A Framework for Model-Based Adaptive Training

A Training Scenario

Industrial training realisation for the type of application presented in this chapter often involves on-the-job-training with the following simple steps:

  1. an initial introduction to the tasks;
  2. practice with guidance from experts, and,
  3. an assessment that the trainee is proficient at the tasks.

These 3 steps have been mapped to equivalent training units in the Proto-MOBAT experiments [MOBIT 1994, W6-W7].

Step 1 maps to presentation training units.

Step 2 maps to practice and demonstration training units.

Step 3 maps to question and test training units.

The approach in Proto-MOBAT supports the view that learning is most effective with learning by doing. “Doing” is also about getting the trainee to tackle and solve problems. This is the learning by problem solving approach. See examples in Table 4-4 Learning by Doing and Problem Solving. Table 4-2 shows a fixed problem solving method for TU12 in a procedural model. Knowledge for diagnosing the power plant circuit board is represented in an estimated 550 rules. As the associations for TU24 are presenting situation-dependent options in problem solving, coordinated patterns of problem solving exist for this associative model. The learning by problem solving approach is generally expected to be used in knowledge-intensive training tasks, using an associative type domain model. These models contain explicit problem solving methods which provide the trainer with the ability to recognise and facilitate different ways in the problem solving approach used by the trainee.

Learning by Doing and Problem Solving

Table 4-4 Learning by Doing and Problem Solving

© | | Sitemap