| Summary: | 碩士 === 國立高雄第一科技大學 === 電腦與通訊工程所 === 96 === PC clusters have recently received much attention as cost-effective
parallel platforms for scientific computations. A parallel program, which
can be executed on a target cluster system, generally consists of a set of
tasks (i.e. program segments). To effectively harness the computing
power of the target cluster system, techniques for task matching and
scheduling becomes vital important.
Task matching and scheduling is extremely complex and is known
to be NP-complete in the strong sense. Consequently, this thesis
presents a constructive-oriented iterative algorithm, which is based on
the primary principles of ant colony optimization (ACO), to acquire
near-optimal solutions of the task matching and scheduling problem
within a reasonable amount of computation time. The proposed
algorithm concentrates on properly allocating the tasks to the processing
elements of the cluster system and sequencing the execution of the
tasks. The main characteristic of this algorithm is the use of an
alternatively forward–backward traversing mechanism for guiding
artificial ants to systematically search feasible schedules. For the sake of
improving the search efficiency, moreover, bounding techniques are also
incorporated into the traversing mechanism. The performance of the
proposed algorithm is evaluated by comparing it against existing
techniques, such as DPS, ACO-TMS and genetic algorithms (GAs), in
terms of overall completion time for a set of problem instances.
Experimental results indicate that the algorithm proposed here is a
significant improvement compared with other approaches.
|
|---|
