An Efficient Search Agent Software for Multi-Agent Systems

Multi-Agent Systems (MASs) have stimulated much interest in research because of the advantages found within such systems. They are able to deal with problems that are too large for single-agent systems, and are faster and more reliable. This paper examines how to use MASs in Search Agent Software. Examining how multi-agents function in a virtual system, the interactions between them and how they formulate decisions, forms part of the basis of the research into Artificial Intelligent systems. In this paper, we set out to develop and create Search Agent Software (SAS) that can retrieve requested information from two databases. The clients are able to search for a book using author, title, ISBN and, more significantly for this study, price. SAS will identify the books requested, where they can be found, and will offer the cheapest available. Each of the two databases has the same names of the titles, however each database may contain a different price. The SAS will get the required information quickly, helping the user to find the cheapest and possibly nearest. It will connect to the database using the Agent Interface; there are two agents in the system, known as MASs. These agents collect the required data and send it to the Agent Handler. This paper merges different program technologies including the following: Jade, Java, Finite State Machine and JFLAP.