A DATA-DRIVEN MULTI-OBJECTIVE PRODUCTION SCHEDULING FRAMEWORK FOR BATTERY PACK MANUFACTURING UNDER MATERIAL ARRIVAL UNCERTAINTY
Keywords:
Battery pack manufacturing, Conformalized quantile regression, Flexible flow shop, Material uncertainty, Multi-objective scheduling, NSGA-II, CVaRAbstract
Material-arrival uncertainty can invalidate otherwise efficient battery-pack production schedules because cells, battery-management systems, enclosures, and cooling components become available at different assembly stages. This paper presents a reproducible data-to-decision framework that connects leakage-controlled arrival-delay learning with four-objective flexible-flow-shop scheduling. Median and 0.90-quantile histogram gradient-boosting models are trained on 180,519 public DataCo supply-chain records using a chronological 70/15/15 split. A split-conformal correction and stratified residual resampling generate scenario-dependent material availability without using post-arrival variables. The manufacturing layer is a transparent synthetic battery-pack benchmark with six assembly stages, two alternative machines per stage, three product families, sequence-dependent setups, due-date weights, and machine-energy tradeoffs. A calibrated quantile-residual NSGA-II (CQR-RNSGA-II) minimizes expected makespan, expected weighted tardiness, energy, and the 0.90 conditional value-at-risk of material-induced waiting. On the held-out logistics period, the learned 0.90 quantile reduces pinball loss by 32.1% relative to a global 90th-percentile buffer and attains 95.9% empirical upper coverage; the conformal correction is exactly zero because the raw discrete-target quantile is already conservative. Across nine independently seeded scheduling instances, the proposed method is best on the normalized four-objective compromise score for 20- and 30-job cases and essentially tied for 12 jobs. Relative to risk-neutral NSGA-II, it reduces material-wait CVaR by 12.1% at 20 jobs and 26.4% at 30 jobs, while revealing a nontrivial risk–tardiness tradeoff. All numerical claims are generated by the supplied code; no proprietary battery-factory data are asserted.References
[1] Kwade A, Haselrieder W, Leithoff R, et al. Current status and challenges for automotive battery production technologies. Nature Energy, 2018, 3: 290–300. DOI: 10.1038/s41560-018-0130-3.
[2] Kölmel A, Sauer A, Lanza G. Quality-oriented production planning of battery assembly systems for electric mobility. Procedia CIRP, 2014, 23: 149–154. DOI: 10.1016/j.procir.2014.10.075.
[3] VDMA Battery Production and PEM of RWTH Aachen University. Battery Module and Pack Assembly Process. 3rd ed. Frankfurt, Germany, 2018.
[4] Kampker A, Heimes H H, Offermanns C, et al. Framework and classification of battery system architectures. World Electric Vehicle Journal, 2023, 14(4): 88. DOI: 10.3390/wevj14040088.
[5] Pinedo M L. Scheduling: Theory, Algorithms, and Systems. 5th ed. Cham, Switzerland: Springer, 2016. DOI: 10.1007/978-3-319-26580-3.
[6] Jain A S, Meeran S. Deterministic job-shop scheduling: Past, present and future. European Journal of Operational Research, 1999, 113(2): 390–434. DOI: 10.1016/S0377-2217(98)00113-1.
[7] Ouelhadj D, Petrovic S. A survey of dynamic scheduling in manufacturing systems. Journal of Scheduling, 2009, 12: 417–431. DOI: 10.1007/s10951-008-0090-8.
[8] Vieira G E, Herrmann J W, Lin E. Rescheduling manufacturing systems: A framework of strategies, policies, and methods. Journal of Scheduling, 2003, 6(1): 39–62. DOI: 10.1023/A:1022235519958.
[9] Bertsimas D, Brown D B, Caramanis C. Theory and applications of robust optimization. SIAM Review, 2011, 53(3): 464–501. DOI: 10.1137/080734510.
[10] Ben-Tal A, El Ghaoui L, Nemirovski A. Robust Optimization. Princeton, NJ, USA: Princeton University Press, 2009. DOI: 10.1515/9781400831050.
[11] Rockafellar R T, Uryasev S. Optimization of conditional value-at-risk. Journal of Risk, 2000, 2(3): 21–41. DOI: 10.21314/JOR.2000.038.
[12] Módos I, Šůcha P, Hanzálek Z. Algorithms for robust production scheduling with energy consumption limits. Computers & Industrial Engineering, 2017, 112: 391–408. DOI: 10.1016/j.cie.2017.08.011.
[13] Mouzon G, Yildirim M B, Twomey J. Operational methods for minimization of energy consumption of manufacturing equipment. International Journal of Production Research, 2007, 45(18–19): 4247–4271. DOI: 10.1080/00207540701450013.
[14] May G, Barletta I, Stahl B, et al. Energy management in production: A novel method to develop key performance indicators for improving energy efficiency. Applied Energy, 2015, 149: 46–61. DOI: 10.1016/j.apenergy.2015.03.065.
[15] Dai M, Tang D, Giret A, et al. Multi-objective optimization for energy-efficient flexible job shop scheduling problem with transportation constraints. Robotics and Computer-Integrated Manufacturing, 2019, 59: 143–157. DOI: 10.1016/j.rcim.2019.04.006.
[16] Brandimarte P. Routing and scheduling in a flexible job shop by tabu search. Annals of Operations Research, 1993, 41: 157–183. DOI: 10.1007/BF02023073.
[17] Pezzella F, Morganti G, Ciaschetti G. A genetic algorithm for the flexible job-shop scheduling problem. Computers & Operations Research, 2008, 35(10): 3202–3212. DOI: 10.1016/j.cor.2007.02.014.
[18] Gao J, Sun L, Gen M. A hybrid genetic and variable neighborhood descent algorithm for flexible job shop scheduling problems. Computers & Operations Research, 2008, 35(9): 2892–2907. DOI: 10.1016/j.cor.2007.01.001.
[19] Taillard E. Benchmarks for basic scheduling problems. European Journal of Operational Research, 1993, 64(2): 278–285. DOI: 10.1016/0377-2217(93)90182-M.
[20] Deb K, Pratap A, Agarwal S, et al. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation, 2002, 6(2): 182–197. DOI: 10.1109/4235.996017.
[21] Romano Y, Patterson E, Candès E J. Conformalized quantile regression. In: Advances in Neural Information Processing Systems, 2019.
[22] Kuleshov V, Fenner N, Ermon S. Accurate uncertainties for deep learning using calibrated regression. In: Proceedings of the 35th International Conference on Machine Learning, PMLR, 2018, 80: 2796–2804.
[23] Pedregosa F. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 2011, 12: 2825–2830.
[24] Li Y, Carabelli S, Fadda E, et al. Machine learning and optimization for production rescheduling in Industry 4.0. International Journal of Advanced Manufacturing Technology, 2020, 110: 2445–2463. DOI: 10.1007/s00170-020-05850-5.
[25] Camur M C, Ravi S K, Saleh S. Enhancing supply chain resilience: A machine learning approach for predicting product availability dates under disruption. arXiv:2304.14902, 2023. DOI: 10.48550/arXiv.2304.14902.
[26] Constante F, Silva F, Pereira A. DataCo SMART SUPPLY CHAIN FOR BIG DATA ANALYSIS. Mendeley Data, Version 5, 2019. DOI: 10.17632/8gx2fvg2k6.5.
[27] Li X, Guo X, Tang H, et al. Survey of integrated flexible job shop scheduling problems. Computers & Industrial Engineering, 2022, 174: 108786. DOI: 10.1016/j.cie.2022.108786.
[28] Zhang C, Song W, Cao Z, et al. Learning to dispatch for job shop scheduling via deep reinforcement learning. In: Advances in Neural Information Processing Systems 33, 2020.
[29] Liu R, Piplani R, Toro C. Deep reinforcement learning for dynamic scheduling of a flexible job shop. International Journal of Production Research, 2022, 60(13): 4049–4069. DOI: 10.1080/00207543.2022.2058432.
[30] Zhang L, Feng Y, Xiao Q, et al. Deep reinforcement learning for dynamic flexible job shop scheduling problem considering variable processing times. Journal of Manufacturing Systems, 2023, 71: 257–273. DOI: 10.1016/j.jmsy.2023.09.009.