DocumentCode :
977151
Title :
The parallel assignment problem redefined
Author :
May, C.
Author_Institution :
IBM Thomas J. Watson Res. Center, Yorktown Heights, NY
Volume :
15
Issue :
6
fYear :
1989
fDate :
6/1/1989 12:00:00 AM
Firstpage :
821
Lastpage :
824
Abstract :
The parallel assignment problem is slightly redefined using a subtler cost function that tends to reduce the number of extra assignments required. It is shown that the new problem, like the classical, is NP-hard. The new problem is then solved for the restricted case of assignment from invertible functions of single variables. For this restricted case and optimum solution can be found in linear time for both the classical problem and the new problem. However, the number of extra assignments required for the classical problem is equal to the number of cycles in the dependency graph, while in the new problem it is equal to the number of isolated cycles in the dependency graph which may be less
Keywords :
computational complexity; parallel algorithms; parallel programming; NP-hard; classical problem; cost function; dependency graph; invertible functions; isolated cycles; linear time; optimum solution; parallel assignment problem; restricted case; single variables; Artificial intelligence; Cost function; Couplings;
fLanguage :
English
Journal_Title :
Software Engineering, IEEE Transactions on
Publisher :
ieee
ISSN :
0098-5589
Type :
jour
DOI :
10.1109/32.24735
Filename :
24735
Link To Document :
بازگشت