在庫存管理問題中，需求分配參數通常被假設為已知，然而在實務情境中，需求分配往往難以精確估計，因此，本研究探討當需求分配已知且參數未知時，決策者如何利用銷售數據來學習需求分配。由於庫存限制，需求資訊的觀察可能不完整，當庫存不足導致缺貨時，無法得知真實需求，這種情況稱為需求審查。相反地，當庫存充足時，雖然可 以觀察到真實需求，但會增加庫存成本。因此，本研究的核心問題在於如何在需求學習與庫存管理之間取得最佳平衡。傳統文獻大多假設需求為連續型且服從指數分配，然而，卜瓦松分配與常態分配為最符合實際的需求假設，基於此，本研究專注於針對離散型需求與連續型需求進行庫存管理優化。為了解決連續需求帶來的隨機問題，本研究引入了需求到達時間資訊，克服了需求審查情境下無法使用共軛機率分配的困難。此外，為了解決計算複雜度的問題，本研究提出了一套演算法，旨在提供一個近似最佳的庫存決策。研究結果顯示，演算法能有效收斂至參數已知下的最佳解，為學術界與實務管理者在面對需求分配參數未知的庫存管理挑戰時，提供了具體的解決方案。決策者可根據本研究的方法，在不完全需求資訊的情況下，達到庫存管理與需求學習的最佳平衡，從而減少庫存成本並避免缺貨風險。

In inventory management problems, demand distribution parameters are typically assumed to be known; however, in practical situations, demand distributions are often difficult to estimate accurately. Therefore, this study explores how decision-makers can learn the demand distribution using sales data when the demand distribution is known but its parameters are unknown. Due to inventory constraints, observations of demand information may be incomplete. When stockouts occur because of insufficient inventory, the true demand cannot be observed—this situation is referred to as demand censoring. Conversely, when inventory is sufficient, the true demand can be observed, but this increases inventory holding costs. Hence, the core issue of this study lies in how to achieve the optimal balance between demand learning and inventory management.Traditional literature mostly assumes demand to be continuous and exponentially distributed; however, Poisson and normal distributions are more realistic assumptions for demand. Based on this, the study focuses on optimising inventory management for both discrete and continuous demand. To address the stochastic challenges posed by continuous demand, this research introduces demand arrival time information, overcoming the difficulty of using conjugate probability distributions under demand censoring conditions. Furthermore, to tackle computational complexity, the study proposes an algorithm designed to provide near-optimal inventory decisions.The research findings indicate that the algorithm can effectively converge to the optimal solution when the parameters are known, providing a concrete solution for academia and practitioners facing inventory management challenges with unknown demand allocation parameters. Decision-makers can apply the method proposed in this study to achieve the optimal balance between inventory management and demand learning under conditions of incomplete demand information, thereby reducing inventory costs and avoiding stockout risks.

Anupindi, R., Dada, M., & Gupta, S. (1998). Estimation of consumer demand with stock-out based substitution: An application to vending machine products. Marketing Science, 17(4), 406–423.
Araman, V. F., & Caldentey, R. (2009). Dynamic pricing for nonperishable products with demand learning. Operations research, 57(5), 1169-1188.
Azoury, K. S. (1985). Bayes solution to dynamic inventory models under unknown demand distribution. Management Science, 31(9), 1150-1160.
Ban, G.-Y. (2020). Confidence intervals for data-driven inventory policies with demand censoring. Operations Research, 68(2), 309-326.
Bartmann, D., & Bach, M. F. (2012). Inventory control: models and methods, vol. 388. Springer Science & Business Media.
Besbes, O., Chaneton, J. M., & Moallemi, C. C. (2022). The exploration-exploitation trade-off in the newsvendor problem. Stochastic Systems, 12(4), 319-339.
Besbes, O., & Muharremoglu, A. (2013). On implications of demand censoring in the newsvendor problem. Management Science, 59(6), 1407-1424.
Bisi, A., Dada, M., & Tokdar, S. (2011). A censored-data multiperiod inventory problem with newsvendor demand distributions. Manufacturing & Service Operations Management, 13(4), 525-533.
Braden, D. J., & Freimer, M. (1991). Informational dynamics of censored observations. Management Science, 37(11), 1390-1404.
Caro, F., & Gallien, J. (2007). Dynamic assortment with demand learning for seasonal consumer goods. Management Science, 53(2), 276-292.
Chen, B., Jiang, J., Zhang, J., & Zhou, Z. (2024). Learning to order for inventory systems with lost sales and uncertain supplies. Management Science, 70(12), 8631-8646.
Chen, L., & Plambeck, E. L. (2008). Dynamic inventory management with learning about the demand distribution and substitution probability. Manufacturing & Service Operations Management, 10(2), 236-256.
Ding, X., Puterman, M. L., & Bisi, A. (2002). The censored newsvendor and the optimal acquisition of information. Operations Research, 50(3), 517-527.
Gel, E. S., & Salman, F. S. (2022). Dynamic ordering decisions with approximate learning of supply yield uncertainty. International Journal of Production Economics, 243, 108252.
Herbon, A. (2019). Replenishing an inventory system permitting lost sales under poisson demand and short lead times: Explicit performance and an efficient algorithm. Computers & Industrial Engineering, 137, 105998.
Hill, R. M. (1997). Applying bayesian methodology with a uniform prior to the single period inventory model. European Journal of Operational Research, 98(3), 555-562.
Hill, R. M. (2007). Continuous-review, lost-sales inventory models with poisson demand, a fixed lead time and no fixed order cost. European Journal of Operational Research, 176(2), 956-963.
Jain, A., Rudi, N., & Wang, T. (2015). Demand estimation and ordering under censoring: Stock-out timing is (almost) all you need. Operations Research, 63(1), 134-150.
Kitaeva, A. V., Livshits, K. I., & Ulyanova, E. S. (2018). Estimating the demand parameters for single period problem, markov-modulated poisson demand, large lot size, and unobserved lost sales. IFAC-PapersOnLine, 51(11), 882-887.
Lariviere, M. A., & Porteus, E. L. (1999). Stalking information: Bayesian inventory management with unobserved lost sales. Management Science, 45(3), 346-363.
Li, R., & Ryan, J. K. (2011). A bayesian inventory model using real-time condition monitoring information. Production and Operations Management, 20(5), 754-771.
Lovejoy, W. S. (1990). Myopic policies for some inventory models with uncertain demand distributions. Management Science, 36(6), 724-738.
Luo, Z., Guo, P., & Wang, Y. (2023). Manage inventories with learning on demands and buy-up substitution probability. Manufacturing & Service Operations Management, 25(2), 563-580.
Mersereau, A. J. (2015). Demand estimation from censored observations with inventory record inaccuracy. Manufacturing & Service Operations Management, 17(3), 335-349.
Nambiar, M., Simchi-Levi, D., & Wang, H. (2021). Dynamic inventory allocation with demand learning for seasonal goods. Production and Operations Management, 30(3), 750-765.
Pando, V., San-José, L. A., Sicilia, J., & Alcaide-López-de-Pablo, D. (2024). Pricing decision in a newsvendor model with partial backorders under normal probability distribution for the demand. Applied Mathematical Modelling, 132, 57-72.
Pettit, L., & Young, K. (1999). Bayesian analysis for inverse gaussian lifetime data with measures of degradation. Journal of Statistical Computation and Simulation, 63(3), 217-234.
Raiffa, H., & Schlaifer, R. (2000). Applied Statistical Decision Theory, vol. 78. John Wiley & Sons.
Scarf, H. (1959). Bayes solutions of the statistical inventory problem. The Annals of Mathematical Statistics, 30(2), 490-508.
Schulte, B., & Sachs, A.-L. (2020). The price-setting newsvendor with poisson demand. European Journal of Operational Research, 283(1), 125-137.
van Jaarsveld, W., & Arts, J. (2024). Projected inventory-level policies for lost sales inventory systems: Asymptotic optimality in two regimes. Operations Research, 72(5), 1790-1805.