A Study of Data-driven Methods for Inventory Optimization

Lee Yeung Ping; Patrick Wong; Tan Cheng Han

REVIEW 4 major objections 6 minor 28 references

A Study of Data-driven Methods for Inventory Optimization

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that a Deep Q-Network reinforcement-learning agent is the most promising of three data-driven methods for supermarket inventory optimization across lost-sales, dual-sourcing, and multi-echelon models.

desk verdict Self-contradictory comparison: DRL training curves fail while the Conclusion calls DRL most promising, and the biased reward design undermines the ranking. read the letter →

arxiv 2505.08673 v1 pith:Y5QU36KK submitted 2025-05-13 cs.AI

Lee Yeung Ping , Patrick Wong , Tan Cheng Han This is my paper · ORCID

classification cs.AI

keywords inventoryoptimizationdeepreinforcementlearningDQNtimeseriesforecastingrandomforestlostsalesmodeldualsourcingmulti-echelon

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish which of three data-driven methods—time-series forecasting, Random Forest with Gradient Boosting as a companion, and a Deep Q-Network (DQN) reinforcement-learning agent—should be trusted with supermarket inventory decisions. Its conclusion is that all three methods can reduce lost sales and coordinate stock across supply-chain levels, and that DQN is the most promising in the lost-sales model and the most efficient in the multi-echelon model. If that ranking were correct, it would give retailers a concrete reason to favor reinforcement-learning systems over conventional forecasting. The paper's own discussion sections record that the DQN reward curves plateaued at large negative values and that the agent did not learn an optimal policy, so the reported advantage sits in tension with the training evidence.

What carries the argument

The load-bearing machinery is the custom simulation environment (`LostSalesInventoryEnv`) plus the DQN agent that learns inside it. The environment is built on a private supermarket dataset with hand-set costs (10 per unit ordered, 1 per day held, 50 per stockout, 5 per price reduction), a three-action space (order stock, do nothing, reduce prices), and a step function that assumes an order of 50 units and sets estimated lost sales to zero whenever the 'order stock' action is taken. The DQN agent uses two hidden layers of 64 ReLU neurons with dropout, experience replay, and epsilon-greedy exploration to produce the reward curves. The time-series forecaster and the tree regressors supply the demand forecasts that the other two comparison arms are built on.

What would settle it

Re-run the lost-sales experiment and log cumulative reward per episode across all 100 episodes, then compare the trained DQN against a simple order-up-to policy using the same cost parameters; if the DQN does not beat that policy on total cost, the central claim fails. The fastest check is to inspect the step() function: when 'order stock' is chosen, estimated lost sales is assumed to be 0, so setting that value to the true demand shortfall and re-training would decide whether the DRL advantage is real.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that a DQN agent trained in a custom supermarket inventory simulator achieves the best balance of lost-sales reduction, cost control, and multi-echelon coordination among time-series forecasting, Random Forest/Gradient Boosting, and deep reinforcement learning. In the lost-sales model the authors call DRL 'the most promising'; in the multi-echelon model they say DRL demonstrates 'the highest efficiency'; and in dual sourcing they credit all three methods with improving the regular/express sourcing trade-off. The paper also reports, in its discussion sections, that after parameter tuning the DQN agent produced smoother but still negative, flat reward curves and that the agent may not be learning an optimal policy or was unable to learn a stable policy. That internal report is part of the paper's own evidence and leaves the headline conclusion resting on the simulator's design rather than on observed improvement in rewards.

Load-bearing premise

The DRL results stand on the simulator's hand-set costs and on the rule that placing an order erases estimated lost sales, and the private dataset means none of the comparisons can be reproduced or audited.

Editorial extensions

If this is right

If the ranking is correct, a supermarket manager facing frequent stockouts should prefer a DQN-based replenishment policy over time-series or Random Forest in settings that match the simulated cost structure.
The near-perfect R2 scores reported for Random Forest and Gradient Boosting in the lost-sales and dual-sourcing models would imply that item, supplier, and date features carry almost all information needed to forecast lost sales and replenishment quantities.
The multi-echelon conclusion would imply that reinforcement learning can coordinate ordering decisions across warehouses, distribution centers, and stores more efficiently than static forecasting rules.
Because the DQN reward curves are flat even under the paper's own conclusion, the claim implies that stable, non-improving rewards should be read as evidence of a learned policy rather than as a sign of learning failure.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A direct test of the paper's ranking would be to remove the step-function shortcut that sets estimated lost sales to zero whenever stock is ordered; if the DQN advantage shrinks or reverses, the headline result is an artifact of that simulator rule.
The very high R2 values for the tree regressors suggest the target or a near-duplicate feature may be leaking into training; a temporal train/test split would settle whether the forecast advantage survives out-of-sample.
Because the dataset is private and the environment is bespoke, an independent replication on a public inventory benchmark (such as the beer-game setting the paper cites) would show whether DQN's edge transfers beyond this simulator.
One practical consequence the authors leave implicit is that the same comparison pipeline could be re-run with actual margin, holding, and stockout costs from a real store to turn the qualitative ranking into a dollar-denominated decision rule.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The short version: this is a careful, methodical application of three off-the-shelf methods (Prophet, Random Forest/Gradient Boosting, DQN) to three standard inventory models on a private Hong Kong supermarket dataset. The writing is detailed about code and parameters, and the paper's own discussion sections openly admit the DRL agents did not learn a good policy. But the Conclusion then claims DRL is 'most promising' and 'highest efficiency,' directly contradicting the reported flat and negative reward curves. That contradiction is load-bearing, so the paper cannot stand as a comparison.

What is genuinely useful: the authors document their environment construction, including cost parameters and action definitions, thoroughly enough that a reader can see exactly what was simulated. The time-series cross-validation tables are honest about horizon-dependent degradation, and the RF/GB results are reported with residuals and learning curves. That is real work, and it would be a decent technical appendix for a course project.

The soft spots are serious, not minor. First, the conclusion does not follow from the evidence. The DQN reward curves plateau at negative values with no upward trend, and Sections 3.5, 4.5, and 5.4 say the agent did not learn an optimal or stable policy. Calling that 'most promising' is unsupported. Second, the DRL environment in Section 3.3.1 zeroes estimated lost sales whenever the agent chooses 'order stock,' which removes stockout cost from exactly the action being rewarded; that alone makes the reward signal a poor basis for ranking methods. Third, the RF/GB R² values near 0.9999 are unexplained and suspicious, likely leakage or a near-trivial target, and without the dataset or code nobody can check. The private dataset and the absence of any baseline policy comparison, even a simple base-stock heuristic, make the numbers unverifiable.

Who is this for? A reader who wants a concrete walk-through of setting up Prophet, RF/GB, and DQN for three inventory models, with honest limitations in the body, might skim it. It is not a contribution to the research literature: Gijsbrechts et al. (2022), which the paper itself cites, already benchmarks deep RL on exactly these three model classes. I would not bring it to a reading group or cite it.

Recommendation: desk reject. The central evaluative claim is internally contradicted by the paper's own results, and the environment design flaw plus missing data and code make the numbers unverifiable. If the authors fixed the conclusion and added a base-stock baseline plus released code and data, it might become a modest empirical study, but as is it does not deserve referee time.

Referee Report

4 major / 6 minor

Summary. The paper compares three data-driven methodologies—Time Series (Prophet), Random Forest/Gradient Boosting, and Deep Reinforcement Learning (DQN)—applied to three inventory models (lost sales, dual-sourcing, and multi-echelon) in a supermarket context. For each model the authors report forecast metrics, training curves, and qualitative comparisons, and they conclude that all three methods are effective while DRL is the most promising (lost-sales) and shows the highest efficiency (multi-echelon). The study relies on a private, undisclosed supermarket dataset and hand-coded simulation environments, with no theoretical derivation and no reproduction package.

Significance. If the central evaluative claim were supported, the paper would provide practical guidance for choosing among forecasting and learning methods for supermarket inventory management. The paper has the merit of covering three inventory models and three algorithm families, and the Prophet, Random Forest, and Gradient Boosting forecasting pipelines are described in enough detail to be reimplemented. However, the central claim is contradicted by the paper's own reported evidence: the DQN training curves are flat or declining and negative, and the text repeatedly states that the agent did not learn an optimal or stable policy. In addition, the DRL environment's reward design and the absence of any baseline comparison make the claimed ranking of methods unsupported. The private dataset and arbitrary cost parameters further limit the significance of the empirical comparison.

major comments (4)

[Section 6 vs. Sections 3.4, 3.5, 4.5, 5.4] The conclusion states that DRL exhibits the most promising results in the lost-sales model and the highest efficiency in the multi-echelon model, but this claim is contradicted by the paper's own evidence. Section 3.4 states that the DQN is not successful in learning the optimal inventory management policy; Section 3.5 reports no obvious upward trend for the total rewards after parameter tuning; Section 4.5 states that the agent is unable to learn in a stable policy; and Section 5.4 concedes that an upward and increasing trend in the total rewards is expected for cost optimization but was not observed. Figures 10, 24, and 37 show total rewards that are negative and flat (or declining) over episodes. A flat or declining reward trajectory cannot support most promising or highest efficiency unless DRL is benchmarked against a competent baseline in the same cost environment, which is not reported. This internal contradiction is load-bearing for the paper's central conclusion.
[Section 3.3.1, step() function] The DRL environment's reward function encodes the policy it is supposed to evaluate. When the agent takes the order stock action, the step() function assumes order quantity and estimated lost sales as 50 units and 0 respectively, meaning that the stockout cost term (defined later in the same subsection) is identically zero for exactly the action the agent is incentivized to choose. This makes the reward signal a direct function of the agent's action rather than a neutral measure of inventory performance, so the claim that DRL minimizes lost sales is circular with respect to the environment definition. The cost parameters (order cost 10, holding cost 1, stockout cost 50, price reduction cost 5) are asserted without justification or sensitivity analysis, and no comparison to a base-stock or other standard policy is provided to calibrate the reward scale.
[Sections 3.3.2, 4.3.2, 5.3.2] The DRL results are not compared to any control policy or to the forecasting methods on the same cost metric. The total-reward values are reported in arbitrary units (e.g., approximately -726,000 to -729,000 in Figure 24), while the Time Series and Random Forest sections report MAPE, MSE, and R2, which are not directly comparable. Without a common baseline (such as a myopic order-up-to policy evaluated in the same simulation environment), the relative ranking of DRL against TS and RF is unsupported. Sections 4.4 and 5.4 offer only qualitative statements about DRL's ability to learn and adapt, which cannot establish the degree of efficiency claimed in the conclusion.
[Sections 3.1.2, 4.1.2, 5.2.2 (Tables 1, 3, 4, 9, 11)] The empirical evidence base is weak and partially inconsistent. The Prophet cross-validation tables report high MAPE values (e.g., 0.20-0.69 in Table 1, 0.43-1.08 in Table 4, 0.43-0.45 in Table 9), indicating substantial forecast error, yet the conclusion claims these methods minimize lost sales and optimize inventory levels. The Random Forest and Gradient Boosting R2 values in the lost-sales model (0.9999 in Table 3) are implausibly high for real sales data and suggest possible target leakage or an effectively trivial prediction task; this possibility is not discussed. In the multi-echelon model the same methods achieve only R2 around 0.69 and 0.66 (Table 11), which is a more credible range but requires a very different interpretation of their accuracy. The paper does not reconcile these discrepancies, and the private dataset is not described or released, so none of the numerical comparisons can be audited or reproduced.

minor comments (6)

[Section 4.2.2 (Table 7)] The entry 0..9999428256116194 for the Gradient Boosting R2 score contains a typographical double decimal point; it should read 0.9999428256116194.
[Section 5.4] The text refers to DLR graphs where DRL is intended; this typo appears in the paragraph beginning with the sentence about optimizing inventory level in all echelons.
[Sections 3.1.2, 4.1.2, 5.1.2] The descriptions of Figures 1-5, 14-19, and 28-32 are lengthy and repetitive; the x-axis and y-axis meanings are stated multiple times within each subsection. Condensing these descriptions would improve readability.
[Section 2.2.1] The reference to Liashenko and Yakymchuk in the text has no entry in the reference list with a complete year or venue; a full citation should be added.
[Section 3.3.1] The phrase the number of epsilon is greater than epsilon min should be reworded to the current epsilon value is greater than epsilon_min for clarity.
[Abstract and Section 1] Several sentences are grammatically incomplete or truncated, e.g., The main purpose is to analyse efficient methods for the data-driven. and particularly effective inventory management solutions.This research is designed. A careful editorial pass is needed.

Circularity Check

1 steps flagged · score 4.0 of 10

The conclusion that DRL is 'most promising' in the lost-sales model is built into the simulator: the step() function sets estimated lost sales to zero for the 'order stock' action, so minimizing lost sales is an assumption of the reward design rather than a learned result.

self definitional [Section 3.3.1 (Deep Reinforcement Learning Code Design, step() description) and Section 6 (Conclusion).]
"If the “order stock” action is taken, the order quantity and the estimated lost sales would be assumed as 50 units and 0 respectively. ... The estimated lost sales cost is multiplied by the current price per unit to determine the lost sales cost. Moreover, the reward is lowered by a defined stockout cost if the quantity sold is less than the total of the estimated lost sales and the current estimated demand."

The reward function defines away the very quantity the paper later claims DRL minimizes: taking the 'order stock' action sets estimated lost sales to 0, which zeroes the lost-sales cost term in the reward. The Conclusion's claim that DRL exhibits 'the most promising results' at minimizing lost sales therefore reflects this hand-coded simulator assumption, not an empirically learned policy. The paper's own Sections 3.4 and 3.5 state that the DQN 'is not successful in learning the optimal inventory management policy' and that after tuning 'there is no obvious upward trend for the total rewards,' confirming the praise is not supported by observed learning behavior.

full rationale

This is an empirical study with no theoretical derivation chain or uniqueness theorem, so the main circularity risk is in the evaluation design. One self-definitional step is present: the DRL environment's reward function sets estimated lost sales to zero whenever the agent chooses the 'order stock' action, and the lost-sales cost is computed as estimated lost sales times price; the Conclusion's verdict that DRL minimizes lost sales is therefore true by construction of the reward rather than by agent performance. The paper also repeatedly concedes that the DQN did not learn a stable or optimal policy and that total rewards showed no upward trend, which makes the 'most promising' verdict an internal contradiction, though that is a correctness concern rather than circularity. No self-citation chain, fitted-parameter-as-prediction, or renamed-known-result was found; the Time Series and Random Forest sections report standard holdout metrics with no evidence that the test target was used to construct the features. The score of 4 reflects one definitional circularity in the central DRL claim without additional circular steps.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central comparison is supported by hand-set costs, an undisclosed dataset, and untested modeling assumptions. The DRL environment's action semantics (order action zeroes lost sales) create a favorable bias, and the unexplained near-perfect RF scores indicate a likely leakage problem. No invented physical entities are introduced.

free parameters (7)

Prophet hyperparameters (changepoint_prior_scale, seasonality_prior_scale, horizon, seasonality toggles) = changepoint_prior_scale=0.1 (lost sales) / 0.05 (dual-sourcing, multi-echelon); seasonality_prior_scale=10…
Set by hand in Section 3.1.1 and 4.1.1 to control trend flexibility and seasonality; the time-series comparison depends on them.
DQN environment cost parameters = order cost=10, holding cost=1, stockout cost=50, price reduction cost=5; dual-sourcing order costs 10/15…
Chosen without calibration in Sections 3.3.1, 4.3.1, 5.3.1; they define the reward that determines the DRL comparison.
DQN episode and network hyperparameters = episodes=100; batch size=10; learning rate=0.001 (later 0.005); epsilon decay=0.995 (later 0.885); epsilon min=0.01…
Set by hand in Section 3.3.1 and adjusted in Discussions 3.5, 4.5, 5.5; no sensitivity analysis around these choices.
DRL environment structural constants = order quantity=50 units; estimated lost sales=0 for order action; initial inventory=100; lead times=5 and 10 days…
Assumed in the step() functions (Sections 3.3.1, 4.3.1, 5.3.1); these force the reward consequences the agent learns.
Random Forest / Gradient Boosting hyperparameter search ranges = n_estimators 50-200; max_depth 2-10; learning rate 0.01-0.1; min_samples_split 2-20; min_samples_leaf 1-20
Search ranges in Section 5.2.1; best values are not reported, so the RF/GB comparisons are not reproducible from the text.
Random Forest feature set = Lost-sales model: Supplier ID Encoded, Item ID Encoded, Shelf-life, Replenishment Year, Quantity Sold x Price…
Feature choices in Sections 3.2.1 and 4.2.1; the near-1 R2 values (Tables 3, 7) suggest possible leakage that the paper does not investigate.
Order decision threshold = minimum order quantity threshold = 10 units
Section 4.2.1 sets the threshold that determines which forecasts become orders, affecting the dual-sourcing decision table.

assumptions (5)

domain assumption The undisclosed supermarket dataset is accurate and representative.
No source, collection method, size, or sampling is given; all metrics and conclusions depend on this data (Sections 3.1.1, 3.2.1, 3.3.1).
ad hoc to paper The DRL environment's action semantics are a valid model of supermarket inventory.
Section 3.3.1 assumes ordering sets estimated lost sales to zero and uses fixed order, holding, and stockout costs; these assumptions are not justified by real data.
domain assumption Demand in the dual-sourcing model is normally distributed with mean mu and variance sigma squared.
Section 2.1.2 invokes Fong et al. (2000); the paper does not test normality on its dataset.
domain assumption DQN can learn a useful policy in 100 episodes with the chosen hyperparameters.
The paper assumes training is sufficient, but its own reward curves (Figures 10, 24, 37) show no improvement, so this assumption fails in practice.
domain assumption The RF/GB near-perfect R2 values are genuine and not caused by data leakage.
Tables 3 and 7 report R2 approximately 0.9999 for lost sales and dual-sourcing without feature-level leakage checks; the paper does not establish this.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Study of Data-driven Methods for Inventory Optimization." pith.science (2026). https://pith.science/paper/Y5QU36KK

@misc{pith2026250508673,
  author       = {Pith},
  title        = {Pith review of: A Study of Data-driven Methods for Inventory Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y5QU36KK}},
  note         = {Machine review of arXiv:2505.08673}
}

read the original abstract

This paper shows a comprehensive analysis of three algorithms (Time Series, Random Forest (RF) and Deep Reinforcement Learning) into three inventory models (the Lost Sales, Dual-Sourcing and Multi-Echelon Inventory Model). These methodologies are applied in the supermarket context. The main purpose is to analyse efficient methods for the data-driven. Their possibility, potential and current challenges are taken into consideration in this report. By comparing the results in each model, the effectiveness of each algorithm is evaluated based on several key performance indicators, including forecast accuracy, adaptability to market changes, and overall impact on inventory costs and customer satisfaction levels. The data visualization tools and statistical metrics are the indicators for the comparisons and show some obvious trends and patterns that can guide decision-making in inventory management. These tools enable managers to not only track the performance of different algorithms in real-time but also to drill down into specific data points to understand the underlying causes of inventory fluctuations. This level of detail is crucial for pinpointing inefficiencies and areas for improvement within the supply chain.

Figures

Figures reproduced from arXiv: 2505.08673 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 20 canonical work pages

[1]

Neural inventory control in networks via hindsight differentiable policy optimization

Matias Alvo, Daniel Russo, and Yash Kanoria. Neural inventory control in networks via hindsight differentiable policy optimization. arXiv preprint arXiv:2306.11246, 2023

arXiv 2023
[2]

Deep reinforcement learning: A brief survey

Kai Arulkumaran, Marc Peter Deisenroth, Miles Brundage, and Anil Anthony Bharath. Deep reinforcement learning: A brief survey. IEEE Signal Processing Magazine, 34 0 (6): 0 26--38, 2017

2017
[3]

Lost-sales inventory theory: A review

Marco Bijvank and Iris FA Vis. Lost-sales inventory theory: A review. European Journal of Operational Research, 215 0 (1): 0 1--13, 2011

work page 2011
[4]

Optimal policy structure characterization for a two-period dual-sourcing inventory control model with forecast updating

Ali Cheaitou, Christian van Delft, Zied Jemai, and Yves Dallery. Optimal policy structure characterization for a two-period dual-sourcing inventory control model with forecast updating. International Journal of Production Economics, 157: 0 238--249, 2014

work page 2014
[5]

Learning to order for inventory systems with lost sales and uncertain supplies

Boxiao Chen, Jiashuo Jiang, Jiawei Zhang, and Zhengyuan Zhou. Learning to order for inventory systems with lost sales and uncertain supplies. Management Science, 70 0 (12): 0 8631--8646, 2024

work page 2024
[6]

Strategy, planning, and operation

Sunil Chopra and Peter Meindl. Strategy, planning, and operation. Supply Chain Management, 15 0 (5): 0 71--85, 2001

work page 2001
[7]

Supermarket policies on less-healthy food at checkouts: natural experimental evaluation using interrupted time series analyses of purchases

Katrine T Ejlerskov, Stephen J Sharp, Martine Stead, Ashley J Adamson, Martin White, and Jean Adams. Supermarket policies on less-healthy food at checkouts: natural experimental evaluation using interrupted time series analyses of purchases. PLoS medicine, 15 0 (12): 0 e1002712, 2018

work page 2018
[8]

Centralized planning models for multi-echelon inventory systems under uncertainty

Awi Federgruen. Centralized planning models for multi-echelon inventory systems under uncertainty. Handbooks in operations research and management science, 4: 0 133--173, 1993

work page 1993

Show all 28 references

[9]

The new science of retailing: how analytics are transforming the supply chain and improving performance

Marshall Fisher and Ananth Raman. The new science of retailing: how analytics are transforming the supply chain and improving performance. Harvard Business Review Press, 2010

2010
[10]

Analysis of a dual sourcing inventory model with normal unit demand and erlang mixture lead times

Duncan KH Fong, Virginia M Gempesaw, and J Keith Ord. Analysis of a dual sourcing inventory model with normal unit demand and erlang mixture lead times. European Journal of Operational Research, 120 0 (1): 0 97--107, 2000

2000
[11]

Deep reinforcement learning in inventory management

Kevin Geevers. Deep reinforcement learning in inventory management. Master's thesis, University of Twente, 2020

2020
[12]

Variable selection using random forests

Robin Genuer, Jean-Michel Poggi, and Christine Tuleau-Malot. Variable selection using random forests. Pattern recognition letters, 31 0 (14): 0 2225--2236, 2010

2010
[13]

Can deep reinforcement learning improve inventory management? performance on lost sales, dual-sourcing, and multi-echelon problems

Joren Gijsbrechts, Robert N Boute, Jan A Van Mieghem, and Dennis J Zhang. Can deep reinforcement learning improve inventory management? performance on lost sales, dual-sourcing, and multi-echelon problems. Manufacturing & Service Operations Management, 24 0 (3): 0 1349--1368, 2022

2022
[14]

Forecasting the forecastability quotient for inventory management

Arthur V Hill, Weiyong Zhang, and Gerald F Burch. Forecasting the forecastability quotient for inventory management. International Journal of Forecasting, 31 0 (3): 0 651--663, 2015

2015
[15]

The application of time-series forecasting models in grocery retail industry

O Liashenko and B Yakymchuk. The application of time-series forecasting models in grocery retail industry
[16]

Reinforcement learning with intrinsically motivated feedback graph for lost-sales inventory control

Zifan Liu, Xinran Li, Shibo Chen, Gen Li, Jiashuo Jiang, and Jun Zhang. Reinforcement learning with intrinsically motivated feedback graph for lost-sales inventory control. arXiv preprint arXiv:2406.18351, 2024

2024 arXiv
[17]

Deep inventory management

Dhruv Madeka, Kari Torkkola, Carson Eisenach, Anna Luo, Dean P Foster, and Sham M Kakade. Deep inventory management. arXiv preprint arXiv:2210.03137, 2022

2022 arXiv
[18]

Applied time series analysis: A practical guide to modeling and forecasting

Terence C Mills. Applied time series analysis: A practical guide to modeling and forecasting. Academic press, 2019

2019
[19]

Human-level control through deep reinforcement learning

Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, et al. Human-level control through deep reinforcement learning. nature, 518 0 (7540): 0 529--533, 2015

2015
[20]

Asynchronous methods for deep reinforcement learning

Volodymyr Mnih, Adria Puigdomenech Badia, Mehdi Mirza, Alex Graves, Timothy Lillicrap, Tim Harley, David Silver, and Koray Kavukcuoglu. Asynchronous methods for deep reinforcement learning. In International conference on machine learning, pages 1928--1937. PmLR, 2016

1928
[21]

A data-driven approach to predict the success of bank telemarketing

S \'e rgio Moro, Paulo Cortez, and Paulo Rita. A data-driven approach to predict the success of bank telemarketing. Decision Support Systems, 62: 0 22--31, 2014

2014
[22]

A deep q-network for the beer game: Deep reinforcement learning for inventory optimization

Afshin Oroojlooyjadid, MohammadReza Nazari, Lawrence V Snyder, and Martin Tak \'a c . A deep q-network for the beer game: Deep reinforcement learning for inventory optimization. Manufacturing & Service Operations Management, 24 0 (1): 0 285--304, 2022

2022
[23]

How many trees in a random forest? In Machine Learning and Data Mining in Pattern Recognition: 8th International Conference, MLDM 2012, Berlin, Germany, July 13-20, 2012

Thais Mayumi Oshiro, Pedro Santoro Perez, and Jos \'e Augusto Baranauskas. How many trees in a random forest? In Machine Learning and Data Mining in Pattern Recognition: 8th International Conference, MLDM 2012, Berlin, Germany, July 13-20, 2012. Proceedings 8, pages 154--168. ...

2012
[24]

Approximate Dynamic Programming: Solving the curses of dimensionality, volume 703

Warren B Powell. Approximate Dynamic Programming: Solving the curses of dimensionality, volume 703. John Wiley & Sons, 2007

2007
[25]

Mastering the game of go with deep neural networks and tree search

David Silver, Aja Huang, Chris J Maddison, Arthur Guez, Laurent Sifre, George Van Den Driessche, Julian Schrittwieser, Ioannis Antonoglou, Veda Panneershelvam, Marc Lanctot, et al. Mastering the game of go with deep neural networks and tree search. nature, 529 0 (7587): 0 484-...

2016
[26]

Improving organizational decision-making using random forests and big data analytics

P Sruthi, Krishna Kant Dixit, Bantu Mahesh, R Padmavathy, Muntather Almusawi, et al. Improving organizational decision-making using random forests and big data analytics. In 2024 International Conference on IoT, Communication and Automation Technology (ICICAT), pages 1392--139...

2024
[27]

Missforest—non-parametric missing value imputation for mixed-type data

Daniel J Stekhoven and Peter B \"u hlmann. Missforest—non-parametric missing value imputation for mixed-type data. Bioinformatics, 28 0 (1): 0 112--118, 2012

2012
[28]

Mining data with random forests: A survey and results of new tests

Antanas Verikas, Adas Gelzinis, and Marija Bacauskiene. Mining data with random forests: A survey and results of new tests. Pattern recognition, 44 0 (2): 0 330--349, 2011

2011

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Neural inventory control in networks via hindsight differentiable policy optimization

Matias Alvo, Daniel Russo, and Yash Kanoria. Neural inventory control in networks via hindsight differentiable policy optimization. arXiv preprint arXiv:2306.11246, 2023

arXiv 2023

[2] [2]

Deep reinforcement learning: A brief survey

Kai Arulkumaran, Marc Peter Deisenroth, Miles Brundage, and Anil Anthony Bharath. Deep reinforcement learning: A brief survey. IEEE Signal Processing Magazine, 34 0 (6): 0 26--38, 2017

2017

[3] [3]

Lost-sales inventory theory: A review

Marco Bijvank and Iris FA Vis. Lost-sales inventory theory: A review. European Journal of Operational Research, 215 0 (1): 0 1--13, 2011

work page 2011

[4] [4]

Optimal policy structure characterization for a two-period dual-sourcing inventory control model with forecast updating

Ali Cheaitou, Christian van Delft, Zied Jemai, and Yves Dallery. Optimal policy structure characterization for a two-period dual-sourcing inventory control model with forecast updating. International Journal of Production Economics, 157: 0 238--249, 2014

work page 2014

[5] [5]

Learning to order for inventory systems with lost sales and uncertain supplies

Boxiao Chen, Jiashuo Jiang, Jiawei Zhang, and Zhengyuan Zhou. Learning to order for inventory systems with lost sales and uncertain supplies. Management Science, 70 0 (12): 0 8631--8646, 2024

work page 2024

[6] [6]

Strategy, planning, and operation

Sunil Chopra and Peter Meindl. Strategy, planning, and operation. Supply Chain Management, 15 0 (5): 0 71--85, 2001

work page 2001

[7] [7]

Supermarket policies on less-healthy food at checkouts: natural experimental evaluation using interrupted time series analyses of purchases

Katrine T Ejlerskov, Stephen J Sharp, Martine Stead, Ashley J Adamson, Martin White, and Jean Adams. Supermarket policies on less-healthy food at checkouts: natural experimental evaluation using interrupted time series analyses of purchases. PLoS medicine, 15 0 (12): 0 e1002712, 2018

work page 2018

[8] [8]

Centralized planning models for multi-echelon inventory systems under uncertainty

Awi Federgruen. Centralized planning models for multi-echelon inventory systems under uncertainty. Handbooks in operations research and management science, 4: 0 133--173, 1993

work page 1993