| Article ID: | iaor1992785 |
| Country: | Japan |
| Volume: | E-73 |
| Issue: | 7 |
| Start Page Number: | 1207 |
| End Page Number: | 1214 |
| Publication Date: | Jul 1990 |
| Journal: | Transactions of the Institute of Electronics, Information and Communication Engineers |
| Authors: | Kim Chonggun, Kameda Hisao |
| Keywords: | performance, scheduling, communications, lagrange multipliers, optimization, queues: theory |
Optimal static load balancing of multi-class jobs in a distributed computer system model is considered. This model is an extension of the Tantawi and Towsley single job class model to a multiple job class model. Some properties of the optimal solution are shown. On the basis of these properties, a straight-forward and efficient algorithm for optimal load balancing of multi-class jobs is derived. The authors compare the performance of the present algorithm and two other well known algorithms for multi-class jobs, the flow deviation algorithm and the Dafermos algorithm. The algorithm and the flow deviation algorithm both require a comparable amount of storage that is far less than that required by the Dafermos algorithm. In the results of numerical experiments the present algorithm and the Dafermos algorithm required mutually comparable computation times for obtaining the optimal solution which were far less than that of the flow deviation algorithm.