An Agent Federation for Supply Chain Integration

Supply chain management represents a critical competency in today’s global business environment, but neither commerce through EDI technology nor current Web applications satisfy combined requirements of process integration and flexibility. The present research builds on prior agent work to argue multi-agent systems offer tremendous potential and capability to overcome such limitations. The paper discusses enterprise supply chain integration and outlines the knowledge engineering techniques employed to implement a corresponding agent federation for high-technology product procurement. It highlights the use and utility of this supply chain agent federation and presents results of early verification and validation testing. Supply Chain Integration Supply chain management (see Porter and Millar 1985) represents a critical competency in today’s global business environment, and a number of effective practices (e.g., just-in-time deliveries, electronic data interchange (EDI), supplier inventory management) are employed to improve the competitiveness and efficiency of enterprises around the world. Our two decades of experience with EDI suggest commercial processes employed by buyers and sellers must be mutually-compatible in order for business exchanges and transactions to occur effectively, and for rapid purchase and responsive order fulfillment, buyer and seller processes must also be closely integrated (Sokol 1996). The process-integration task is now routinely enabled through EDI technology for many major corporations and government agencies. But not only can it be expensive, this technological approach requires a relatively stable and predictable commercial environment to effect supply chain integration. Such conditions are increasingly unlikely in the fast-paced, hypercompetitive business environment of today (see D’Aveni 1994). And modern enterprises now also require flexibility to quickly change vendors and adapt to dynamic economic conditions and diverse commercial practices (see Davidow and Malone 1992). Commerce through EDI technology does not satisfy the flexibility requirement well. One approach is to replace EDI with more flexible technology. For instance, with the continuing surge of activities on the World Wide Web (Web) and corresponding research on electronic commerce, many firms are moving to Web-based support for commercial transactions (e.g., electronic catalogs, virtual supply chains and storefronts, intranets/extranets). Indeed, the Internet offers good potential to "revolutionize procurement" and related commercial activities (Gebauer et al. 1998), and Web-based transactions are beginning to supplant the traditional EDI for some business-to-business and business-to-consumer commerce. However, much of the capability for supply chain integration is being lost during the transition from EDI to Web technology. Whereas EDI effectively compels buyers and sellers to integrate their supply chain processes, Web-based supply chain technologies are noticeably one-sided; that is, the latter sites and applications are predominately developed for either the buyer or seller, but not both. Commerce though current Web technologies does not satisfy the integration requirement well. The present research builds on work by Barbuceanu and Fox (1993) and others (e.g., Collins et al. 1998, Gini and Boddy 1998, Rodriguez-Aguilar et al. 1998, Walsh et al. 1998, Wurman et al. 1998) to argue intelligent agents offer tremendous potential and capability for both buyer-seller process integration and supply chain flexibility. And multiagent systems may be able to both disintermediate and integrate the supply chain in a cost-effective manner. With the capability to formalize and embed domain-specific knowledge and market-specific expertise in multi-agent systems, this emerging technology offers potential to substitute federations of intelligent agents for the bureaucratic and expensive intermediaries now employed along most major enterprise supply chains. It may also provide the same kinds of value-added services expected from other intermediaries, but without the attendant cost and time associated with human labor. From: AAAI Technical Report WS-99-02. Compilation copyright © 1999, AAAI (www.aaai.org). All rights reserved. Agent Supply Chain Integration This present work builds upon and extends research by Mehra and Nissen (1998) to implement and demonstrate an agent development environment, and by Nissen and Mehra (1998) that explores process redesign enabled by intelligent-agent technology. Two primary processes are involved with a supply chain—customer purchasing and vendor order fulfillment. Although these customer and vendor processes can be viewed as separate, intraorganizational activities within each of the respective buying and selling enterprises, a strong case can be made for viewing such activities together, as an integrated, interorganizational supply chain process. Clearly, before agents can be equipped with knowledge required for supply chain integration, their designers must first acquire the requisite knowledge. This is the essence of knowledge engineering, the requirement for which is not much different for multi-agent systems than their expert system counterparts from two decades ago. We present and discuss two, complementary process instances— procurement and order fulfillment—in terms of a single, integrated whole; that is, both purchasing and order fulfillment are modeled as a single process that spans organizational boundaries. Specifically, the university purchasing process examined through this investigation pertains to work done by the procurement department at a graduate educational institution on the West Coast. As an educational institution, this university facility is subject to the full complement of procurement policies and procedures that govern the purchasing activities of most university organizations. The commercial order fulfillment process examined through this investigation pertains to work done by the product and licensing department at a leading high-technology company on the East Coast. This company is a leader in its product market and maintains an active research and development activity that drives frequent introductions and updates. Therefore it maintains the kind of rapid product evolution that has been problematic for procurement in large enterprises such as the university. The high-level process delineated in Figure 1 depicts the integration of the user, university procurement department and commercial contractor. The author has invested several months' effort acquiring detailed knowledge associated with this process. Knowledge acquisition techniques include document analysis, personnel interviews and direct process observation. One advantage of research in the procurement domain is the process activities tend to be well documented, including, for example, detailed procedures used to guide the day-to-day activities of university procurement specialists. This expedites the knowledge engineering process considerably. ID rqmts User Procurement Contractor