Optimizing Flexibility and Commitments: Partially-Adaptive Multi-Stage Matrix Assembly Layout Design Problem
Abstract
Matrix assembly layouts provide manufacturing flexibility under product variety and demand volatility through locally reconfigurable production stations. Periodic redesign exploits this flexibility, yet existing models allow only annual revisions. Finer calendars expose a flexibility-commitment trade-off: advance layout commitments support coordination, supplier agreements, and discounts. We introduce the layout revision timing problem, jointly choosing revision points and committed layout sequences under demand uncertainty while production adapts within those layouts. We formulate it as a partially adaptive multistage stochastic program with two complementary extensions. Our cost-based extension places the economic value of advance commitment in the objective rather than prescribing revision budgets; our scenario-adaptive extension adapts revision timing to revealed information. For station-use minimization, we establish a cost decomposition across commitment intervals, yielding an exact dynamic program that reuses station subproblem solutions across model variants and parameter settings. As a more general alternative, we develop an integer L-shaped method strengthened by monotonicity-based cuts for budgeted formulations and enhanced cuts exploiting this decomposition. On automotive instances, including scenario trees with nearly 10,000 nodes, our combined model's average savings in expected station-use cost at the base-case discount reach 4.99% versus full adaptivity and 28.65% versus full commitment. Scenario adaptivity is most valuable when demand paths diverge and discounts justify commitment on some paths but not others, as in electrification-driven scenarios.
Related
- A study on a tightening method for optimizing bolt preload distribution in high-precision assembly of curvic couplings for aerospace applications
- Partially Adaptive Robust Optimization for Uncertain Fluid Processing Networks
- Sequential Certification of Partially Identified Product Targets: Information Limits and Adaptive Sampling
- Optimal Service Commitments in Traveling Salesman Problems with Stochastic Demand
- Length spectrum rigidity and flexibility of spheres of revolution with one equator
BibTeX
@misc{bayraktar2026optimizing,
title = {{Optimizing Flexibility and Commitments: Partially-Adaptive Multi-Stage Matrix Assembly Layout Design Problem}},
author = {Baturhan Bayraktar and Merve Bodur and Martin Grunow and Rainer Kolisch},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.27401},
eprint = {2610.27401},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.27401},
}
RIS
TY - UNPB
TI - Optimizing Flexibility and Commitments: Partially-Adaptive Multi-Stage Matrix Assembly Layout Design Problem
AU - Baturhan Bayraktar
AU - Merve Bodur
AU - Martin Grunow
AU - Rainer Kolisch
PY - 2026
DA - 2026/10/19
PB - arXiv
DO - 10.48550/arxiv.2610.27401
UR - https://arxiv.org/abs/2610.27401
AB - Matrix assembly layouts provide manufacturing flexibility under product variety and demand volatility through locally reconfigurable production stations. Periodic redesign exploits this flexibility, yet existing models allow only annual revisions. Finer calendars expose a flexibility-commitment trade-off: advance layout commitments support coordination, supplier agreements, and discounts. We introduce the layout revision timing problem, jointly choosing revision points and committed layout sequences under demand uncertainty while production adapts within those layouts. We formulate it as a partially adaptive multistage stochastic program with two complementary extensions. Our cost-based extension places the economic value of advance commitment in the objective rather than prescribing revision budgets; our scenario-adaptive extension adapts revision timing to revealed information. For station-use minimization, we establish a cost decomposition across commitment intervals, yielding an exact dynamic program that reuses station subproblem solutions across model variants and parameter settings. As a more general alternative, we develop an integer L-shaped method strengthened by monotonicity-based cuts for budgeted formulations and enhanced cuts exploiting this decomposition. On automotive instances, including scenario trees with nearly 10,000 nodes, our combined model's average savings in expected station-use cost at the base-case discount reach 4.99% versus full adaptivity and 28.65% versus full commitment. Scenario adaptivity is most valuable when demand paths diverge and discounts justify commitment on some paths but not others, as in electrification-driven scenarios.
ER -