Approximating the Non-contiguous Multiple Organization Packing Problem
Résumé
We present in this paper a -approximation algorithm for scheduling rigid jobs on multi-organizations. For a given set of n jobs, the goal is to construct a schedule for N organizations (composed each of m identical processors) minimizing the maximum completion time (makespan). This algorithm runs in O(n(N + log(n))log(np max )), where p max is the maximum processing time of the jobs. It improves the best existing low cost approximation algorithms. Moreover, the proposed analysis can be extended to a more generic approach which suggests different job partitions that could lead to low cost approximation algorithms of ratio better than .