INFERENCE AND REFLECTION IN THE OBJECT-CENTERED REPRESENTATION LANGUAGE NOOS

This paper explains the inference and reflection capabilities of NOOS, an object-centered representation language designed to integrate prob lem solving and learning. Problem solving and learning in NOOS are mod elled by means of concepts, tasks, methods and metalevels. Metalevels allow NOOS to reason about own problem solving. Using metalevels, NOOS can reason about preferences in order to make decisions about sets of alternatives present in domain knowledge and problem solving knowledg e. Reflection in NOOS is provided by inference processes that involve metalevels. Basic reflective capabilities include reasoning about alte rnative methods to solve a task, reasoning about what is known by the system itself, and reasoning about the existence of solutions. A forma l model of NOOS inference using Descriptive Dynamic Logic is also pres ented.