Demand Estimation from Censored Observations with Inventory Record Inaccuracy

  title={Demand Estimation from Censored Observations with Inventory Record Inaccuracy},
  author={Adam J. Mersereau},
  journal={Manuf. Serv. Oper. Manag.},
A retailer cannot sell more than it has in stock; therefore, its sales observations are a censored representation of the underlying demand process. When a retailer forecasts demand based on past sales observations, it requires an estimation approach that accounts for this censoring. Several authors have analyzed inventory management with demand learning in environments with censored observations, but the authors assume that inventory levels are known and hence that stockouts are observed… 

Figures and Tables from this paper

On the Hardness of Inventory Management with Censored Demand Data

A repeated newsvendor problem where the inventory manager has no prior information about the demand, and can access only censored/sales data is considered, and a simple randomized policy, termed the Exponentially Weighted Forecaster, combined with a carefully designed cost estimator achieves optimal scaling of the expected regret.

Inventory Auditing and Replenishment Using Point‐of‐Sales Data

Events such as spoilage, expiration, employee theft, and customer shoplifting reduce available inventories in retail stores without these reductions being reflected in inventory records. As a result,

Managing Inventory with Cash Register Information: Sales Recorded but Not Demands

type="main" xml:id="poms12511-abs-0001"> Inventory inaccuracy is common in many businesses. While retailers employ cash registers to enter incoming orders and outgoing sales, inaccuracy arises

Estimation of potential lost sales in retail networks of high-value substitutable products

Abstract Sales data reveal only partial information about demand due to stockout-based substitutions and lost sales. We develop a data-driven algorithm to estimate stockout-based lost sales and

The Exploration-Exploitation Trade-off in the Newsvendor Problem

When an inventory manager attempts to construct probabilistic models of demand based on past data, demand samples are almost never available: only sales data can be used. This demand censoring

Impacts of Inventory Record Inaccuracy on Retailers’ Internal Operations

It is proved that the optimal restocking policy is a threshold policy in a three-dimensional state space that defines the Bayesian belief dynamics of the shelf stock level and compared with the best performing threshold-based heuristic policies in both time and probability domains.

Newsvendor Decisions with Two-Sided Learning

A substantial collection of work in operations management considers settings where a firm faces uncertain demand that depends on parameters that are ex ante unknown but can be learned by observing

Fixing Phantom Stockouts: Optimal Data‐Driven Shelf Inspection Policies

This paper proposes two partially-observable Markov decision models for the shrinkage problem and the shelf execution failure problem, and shows that there exists a closed-form condition to determine the optimal inspection threshold based on the number of consecutive zero-sales periods.

Estimating Demand with Constrained Data and Product Substitutions

This chapter shows why this is rarely a good assumption and provides a review of the statistical techniques that have been developed to correct product demand distribution estimates that are biased due to truncated demand or product substitution effects.



Information-Sensitive Replenishment when Inventory Records Are Inaccurate

A periodic review inventory system with imperfect inventory records and unobserved lost sales is considered, and it is found that the myopic heuristic is likely sufficiently good in practical settings targeting high service levels.

A Nonparametric Asymptotic Analysis of Inventory Planning with Censored Demand

Stochastic inventory planning with lost sales and instantaneous replenishment where knowledge of the demand distribution is not available is studied and nonparametric adaptive policies that generate ordering decisions over time are proposed.

Stalking Information: Bayesian Inventory Management with Unobserved Lost Sales

Retailers are frequently uncertain about the underlying demand distribution of a new product. When taking the empirical Bayesian approach of Scarf 1959, they simultaneously stock the product over

Retail Inventory Management When Records Are Inaccurate

This paper considers an intelligent inventory management tool that accounts for record inaccuracy using a Bayesian belief of the physical inventory level, and shows that a probability distribution on physical inventory levels is a sufficient summary of past sales and replenishment observations and that this probability distribution can be efficiently updated as observations are accumulated.

Demand estimation in lost sales inventory systems

This article considers the problem of estimating parameters of the demand distribution in lost sales inventory systems. In periods when lost sales occur demand is not observed; one knows only that

Bounds and Heuristics for Optimal Bayesian Inventory Control with Unobserved Lost Sales

This paper considers a finite-horizon inventory control problem for a nonperishable product with unobserved lost sales and a demand distribution having an unknown parameter and proposes two heuristics that outperform the myopic policies by a wide margin.

A Censored-Data Multiperiod Inventory Problem with Newsvendor Demand Distributions

It is shown that the Weibull is the only newsvendor distribution for which the optimal solution can be expressed in scalable form and for the special case of exponential demand the cost function is convex, so that for the storable inventory case, the optimal policy can be found using simple one-step look-ahead recursions whereas for the perishable case the optimal Policy can be expression by exact closed-form formulas.

Dynamic Inventory Management with Learning About the Demand Distribution and Substitution Probability

It is proved that with nonperishable inventory, the famous “ stock more” result is often reversed to “stock less,” in that the Bayesian optimal inventory level with unobserved lost sales is lower than the myopic inventory level.

Estimating negative binomial demand for retail inventory management with unobservable lost sales

The importance of effective inventory management has greatly increased for many major retailers because of more intense competition. Retail inventory management methods often use assumptions and

Adaptive Data-Driven Inventory Control with Censored Demand Based on Kaplan-Meier Estimator

This work proposes a new class of nonparametric adaptive data-driven policies for stochastic inventory control problems on the distribution-free newsvendor model with censored demands and obtains new results on the asymptotic consistency of the Kaplan-Meier estimator for discrete random variables that extend existing work in statistics.