Research Of Multi-Agent Task Assignment Based On Granular Computing

Multi-Agent system is a relatively new branch of computer science. It is one of the major research fields of Distributed Artificial Intelligence while Multi-agent Task Allocation is its important sub-branch. Multi-agent Collaboration involves a task allocation, in which case people are confronted with abominable complications.Granular Computing is a new research focus in the field of artificial intelligence, and as a new approach to information and knowledge has been adopted and applied in recent years by more and more researchers in different areas. This thesis is a tentative try to look into a specific mandate tablets method for multi-agent task allocation, in the light of the relevant theory about the tablets and drawing on the concept of information tablets. This method aims at translating specific tasks tablets into several atomic tasks tablets space, resulting in a multi-agent task model based on the granular computing. According to the new model this thesis will further come up with a calculation on the basis of the tablets of the improved multi-agent task allocation algorithm, in order to improve the traditional design for the initial fitness function and optimization of the design groups in genetic algorithms, thereby applying an improved genetic algorithm to the solution of the task distribution of the Multi-Agent SystemThis paper has put the new calculation into use, through experiments, in the website classification system algorithm for the realization of a multi-task allocation. The analysis of the experiments the results has proved that the algorithm presented in this paper in dealing with the relatively large size of the initial groups, can effectively improve computational efficiency, better allocate the tasks of multi-agent system.