PDF pending restoration
ERIC Number: ED232856
Record Type: Non-Journal
Publication Date: 1982
Reference Count: N/A
Cognitive Processes and Knowledge Structures Used in Solving Physics Problems. Final Technical Report.
Novak, Gordon S., Jr.
Cognitive processes and knowledge structures used in physics problem-solving at the high school or freshmen college level were investigated by analysis of expert and novice human problem-solving behavior and by development of computer programs which can solve informally-stated physics problems. Computer representations for the information developed by a problem solver while solving a problem, a computer language for specifying such representations (GLISP), and a compiler for translating programs written in terms of such representations into executable form were developed. In addition, a set of rules for problem analysis were also developed. These rules determine the general context of a problem, fill in features missing from the problem statement, and connect the features of the problem with the appropriate object models and methods for solving the problem. These representations and rules have been used successfully to analyze mechanics problems. Current research involves working on methods by which a program, starting with general formulations of physical laws, can learn expertise in solving particular kinds of problems through practice. (An example of knowledge representation for units of measurement and an example of problems and problem analysis rules are appended to this report). (Author/JN)
Publication Type: Reports - Research
Education Level: N/A
Sponsor: National Science Foundation, Washington, DC.
Authoring Institution: Texas Univ., Austin. Dept. of Computer Science.