Agent-based e-supply chain decision support

Abstract

The global Internet economy is expected, over time, to give rise to increasingly agile practices where new supply chain arrangements are dynamically set up in response to changing business conditions and demands for highly customized products and services. To implement these "e-supply chain" practices, enterprises need the ability to rapidly evaluate new business opportunities and dynamically identify the best potential supply chain partners to respond to them. They also need the ability to effectively coordinate production and delivery of goods or services across the resulting value chains. In this article, we introduce MASCOT, an agent-based decision support environment for dynamic supply chain management. MASCOT agents help users distributed across multiple organizations and across different levels within a given enterprise to collaborate on the development and revision of supply chain solutions, as they evaluate different business opportunities (e.g., requests for bids from potential customers) and different sourcing options. This article also discusses new coordination protocols, developed within the context of MASCOT and aimed at better exploiting the power of finite capacity scheduling functionality across the supply chain. Empirical results are presented quantifying the benefits afforded by these new protocols under different loads and supply chain configurations.