Title :
A new type of intelligent navigation node of hypertext
Author_Institution :
Dept. of Math., Hangzhou Inst. of Commerce, Zeijiang, China
Abstract :
To solve the problem that is difficult for some current navigation devices to discover navigation knowledge automatically, this paper adds the mechanism of machine learning to hypertext, and proposes a new type of intelligent navigation node of hypertext, INNGRL, in which a genetic-based reinforcement learning system is embedded. In a class of process control systems with a hypertext structure, INNGRL can in real-time and automatically discover rules, optimize them, correct mistaken ones, and can evaluate the utility of these rules. In the simulation of a process control, it has discovered 130 rules. The test of the quality of these rules shows that the cost of the simulation process is lower than that of the practical process by 13.4%
Keywords :
hypermedia; learning (artificial intelligence); navigation; process control; software agents; INNGRL; genetic-based reinforcement learning system; hypertext; intelligent navigation node; machine learning; navigation devices; process control systems; Business; Computer architecture; Engines; Expert systems; Learning systems; Machine learning; Mathematics; Navigation; Process control; User interfaces;
Conference_Titel :
Intelligent Processing Systems, 1997. ICIPS '97. 1997 IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4253-4
DOI :
10.1109/ICIPS.1997.672963
