Research On The Application Of Mobile Agent In Distirbuted Search

As the widely use of the Internet, the information on the internet has grown explosivelyin recent years. The distributiveness and openness of the massive resource, together with thenarrow and unstable information transmission channel, makes people hard to acquire the val-uable information accurately and quickly.Directing at this situation, the mobile agent technology which can effectively exploit theinternet broadband and process the internet resource by software with the existing hardwareequipment comes into being. It combines software, communication and dispersive technologyand has a good development potential. In dispersive information search, the mobile agenttechnology first transfers the code to the resource location.and then search and process thelocal resource by establishing index to it.Finally, it returns to initial search request with therequired search results. This technology solves the problems persisting in the traditionalsearch engine: slowness in nonlocal search and high system occupation rate.This paper first introduces the present situation of information search technology abroad.Directing at the deficiency of the existing dispersive search technology, the mobile agenttechnology is introduced. Based on the research on mobile agent technology, an individu-al-oriented mobile-agent-based search model were brought up. Within this search model thispaper conducts a research concerning several key technologies, i.e. the combination of mo-bile-agent technology and categorical computation, the route-selecting of mobile-agent trans-ferring and the parallel computation which involves the communication technology,fault-tolerance function and load balancing between mobile-agents.According the particularity of Chinese search, this paper makes an improvement based onthe existing Word segmentation algorithm by introducing the pre-processing of user search.When selecting route, the model avoids the unstable situations in traditional search modelwhich was conditioned by the net and transmission distance but adopts route-selecting strat-egy based on the user’s interest in the required content. In the last part of the paper it isproved with the IBM mobile-agent development tool that the mobile-agent model is superiorover the traditional one, it is applicable in certain area and this model lays foundation for theefficient information search.