Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Algorithms for parallel processor scheduling

Thesis/Dissertation ·
OSTI ID:5897371
This dissertation discusses systems with n independent jobs to be assigned to m identical processors. Two problems with different criteria are studied. The first problem deals with finding the bounds for the makespans of schedules based on the Longest Processing Time first algorithm. Both deterministic and stochastic models are investigated. Worst case analysis is performed for the deterministic models in which the job processing times are assumed to be known deterministically. A bound is obtained for the expected makespan of the schedules in the stochastic models where the processing times are assumed to be uniformly distributed in the interval (0,1). Furthermore, an improved expected bound is developed by approximation. The second problem is focused on minimizing the mean flow time of n jobs on two identical parallel processors where one processor has limited time available. Two mathematical models are presented for obtaining the optimal schedules. A heuristic is developed in order to achieve better computational efficiency than the mathematical models. The worst case analysis for the heuristic is also performed.
Research Organization:
Rutgers--the State Univ., New Brunswick, NJ (USA)
OSTI ID:
5897371
Country of Publication:
United States
Language:
English

Similar Records

Stochastic scheduling of parallel processors
Thesis/Dissertation · Mon Dec 31 23:00:00 EST 1984 · OSTI ID:6891504

Tighter bounds for the multifit processor scheduling algorithm
Journal Article · Tue Jan 31 23:00:00 EST 1984 · SIAM J. Comput.; (United States) · OSTI ID:5108217

Bounds for multifit scheduling on uniform processors
Journal Article · Mon Jan 31 23:00:00 EST 1983 · SIAM J. Comput.; (United States) · OSTI ID:6331025