Multi-Agent System for Web-Based Customization

Abstract

Web-based customization refers to an e-commerce business model whereby customers can individualize their products via the electronic channel. The success of this business model to a great extent depends on the appropriateness of the information system, which supports online interaction between customer and supplier (Franke & Piller, 2003). For example, computer manufacturers offering their products online enable customers to select the type of components to be built in the final product (e.g., type of processor, motherboards, graphic and sound cards, etc.). This type of interaction is referred to as Web-based product configuration. In order to increase the chances that customers find suitable products, Web-based customization offers large variety. However, due to the limited information processing capacity of humans and lack of technical product knowledge, excessive product variety confuses customers (Piller, Koch, Moeslein, & Schubert, 2003). It triggers decision-making difficulties and uncertainty concerning the suitability of choices. In effect, customers are generally unaware of their needs and not capable of making optimal buying decisions. Furthermore, the support provided by common information systems is more consistency check than real assistance, since their main task is to verify the compatibility of components between each other. On the other hand, the resulting variety brings about an increasing complexity of operations and manufacturing-related tasks on the supplier’s side. The complexity that is perceived by customers during the interaction process is called “external complexity,” whereas complexity that is experienced inside manufacturing and operations is referred to as “internal complexity.” Therefore, a suitable information system that supports Web-based customization should simultaneously address both challenges. It should provide customers with appropriate assistance during the search for adequate products and support suppliers in managing variety. In the following, an information system that is capable of coping with both types of complexity will be developed.