These days people are getting a lot of information from networks such as the WWW using many kinds of search systems. But the traditional systems are not so efficient in many ways. At times it becomes very difficult to find the informations you need.
In this thesis we propose a new model for searching the WWW using mobile agent technology. The main idea is based on the fact that an expert has a lot of information in his discipline, and that if we share that information we can get more important information easily. We have applied this idea to our model. We have adopted mobile agent technology for the search, and also applied fuzzy inference to the search algorithm. The system have two types of agents. One is the stationary agent and the other is the mobile agent. The stationary agent dispatches the mobile agent to a remote host, and the dispatched mobile agent executes the given task at the remote host. After execution, the mobile agent sends the result to the stationary agent. And then, the agent returns to the local host and disposes itself. We also have implemented a host name server, to which a user could register or request. Our system can alleviate many problems present in traditional systems and by using this system we can search the WWW more efficiently than in traditional systems.