Research On Constructing Distance Tutoring Adaptive Instruction Flat Model

Since thy,emergence of computer in 50's, computer experts and education experts have been studying to apply computer to education in order to improve the education effect. Then Computer-Assisted Instruction (CAI) emerges. At the early stage of CA1, computer is only regarded as a teaching tool, applying the same teaching content to the entire students, which submerges student's individuality. In order to improve CAI's effect, it requires that CAI system has capability in knowing of student's learning ability, learning base and current knowledge level, and producing the optimal teaching strategy according to the different traits of every student, by which gives specific individual instruct to every student. For this purpose, CAI system applies artificial intelligence technology, boosting the intelligent ability of teaching system, which come into bring Intelligent Tutoring System (ITS). In the current network stage, the ITS on stand-alone PC has not been suitable. People are willing to obtain education in everywhere, on every time. Modern distance education based computer network has become the main stream of CAI.There have been many modern distance-tutoring platforms in use. Especially in recent years, modern distance-tutoring platforms come forth rapidly. However, the quality of education of the platforms' should be improved. There are three directions in improving the modern distance-tutoring platform's education effect: (1) the structure of tutoring content; (2) making of tutoring media; (3) design of tutoring platform. As the design of tutoring platform is considered, in the network environment, existing distance tutoring platforms have been short of interacting and individualized instruction. One of feasible shortcuts to solve the individualized instruction problem in distributed environment is to incorporate the mature technology in stand-alone intelligent tutoring system and Distributed Artificial Intelligence technology, employing agent theories to prove distance-tutoring platform's intelligence. In teaching process, a complete teaching process includes teaching and reviewing. There are many researches focusing on teaching, research in reviewing is short of comparatively. Providing with intelligent, personalized review in distance-tutoring platform is one of the important parts of improving teaching effect, achievingindividualized teaching. Based on this idea, in this paper, we design a distance tutoring adaptive instruction flat with agent theories. Referring to the stand-alone intelligent tutoring system and applying agent theories put a three-layer structure model forward. The three layers are behavior layer, strategy-programming layer and resource layer. Behavior layer accomplishes teaching task, collecting feedback, and cooperate with teachers; strategy-programming layer forms the appropriate individual teaching strategy based on every registered student's characters; resource layer is the main store of all kinds of information resources, learning records and rules. This model can dynamic produce appropriate learning materials to individuals according to different student's learning level and learning stage; since we adopt a layer management and local adjust to students models and rules, the loads of network are lightened, avoiding the block of network. In addition, we import review in teaching platform model. By review agent diagnoses the lack of learning and gives a specific individual hint and instruct, which guarantee the teaching quality greatly.In DTAIF model, our research focus on the design of student model, applying knowledge space theories to form the knowledge tree of DTAIF model in knowledge representation. Using Bayes approach, we can find the learning deficiency.Finally, based on the evaluating review process in DTAIF, we establish an experiment system. This system completes the operation questions evaluating and analyzing the results, giving students' specific reviewing instruction.