Multi-Agent-Based Educational Environment

---

Wooldridge and Jennings (Wooldridge, 1994) gave one of the most comprehensive definitions of agents:

…a hardware or (more usually) a software-based computer system that enjoys the following properties: autonomy—agents operate without the direct intervention of humans or others, and have some kind of control over their actions and internal state; social ability—agents interact with other agents (and possibly humans) via some kind of agent-communication language; reactivity— agents perceive their environment and respond in a timely fashion to changes that occur in it; pro-activeness—agents do not simply act in response to their environment, they are able to exhibit goal-directed behavior by taking initiative.

We call agents used in an education environment the “educational agents.” The role of the educational agent is to provide task-related feedback and assistance to the learner and to guide the learner through the learning process and help the learner reach his or her learning goals. In an education environment, multiple agents are usually involved, and each agent plays different roles. There are two aspects to be considered in designing and building educational agents:

• Reusability: Reuse agents in different kinds of systems and environments.

• Interaction: In an environment containing multiple educational agents, tutor agents interact with each other and customize their behaviors based on the behaviors of other agents in the environment.

Norrie and Gaines proposed the following agents in an agent-oriented model for an education environment (Norrie, 1995):

• Knowledge Agent has knowledge in a particular area.

• Knowledge Server Agent stores, retrieves, and manages knowledge; answers queries; and provides information by inferring or reasoning using the stored knowledge bases.

• Interface Agent serves as an interface to learners, monitors and learns from the user’s actions, and then functions as an intelligent assistant.

• Coach or Tutor Agent provides guidance to assist in the learning process.

• Mediator Agent coordinates the activities of other agents and resolves conflicts between them.

• Knowledge Management Agent provides the high-level coordination of knowledge activities, such as creation, assembly, manipulation, and interpretation of knowledge, within either an individual or a collective project.

• Information Search Agent searches for specific information and sends the results back to learners.

• Directory Agent points to an appropriate agent, service, or resource.

• Mentor Agent is envisaged as acting in a rather analogous way in the learning environment, as a kind of coach for the higher-level strategies of learning.

Bruff and Williams illustrated an agent-based intelligent tutoring system architecture with the following three kinds of agents:

• Knowledge Management Agent responds to requests from other agents.

• Student Agent is assigned to each student and manages the evolution of a student model, which may include a representation of the student’s current knowledge and history about the topic and the student’s personal goals and preferences, etc. The student agent’s goals will typically vary from student to student or from time to time, even for the same student, and can be customized by a third party such as a human tutor. These goals determine the learning strategies and tasks to be used during a given learning session. The learning strategies together with the database describing the current state of the agent and its knowledge about the student’s capabilities will largely control the agent’s behaviors, that is, a customized agent for each individual learner.

• Inference Agent provides preset inference mechanisms, which include a group of agents, such as deduction, abduction, and induction agent, belief revision agent, possibility reasoning agent, nonmonotonic reasoning agent, and theory extraction agent, etc.

In general, all these three kinds of agents can be called tutoring agents, which are able to interact and cooperate with the student for tutoring and learning purposes. In the above architecture, we have three kinds of agents, and we assign more functions to each agent. Many different architectures have been proposed for an agent-based tutoring system. For example, Silveira and Viccari (Silveira, 1998) proposed several different agents: curriculum manager (the agent responsible for registering and controlling the curriculum attended by the students); agent communicator manager (the agent responsible for the agent’s society administration and for controlling the communication between agents); interface communication (the agents responsible for peer-to-peer communications between the student’s environment and the network environment); and presentation manager (the agent responsible for the browser control in the student’s environment). In a finer model, every tutoring agent will perform only one tutoring function. All of these functions should be performed as session-based actions. Tutoring functions may include the following (Morin, 1998):

• Select a subject element.

• Format and present a subject element.

• Format and present an explanation of a subject element.

• Compare different concepts.

• Select, format, and present an example.

• Answer a student’s question.

• Evaluate the student’s answer to a system-asked question.

• Send feedback to a student about his answer to a system-asked question.

• Diagnose a student’s behaviors.

• Update student model.

Tutoring systems for different courses or topics, or for students with different backgrounds, may have different preferences or requirements on system architecture (types and amount of agents, their responsibilities and interactions), and there is no existing universal architecture that will fit all. When designing agents for a learning environment, we have to understand the requirements of the to-be-built learning system first, consider the backgrounds and goals of its users, determine the types and numbers of agents we need, predict the interactions among them, and assign tasks accordingly.

---