Optimal Task Assignment in Heterogeneous Distributed Computing Systems

  • Authors:
  • Muhammad Kafil;Ishfaq Ahmad

  • Affiliations:
  • -;-

  • Venue:
  • IEEE Concurrency
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

A distributed system comprising networked heterogeneous processors requires an efficient tasks-to-processor assignment to achieve a fast turnaround time. This problem is known to be NP-hard, except in a few cases. Although the literature offers heuristics that can yield suboptimal solutions in a reasonable amount of time, little is available for optimal solutions of moderate-sized problems. In this article, the authors develop techniques for optimal solutions for assigning arbitrary task graphs to arbitrary processor networks. The basis of their research is the best-first search technique known as the A* algorithm, from the area of artificial intelligence. The algorithm, in its traditional form, guarantees optimal solutions, but does not work for large problems, due to its high time and space complexity. The authors propose algorithms based on the A* technique that are considerably faster and memory efficient, and still yield optimal solutions.