|
Activity Number:
|
241
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing*
|
| Abstract - #300210 |
|
Title:
|
Statistical Methods for Information Management in Large Logistics Networks
|
|
Author(s):
|
Thomas Fender*+
|
|
Affiliation(s):
|
Universitaet Dortmund
|
|
Address:
|
Vogelpothsweg 87, Dortmund, International, D-44221, Germany
|
|
Keywords:
|
information management ; dimension reduction ; logistics networks ; robust methods ; identification of outliers
|
|
Abstract:
|
Modeling and simulation are very important techniques for the analysis and development of new structures or operations in large logistics networks. Any reasonable application of these kind of techniques assumes data of good quality, i.e. valid and structured data. Identification of information, data acquisition and data condensation are the bases for the extraction of the needed information out of data sets from logistics systems. Due to the highdimensionality and the heterogeneity of the data there is the necessity for statistical information management with modern robust methods. The special challenge lies in identifying classes of similar problems within these high dynamical systems and to assign adequate (probably new) statistical methods to solve the problems. For an example from the context of procurement channels we investigate the applicability and adequacy of statistical methods such as techniques for identification of outliers and methods for dimension reduction which seem promising for the use in modeling of large logistics networks.
|
