Coursework submitted in partial fulfillment of the requirements for the degree of Computing. Plagiarism is not permitted and is punishable by law.
However, feel free to get inspired by this project. If it helped you, star this repo to show your appreciation :)
Manufacturing companies hardly ever handle all the processes that are behind the production of a good, which span from the extraction of raw materials through to the delivery of the final product. Instead, they are commonly part of a supply chain: they buy goods, perform some work on these, and sell them to the next customer. In other words, each company does not only act as the manufacturer, but also as the customer and the supplier. The dynamics of the enterprise market make the coordination for this type of system uncertain: materials delivery can be delayed, production can experience downtime, customers can cancel orders, and other issues can cause deviations from the plan (Umet al., 2010).
Multi-agent systems (MAS) reveal extremely suitable for this type of problem: agents possess features that can power autonomous, adaptive, collaborative and intelligent systems that allow companies to automate their inter-processes, such as providing automatic partner discovery, interaction and coordination, as well as their intra-processes, such as managing the production schedule and stock levels. Agents can react and adapt to changes, making optimal decisions based on complex global criteria that might not be entirely known and may involve conflicting interests. Each agent actively participates in the collaboration process: they can determine the effects that their decisions have on other agents and coordinate with them to choose an alternative solution that is optimal for the entire supply chain.
This project aimed to design and implement a PC supply chain multi-agent system in JADE; this system consists of three main separate agents: the customer, the manufacturer and the supplier. The project focuses on the manufacturer agent and the development of a strategy that maximises its profit. Moreover, particular attention was put into the implementation of an ontology and the effective use of communication protocols.