AN INVESTIGATION OF ALTERNATIVE DECISION RULE MODELS FOR PRODUCTION PLANNING UNDER CHANCE-CONSTRAINED SALES.
Item
-
Title
-
AN INVESTIGATION OF ALTERNATIVE DECISION RULE MODELS FOR PRODUCTION PLANNING UNDER CHANCE-CONSTRAINED SALES.
-
Identifier
-
AAI8112341
-
identifier
-
8112341
-
Creator
-
AFFISCO, JOHN FRANK.
-
Contributor
-
Georghios P. Sphicas | Seymour Kwerel
-
Date
-
1980
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Business Administration, Management
-
Abstract
-
This study deals with models to determine aggregate production plans when future demand is of a known stochastic nature. Specifically it proceeds along two directions. First, existing decision rule models, in which production depends solely upon previously realized sales feedback elements and an adjustment variable, which had been tested only under normal stationary demand are tested under exponential stationary demand. Second, two new forms of production decision rules are proposed and the corresponding decision rule models are formulated for the cases of normal and two parameter exponential demand. Both forms include, in addition to previous sales feedback elements, present forecasted demand elements.;The performance of these models under trend and seasonal demand patterns is investigated in a series of simulation studies. For each demand type two random sets of fifty replicates of demand over the respective planning horizon are generated. In each case the first set is used to select the best parameter value for the Type 1 quadratic programming model. Then the second set is used to choose the best from the competing models with the previously determined parametric value. Finally, for each pattern of demand, production plans determined through the best stochastic decision rule model are compared to plans developed by the HMMS Linear Decision Rule. These comparisons are effected through the use of the Wilcoxon Matched-Pairs Signed-Ranks Test.;For both normal trend and seasonal demand with 95% service level the stochastic decision rule models proved to be most competitive with the HMMS approach. In all three exponential demand cases the HMMS results significantly outperform the best decision rule models when the service level is 0.95. However when a less stringent service level of 75% was assumed in the stationary case the best decision rule model compared favorably to the HMMS approach.;Further research into the implications of service level for model performance should tell whether the stochastic decision rule approach is an acceptable technique for aggregate planning for exponential demand patterns. In addition, when demand is normally distributed future studies should concentrate on possible horizon effects in the best decision rule models. Ultimately these results should aid the systems analyst in selecting a model for aggregate production planning based upon the nature of the demand that is present.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Business
