DocumentCode :
3163370
Title :
The parallel terraced scan: an optimization for an agent-oriented architecture
Author :
Rehling, John ; Hofstadter, Douglas
Author_Institution :
Center for Res. on Concepts & Cognition, Indiana Univ., Bloomington, IN, USA
Volume :
1
fYear :
1997
fDate :
28-31 Oct 1997
Firstpage :
900
Abstract :
A relatively new area of research concerns computer programs which operate by the action of many small agents, rather than the serial execution of an algorithm. Because they differ drastically from conventional programs, new approaches are required for developing and maintaining them. An optimization technique, the parallel terraced scan, has been applied to one such program: a letter-recognition system called the Examiner. The parallel terraced scan is the exploration of many possible paths, but with more computation devoted to paths which are identified as being more promising. Thus, it resembles pruning of search trees in some respects, but it does not completely abandon paths which are tentatively identified as less fruitful
Keywords :
character recognition; optimisation; parallel architectures; software agents; Examiner; agent-oriented architecture; computer programs; letter-recognition system; parallel terraced scan; search trees; Character recognition; Cognition; Concurrent computing; Data structures; Humans; Joining processes; Optical character recognition software; Optical network units;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Processing Systems, 1997. ICIPS '97. 1997 IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-4253-4
Type :
conf
DOI :
10.1109/ICIPS.1997.672938
Filename :
672938
Link To Document :
بازگشت