Introduction
Supply Chains (SC) are a network of suppliers, warehouses, distribution centres and retailers through which raw materials are acquired, transformed and delivered to customers. Supply chains consist of many self-interested agents aiming to gather necessary raw materials from best suppliers in a timely and efficient manner to optimise their profit. Each partner in the supply chain can change parameters of its own such as production and inventory strategies, and may or may not have power over other suppliers in the chain through contracts. propose that many supply chains ''emerge'' rather than result from purposeful design by a singular entity. They further point out that membership of a supplier base is continuously reassessed and reshuffled as new organisations enter in the trade or existing suppliers modify their offerings.
Given the natural suitability of computational agent-based systems to model, analyse and drive supply chain problems, what started with agent-based simulation of the supply chain in the last decade continued with feasibility studies of agent software integration into enterprise resource and business information systems. Consequently, an industrial interest has started to form in the topic. In 2002 Enterprise Software Company, SAP started an initiative on adaptive supply chain networks using a combination of new technologies including agent-based systems, radio frequency identification (RFID), and web services . Honeywell Laboratories is examining multi agent technology for dynamically distributing supply chain management . The Boeing Company is currently conducting research on automating its service supply chain for perishable products . General Electric developed an enterprise network application where supplier agents bid on part specifications displayed by the company. It is forecasted that more companies will join in the agent-based supply chain trend .
The primary benefit of designing autonomous agents that govern a supply chain is that of most autonomous applications: relieving human agents of their repetitive, inefficient tasks. The autonomous supply chain leads to an improvement in timely communication as agents gather new data and react to changes dynamically. Most research in autonomous supply chain focused on agent architectures and frameworks for supply chain automation, modelling and analysis and examined dynamic production and logistics planning in the supply chain. This stream of research assumed static production strategies but dynamically re-scheduled to optimise. However, showed that changing the production strategy can impact the supply chain performance positively. While Forget et al. used simulation and design of experiments to decide on strategies, [6] used a genetic algorithm to