Using multi-agent approach for the design of an intelligent learning environment

  • Dong Mei Zhang
  • Leila Alem
  • Kalina Yacef
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1441)


This paper presents FILIP, a multi-agent framework for the design of Intelligent Simulation-based Learning Environment (ISLE). Such a learning environment assesses the learner and provides adaptive instruction when the learner is developing his/her operational skill in dynamic and highly risky domains. The FILIP framework offers a great hope as a means of helping learners develop the skill necessary for effective performance. It is geared towards skill development and acquisition in order to develop and support operators performance. FILIP is designed as a multi agent architecture, which includes the seven agents: a simulator; a user interface; a domain expert agent; a learner agent; an instructor agent, a curriculum agent and a skill development agent. FILIP framework is distinguished from other traditional ISLE by its skill development agent and the curriculum agent. The skill development agent is to advise the instructor on issues related to the skill development of the learner so that it can be taken into account in the planning of the instruction. The curriculum agent is made of a Curriculum Formalism for Operational Skill Training (CFOST). It provides information for skill development agent to assess skill development and for the instructor agent to provide adaptive instruction.

As an application of FILIP in the context of Air Traffic Control (ATC) training, ATEEG is described. It adopts a blackboard architecture based on a multi-agent approach, in which major tasks of the training are distributed into agents. The architecture of ATEEG is composed of an air traffic control simulator; a domain expert called here the ideal ATC which represents an ATC expert and provides expert's problem solving strategies and solutions to given traffic situations presented to the learner; a learner which gathers the learner's actions during the execution of a training exercise and derives a bi-dimension learner model; an instructor which adapts instruction to the individual learner's needs at two levels: tailoring training exercises and adapting instructional methods; an ATC curriculum, which provides a hierarchical network of ATC training topics, training situations and instructional methods; and an ATC exercise base that represents existing ATC training exercises in terms of underlying training topics presented, complexity of training situations. An object sever facilities the data flow and communication among ATEEG agents.


multi-agent training system agent modelling intelligent tutoring system simulation-based learning 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    L. Alem & R. Keeling (1996). Intelligent simulation environments: an application to Air Traffic Control training. Proceedings of SimTecT'96, pp 323–328.Google Scholar
  2. [2]
    L. Alem, B.P. Woolf, A. Munro & W. van Joolingen (1996). Proceedings of Simulation-based learning technology workshop, held in conjunction with the Third International Conference in Intelligent Tutoring Systems (ITS'96), Montreal, Canada.Google Scholar
  3. [3]
    E. Costa, A. Perkusich (1996). Modelling the cooperative interactions in a teaching/learning situation. In Intelligent Tutoring Systems, Ed. C. Frasson, G. Gauthier & A. Lesgold.Google Scholar
  4. [4]
    C. Frasson, T. Mengelle, E. Aimeur, G. Gouarderes (1996). An actor-based architecture for intelligent tutoring systems. In Intelligent Tutoring Systems, Ed. C. Frasson, G. Gauthier & A. Lesgold.Google Scholar
  5. [5]
    K. Yacef & L. Alem (1997). Towards an assessment of skill acquisition is student modelling. Proceedings of AI-ED'97, Kobe, JapanGoogle Scholar
  6. [6]
    D. Zhang & L. Alem (1997). From task sequencing to curriculum planning. Proceedings of AI-ED'97, Kobe, Japan.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1998

Authors and Affiliations

  • Dong Mei Zhang
    • 1
  • Leila Alem
    • 1
  • Kalina Yacef
    • 2
  1. 1.CSIRO Mathematical and Information SciencesNorth RydeAustralia
  2. 2.Thomson Radar Australia Corporation and Laboratoire d'Intelligence Artificielle de Paris 5France

Personalised recommendations