ON SOME STATISTICAL PROBLEMS IN INVENTORY SYSTEMS ASSOCIATED WITH MODELING THE LEAD TIME DEMAND.

Persistent Link:
http://hdl.handle.net/10150/186204
Title:
ON SOME STATISTICAL PROBLEMS IN INVENTORY SYSTEMS ASSOCIATED WITH MODELING THE LEAD TIME DEMAND.
Author:
MYKYTKA, EDWARD FRANK.
Issue Date:
1983
Publisher:
The University of Arizona.
Rights:
Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Abstract:
This dissertation contains a number of varied, yet closely related, results that are relevant to the construction of mathematical and statistical models of inventory systems. Its primary focus is on the sensitivity of some specific inventory models to errors in certain modeling assumptions. Motivation for this research is provided through the development of analytical expressions that show that the deterministic economic order quantity can be quite sensitive to errors in the forecast of the demand rate whenever the lead time is non-zero. Similar results are provided for the stochastic case by means of a carefully designed experiment that shows that the specific form or "shape" of the distribution chosen to represent the stochastic behavior of the lead time demand can have a significant impact on a minimum cost (Q,R) policy. Together, these results refute the "conventional wisdom" that inventory models are generally insensitive to errors in model specification or parameter estimation. Considerable attention is also given to the postulation of a "robust" model for the lead time demand distribution (LTDD). This discussion culminates with the introduction of a new probability distribution, based on a hyperbolic cosine transformation of normal random variables, that appears to be well-suited for modeling the LTDD. Furthermore, it is concluded that the two- and three-parameter versions of the lognormal and inverse Gaussian distributions can also be considered as viable candidates to model the LTDD in a wide variety of inventory systems. A number of new algorithms for computing optimal (Q,R) policies are also introduced. These significantly reduce both the amount and complexity of computation required by the standard iterative method. Two additional sets of analytical results are chronicled in this work. The first allows the LTDD to be characterized (by its first four moments) on the basis of information about the distributions of the lead time and demand rate. The second expresses the linear loss functions (LLF's) for a number of probability distributions whose LLF's are not readily available in the inventory control literature. Complete and intuitive proofs of these results are included.
Type:
text; Dissertation-Reproduction (electronic)
Keywords:
Inventories -- Mathematical models.; Inventory control.; Demand (Economic theory) -- Mathematical models.
Degree Name:
Ph.D.
Degree Level:
doctoral
Degree Program:
Systems and Industrial Engineering; Graduate College
Degree Grantor:
University of Arizona

Full metadata record

DC FieldValue Language
dc.language.isoenen_US
dc.titleON SOME STATISTICAL PROBLEMS IN INVENTORY SYSTEMS ASSOCIATED WITH MODELING THE LEAD TIME DEMAND.en_US
dc.creatorMYKYTKA, EDWARD FRANK.en_US
dc.contributor.authorMYKYTKA, EDWARD FRANK.en_US
dc.date.issued1983en_US
dc.publisherThe University of Arizona.en_US
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.en_US
dc.description.abstractThis dissertation contains a number of varied, yet closely related, results that are relevant to the construction of mathematical and statistical models of inventory systems. Its primary focus is on the sensitivity of some specific inventory models to errors in certain modeling assumptions. Motivation for this research is provided through the development of analytical expressions that show that the deterministic economic order quantity can be quite sensitive to errors in the forecast of the demand rate whenever the lead time is non-zero. Similar results are provided for the stochastic case by means of a carefully designed experiment that shows that the specific form or "shape" of the distribution chosen to represent the stochastic behavior of the lead time demand can have a significant impact on a minimum cost (Q,R) policy. Together, these results refute the "conventional wisdom" that inventory models are generally insensitive to errors in model specification or parameter estimation. Considerable attention is also given to the postulation of a "robust" model for the lead time demand distribution (LTDD). This discussion culminates with the introduction of a new probability distribution, based on a hyperbolic cosine transformation of normal random variables, that appears to be well-suited for modeling the LTDD. Furthermore, it is concluded that the two- and three-parameter versions of the lognormal and inverse Gaussian distributions can also be considered as viable candidates to model the LTDD in a wide variety of inventory systems. A number of new algorithms for computing optimal (Q,R) policies are also introduced. These significantly reduce both the amount and complexity of computation required by the standard iterative method. Two additional sets of analytical results are chronicled in this work. The first allows the LTDD to be characterized (by its first four moments) on the basis of information about the distributions of the lead time and demand rate. The second expresses the linear loss functions (LLF's) for a number of probability distributions whose LLF's are not readily available in the inventory control literature. Complete and intuitive proofs of these results are included.en_US
dc.typetexten_US
dc.typeDissertation-Reproduction (electronic)en_US
dc.subjectInventories -- Mathematical models.en_US
dc.subjectInventory control.en_US
dc.subjectDemand (Economic theory) -- Mathematical models.en_US
thesis.degree.namePh.D.en_US
thesis.degree.leveldoctoralen_US
thesis.degree.disciplineSystems and Industrial Engineeringen_US
thesis.degree.disciplineGraduate Collegeen_US
thesis.degree.grantorUniversity of Arizonaen_US
dc.identifier.proquest8315299en_US
dc.identifier.oclc688639718en_US
All Items in UA Campus Repository are protected by copyright, with all rights reserved, unless otherwise indicated.