Title :
Which organizational knowledge is useful: rare, medium, or well-done?-Comparison of different levels of organizational knowledge in multiagent environments
Author :
Takadama, Keiki ; Terano, Takao ; Shimohara, Katsunori
Author_Institution :
ATR Int., Kyoto, Japan
Abstract :
Focuses on several levels of organizational knowledge and investigates the characteristics of each level in multiagent environments. A careful investigation of the characteristics has revealed the following implications: (1) a moderately structured level of organizational knowledge improves solutions in an individual evaluation and reduces computational costs in an organizational evaluation as a comparison when reusing roughly or well structured levels of organizational knowledge; and (2) the moderately structured level of organizational knowledge has great, potential for further improvement of solutions or computational costs even in strict cases where good solutions or small computational costs have been already found
Keywords :
learning (artificial intelligence); multi-agent systems; computational costs; individual evaluation; knowledge reuse; knowledge structured level; learning classifier system; multiagent environments; multiagent learning; organizational knowledge; Analytical models; Business; Computational efficiency; Computational modeling; Costs; Humans; Knowledge management; Problem-solving;
Conference_Titel :
Industrial Electronics Society, 2000. IECON 2000. 26th Annual Confjerence of the IEEE
Conference_Location :
Nagoya
Print_ISBN :
0-7803-6456-2
DOI :
10.1109/IECON.2000.972457