Parallel randomized search for distributed memory machines:

Abstract: "For the computation on parallel machines, each processor should be given even computational load. Generally speaking, there are two types of load balancing methods: dynamic ones and static ones. Static methods need the estimation of the task size and the communication cost, which is...

Full description

Saved in:
Bibliographic Details
Main Authors: Iwayama, Noboru (Author), Satoh, Ken 1959- (Author)
Format: Book
Language:English
Published: Tokyo, Japan 1991
Series:Shin-Sedai-Konpyūta-Gijutsu-Kaihatsu-Kikō <Tōkyō>: ICOT technical report 696
Subjects:
Summary:Abstract: "For the computation on parallel machines, each processor should be given even computational load. Generally speaking, there are two types of load balancing methods: dynamic ones and static ones. Static methods need the estimation of the task size and the communication cost, which is difficult in most cases. For Dynamic methods, communication problems should be resolved for efficient computing especialy [sic] on distributed memory machines. We give a different type of load balancing method for problems searching a single solution on distributed memory machines. That is a parallelization of randomized search
Our method requires no communication during computation, and all processors are always busy. This parallelization on distributed memory machines is profitable for the problems which the above methods don't handle well.
Physical Description:8 S.

There is no print copy available.

Interlibrary loan Place Request Caution: Not in THWS collection!