Parallel computing model for problem solver towards ITSs: Epistemological articulation of human problem solving

Abstract

Study on building powerful and robust problem solver for ITSs is described. The proposed problem solver is based on parallel computing technology to simulate some of the intelligent characteristics of human problem solving including the concurrency of thought, the case based reasoning, and the awareness of the key to problem solving. We propose the multiagent problem solver which divides a given problem into several small sub-problems and solves them concurrently. This system is especially useful for ITS which teaches procedural problem solving to novice students, because it is capable of showing the reason of achieving an idea for solution by thinking aloud by the “trial-and-error” method.