Electronic Catalogs for Supply Chain Management

Aryya Gangopadhyay
Information Systems
University of Maryland Baltimore County

2:00pm Friday November 12, 1999
Lecture Hall V, ECS

Electronic catalogs are an effective means for companies in the supply chain to disseminate product information. Paper catalogs are costly to publish and maintain for manufacturers and retailers, and difficult to browse, search through, and compare products for consumers. Electronic catalogs have been sought after as a technological solution to these ubiquitous problems for all trading partners in the supply chain of products and services. Proprietary catalogs in electronic media such as CD-ROMs and the Web have been found to be limited in their ability to provide effective browsing capabilities and supporting inexact queries. Current EDI-based electronic catalogs suffer from the limitation of requiring their creators to provide product information in terms of only twenty free-form text characters. This has led to the difficulty in correctly specifying the desired product specification, which in turn has caused confusion in communications between retailers and manufacturers. This problem has been exacerbated by the fact that the information about the products in current EDI-based catalogs is insufficient for consumers to make product selection. Since only limited amount of product information can be maintained in the catalog, service providers have to warehouse detailed product data, which in turn leads to the problems of data redundancy, errors, and delays in supplying product information.

In this talkwe describe a methodology and a prototype implementation of an electronic catalog that enables users to search for products by specifying alphanumeric descriptions and/or image characteristics. An alphanumeric query is a search for a word or phrase in attributes having textual or numerical values. Textual attributes could be long narratives such as product descriptions as well as short descriptors such as color. An image query is a search for an image in the system, based on characteristics specified by the user. Image characteristics can be specified by textual descriptions such as color, texture, size, and shapes, or by specifying modifications of a retrieved image. For example, if the system returns a list of images in response to a user request, as a second step the user can further modify the initial search. The system recognizes similarities among images irrespective of their differences in size. At the same time, however, the user can specify specific size characteristics to further narrow down the search space. The feasibility of using the system in supply chain management will be illustrated using a real life simulation.