An overview on agent based modeling and simulation

This review presented the following parts of ABMS research: (1) Theories and techniques. In the part, definitions and elements of ABMS were presented at first, and researches on various modeling frameworks and artificial intelligence based learning methods in agents were explained. ABMS development tools were classified into two categories: desktop tools and large-scale tools, and they were analyzed and compared. Then concepts and methods of data farming were introduced for output data analysis; (2) Applications. In this part, typical researches were classified into six classes. Authors, publishing date and research objectives of each research were explained. (3) Limitations of existing research. In this part, problems on the research of learning methods in agents, experimental methods of data fanning and discrepant frameworks were analyzed and presented; (4) Future research orientations. Research orientations of ABMS were proposed to solve above problems.