is their ability to compensate for disruptions by relocating the remaining operational units between established facilities in the short term. However, existing approaches to modular production network design have neglected the impact of such disruptions. To address this gap, we extend the Capacitated Reliable Facility Location Problem (CRFLP) to the context of modular production. We introduce two mathematical formulations: a facility-indexed and a module-indexed approach. The module-indexed formulation generalizes the CRFLP and represents the scenario space with a significantly lower number of scenarios compared to the standard model, provided an upper bound is imposed on the number of open facilities. Both formulations minimize total network costs while ensuring demand fulfillment, allowing for external procurement at a penalty cost. We propose a two-stage stochastic optimization framework. In the first stage, facility locations and modular production unit allocations are determined. In the second stage, production units may be relocated among open facilities, and customer demand may be reassigned accordingly. Given the combinatorial complexity and exponential growth of the scenario space, we apply Sample Average Approximation for our computational experiments. Our analysis evaluates the impact of relocation flexibility and reliability on modular production network cost. The results indicate that incorporating stochasticity into the selection of modular production facilities improves the performance of the network for moderate and high failure probabilities.
Keywords
- Stochastic Programming
- Network Design
- Location
Status: accepted
Back to the list of papers