An Indexing Framework for Adaptive Arrangement of Mechanics Problems for ITS

Tsukasa HIRASHIMA  Toshitada NIITSU  Kentaro HIROSE  Akihiro KASHIHARA  Jun'ichi TOYODA  

IEICE TRANSACTIONS on Information and Systems   Vol.E77-D   No.1   pp.19-26
Publication Date: 1994/01/25
Online ISSN: 
Print ISSN: 0916-8532
Type of Manuscript: Special Section PAPER (Special Issue on Intelligent CAI and Hypermedia)
ITS,  adaptive arrangement of problems,  indexing of problems,  mechanics problem,  

Full Text: PDF(691KB)>>
Buy this Article

This paper describes an indexing framework for adaptive arrangement of mechanics problems in ITS (Intelligent Tutoring System). There have been some studies for adaptive arrangement of problems in ITS. However, they only choose a solution method in order to characterize a problem used in the practice. Because their target domains have been sufficiently formalized, this kind of characterization has sufficed to describe the relations between any two problems of such a class. In other words, here, it is enough to make students understand only the solution methods for the given class of problems. However, in other domains, it is also important to understand concepts used in the problems and not only to understand solution methods. In mechanics problems, concepts such as mechanical objects, their attributes, and phenomena composed of the objects and the attributes also need to be taught. Therefore, the difference between solution methods applied is not sufficient to describe the difference between two given problems. To use this type of problems properly in the practice, it is necessary to propose an advanced new characterization framework. In this paper, we describe a mechanics problem with three components: (1) surface structure, (2) phenomenon structure, (3) solution structure. Surface structure describes surface features of a problem with mechanical objects, their configuration, and each object's attributes given or required in the problem. Phenomenon structure is described by attributes and operational relations among them included in the phenomenon specific to the surface structure. Solution structure is described by a sequence of operational relations which compute required attributes from given attributes. We call this characterizing indexing because we use it as index of each problem. This paper also describes an application of the indexing to arrangement of problems. We propose two mechanisms of control: (a) reordering of a problem sequence, and (b) simplifying of a problem. By now, we have implemented basic functions to realize the mechanisms except for the part of interface.