High-dimensional stochastic control models for newsvendor problems and deep learning resolution. (English) Zbl 07898272
Summary: This paper studies continuous-time models for newsvendor problems with dynamic replenishment, financial hedging and Stackelberg competition. These factors are considered simultaneously and the high-dimensional stochastic control models are established. High-dimensional Hamilton-Jacobi-Bellman (HJB) equations are derived for the value functions. To circumvent the curse of dimensionality, a deep learning algorithm is proposed to solve the HJB equations. A projection is introduced in the algorithm to avoid the gradient explosion during the training phase. The deep learning algorithm is implemented for HJB equations derived from the newsvendor models with dimensions up to six. Numerical outcomes validate the algorithm’s accuracy and demonstrate that the high-dimensional stochastic control models can successfully mitigate the risk.
MSC:
| 90B05 | Inventory, storage, reservoirs |
| 91A80 | Applications of game theory |
| 90B06 | Transportation, logistics and supply chain management |
| 68T05 | Learning and adaptive systems in artificial intelligence |
Keywords:
supply chain management; newsvendor models; stochastic control; dynamic replenishment; financial hedging; Stackelberg game; deep learningSoftware:
DeepXDEReferences:
| [1] | Bach, F., Breaking the curse of dimensionality with convex neural networks, The Journal of Machine Learning Research, 18, 629-681, 2017 · Zbl 1433.68390 |
| [2] | Baydin, AG; Pearlmutter, BA; Radul, AA; Siskind, JM, Automatic differentiation in machine learning: A survey, Journal of Machine Learning Research, 18, 1-43, 2018 · Zbl 06982909 |
| [3] | Beck, CEW; Jentzen, A., Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations, Journal of Nonlinear Science, 29, 1563-1619, 2019 · Zbl 1442.91116 · doi:10.1007/s00332-018-9525-3 |
| [4] | Berling, P., Real options valuation principle in the multi-period base-stock problem, Omega, 36, 1086-1095, 2008 · doi:10.1016/j.omega.2006.05.007 |
| [5] | Blomvall, J.; Hagenbjörk, J., Reducing transaction costs for interest rate risk hedging with stochastic programming, European Journal of Operational Research, 302, 1282-1293, 2022 · Zbl 1507.91217 · doi:10.1016/j.ejor.2022.02.004 |
| [6] | Chen, X.; Sim, M.; Simchi-Levi, D.; Sun, P., Risk aversion in inventory management, Operations Research, 55, 828-842, 2007 · Zbl 1167.90317 · doi:10.1287/opre.1070.0429 |
| [7] | Chen, L.; Song, JS; Zhang, Y., Serial inventory systems with Markov-modulated demand: derivative bounds, asymptotic analysis, and insights, Operations Research, 65, 1231-1249, 2017 · Zbl 1380.90012 · doi:10.1287/opre.2017.1615 |
| [8] | Chen, X.; Zhang, H.; Zhang, M.; Chen, J., Optimal decisions in a retailer Stackelberg supply chain, International Journal of Production Economics, 187, 260-270, 2017 · doi:10.1016/j.ijpe.2017.03.002 |
| [9] | Choi, S.; Ruszczyński, A., A multi-product risk-averse newsvendor with exponential utility function, European Journal of Operational Research, 214, 78-84, 2011 · Zbl 1218.91053 · doi:10.1016/j.ejor.2011.04.005 |
| [10] | Duchi, J.; Hazan, E.; Singer, Y., Adaptive subgradient methods for online learning and stochastic optimization, Journal of Machine Learning Research, 12, 2121-2159, 2011 · Zbl 1280.68164 |
| [11] | Ding, Q.; Dong, L.; Kouvelis, P., On the integration of production and financial hedging decisions in global markets, Operations Research, 55, 470-489, 2007 · Zbl 1167.91366 · doi:10.1287/opre.1070.0364 |
| [12] | Yu, B., The Deep Ritz Method: A deep learning-based numerical algorithm for solving variational problems, Communications in Mathematics and Statistics, 1, 1-12, 2018 · Zbl 1392.35306 |
| [13] | Han, J.; Jentzen, A., Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations, Communications in Mathematics and Statistics, 5, 349-380, 2017 · Zbl 1382.65016 · doi:10.1007/s40304-017-0117-6 |
| [14] | Elbärchter, D.; Grohs, P.; Jentzen, A.; Schwab, C., DNN expression rate analysis of high-dimensional PDEs: Application to option pricing, Constructive Approximation, 3, 1-69, 2021 |
| [15] | Gallego, G.; Scheller-Wolf, A., Capacitated inventory problems with fixed order costs: Some optimal policy structure, European Journal of Operational Research, 126, 603-613, 2000 · Zbl 0982.90002 · doi:10.1016/S0377-2217(99)00314-8 |
| [16] | Germain, M.; Pham, H.; Warin, X., Approximation error analysis of some deep backward schemes for nonlinear PDEs, SIAM Journal on Scientific Computing, 44, A28-A56, 2022 · Zbl 1490.65231 · doi:10.1137/20M1355355 |
| [17] | Glock, CH; Rekik, Y.; Ries, JM, A coordination mechanism for supply chains with capacity expansions and order-dependent lead times, European Journal of Operational Research, 285, 247-262, 2020 · Zbl 1441.90005 · doi:10.1016/j.ejor.2020.01.048 |
| [18] | Heckmann, I.; Comes, T.; Nickel, S., A critical review on supply chain risk-Definition, measure and modeling, Omega, 52, 119-132, 2015 · doi:10.1016/j.omega.2014.10.004 |
| [19] | Heston, SL, A closed-form solution for options with stochastic volatility with applications to bond and currency options, The Review of Financial Studies, 6, 327-343, 1993 · Zbl 1384.35131 · doi:10.1093/rfs/6.2.327 |
| [20] | Huré, C.; Pham, H.; Warin, X., Deep backward schemes for high-dimensional nonlinear PDEs, Mathematics of Computation, 89, 1547-1579, 2020 · Zbl 1440.60063 · doi:10.1090/mcom/3514 |
| [21] | Jian, H.; Farzad, A.; Dmitry, I.; Hamed, J., A real-option approach to mitigate disruption risk in the supply chain, Omega, 88, 133-149, 2018 |
| [22] | Johansson, B., Keviczky, T., Johansson, M., & Johansson, K. H. (2008). Subgradient methods and consensus algorithms for solving convex optimization problems. In 2008 47th IEEE Conference on Decision and Control, pp 4185-4190. |
| [23] | Kogan, K.; Lou, S., Multi-stage newsboy problem: A dynamic model, European Journal of Operational Research, 149, 448-458, 2003 · Zbl 1033.90003 · doi:10.1016/S0377-2217(02)00450-2 |
| [24] | Lu, L.; Meng, X.; Mao, Z.; Karniadakis, GE, Deepxde: A deep learning library for solving differential equations, SIAM Review, 63, 208-228, 2021 · Zbl 1459.65002 · doi:10.1137/19M1274067 |
| [25] | Ni, J.; Chu, LK; Wu, F.; Sculli, D.; Shi, Y., A multi-stage financial hedging approach for the procurement of manufacturing materials, European Journal of Operational Research, 221, 424-431, 2012 · Zbl 1253.91201 · doi:10.1016/j.ejor.2012.03.031 |
| [26] | Øksendal, B.; Sandal, L.; Ubøe, J., Stochastic Stackelberg equilibria with applications to time-dependent newsvendor models, Journal of Economic Dynamics and Control, 37, 1284-1299, 2013 · Zbl 1402.90010 · doi:10.1016/j.jedc.2013.02.010 |
| [27] | Ou, J.; Feng, J., Production lot-sizing with dynamic capacity adjustment, European Journal of Operational Research, 272, 261-269, 2019 · Zbl 1403.90044 · doi:10.1016/j.ejor.2018.06.030 |
| [28] | Petersen, P.; Voigtlaender, F., Optimal approximation of piecewise smooth functions using deep ReLU neural networks, Neural Networks, 108, 296-330, 2018 · Zbl 1434.68516 · doi:10.1016/j.neunet.2018.08.019 |
| [29] | Qin, Y.; Wang, R.; Vakharia, AJ; Chen, Y.; Seref, M., The newsvendor problem: Review and directions for future research, European Journal of Operational Research, 213, 361-374, 2011 · Zbl 1215.90005 · doi:10.1016/j.ejor.2010.11.024 |
| [30] | Raissi, M.; Perdikaris, P.; Karniadakis, GE, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics, 378, 686-707, 2018 · Zbl 1415.68175 · doi:10.1016/j.jcp.2018.10.045 |
| [31] | Sirignano, J.; Spiliopoulos, K., DGM: A deep learning algorithm for solving partial differential, equations, Journal of Computational Physics, 375, 1339-1364, 2018 · Zbl 1416.65394 · doi:10.1016/j.jcp.2018.08.029 |
| [32] | van der Meer, R.; Oosterlee, CW; Borovykh, A., Optimally weighted loss functions for solving pdes with neural networks, Journal of Computational and Applied Mathematics, 405, 2022 · Zbl 1518.65143 · doi:10.1016/j.cam.2021.113887 |
| [33] | Vishal, G.; Sridhar, S., Hedging inventory risk through market instruments, Manufacturing and Service Operations Management, 7, 103-120, 2005 · doi:10.1287/msom.1040.0061 |
| [34] | Wang, H.; Chen, B.; Yan, H., Optimal inventory decisions in a multiperiod newsvendor problem with partially observed Markovian supply capacities, European Journal of Operational Research, 202, 502-517, 2010 · Zbl 1175.90037 · doi:10.1016/j.ejor.2009.05.042 |
| [35] | Wang, M.; Guo, X.; Wang, S., Financial hedging in two-stage sustainable commodity supply chains, European Journal of Operational Research, 303, 803-818, 2022 · Zbl 1524.90088 · doi:10.1016/j.ejor.2022.02.048 |
| [36] | Wein, LM; Gallien, J., A smart market for industrial procurement with capacity constraints, Management Science, 51, 76-91, 2005 · Zbl 1232.91300 · doi:10.1287/mnsc.1040.0230 |
| [37] | Xie, S.; Li, Z.; Wang, S., Continuous-time portfolio selection with liability: Mean-variance model and stochastic LQ approach, Insurance: Mathematics and Economics, 42, 943-953, 2008 · Zbl 1141.91474 |
| [38] | Yong, J.; Zhou, XY, Stochastic Controls: Hamiltonian Systems and HJB Equations, 1999, Cham: Springer, Cham · Zbl 0943.93002 · doi:10.1007/978-1-4612-1466-3 |
| [39] | Zhang, J.; Qi, L.; Tong, S., Dynamic contract under quick response in a supply chain with information asymmetry, Production and Operations Management, 30, 1273-1289, 2020 · doi:10.1111/poms.13321 |
| [40] | Zhao, L.; Huchzermeier, A., Integrated operational and financial hedging with capacity reshoring, European Journal of Operational Research, 260, 557-570, 2016 · Zbl 1403.90068 · doi:10.1016/j.ejor.2016.12.036 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
