Show simple item record

dc.contributor.authorMeyer, Robert R.
dc.contributor.authorChristou, Ioannis T.
dc.date.accessioned2013-04-11T16:31:52Z
dc.date.available2013-04-11T16:31:52Z
dc.date.issued1995
dc.identifier.citation95-19en
dc.identifier.urihttp://digital.library.wisc.edu/1793/65317
dc.description.abstractWe present an efficient method for assigning any number of processors to tasks associated with the cells of a rectangular uniform grid. Load balancing equi-partition constraints are observed while approximately minimizing the total perimeter of the partition, which corresponds to the amount of interprocessor communication. This method is based upon decomposition of the problem size grows large in all parameters, he error bound associated with this feasible solution approaches zero. We also present computational results from a high level parallel Genetic Algorithm that utilizes this method, and make comparisons with other methods. On a network of workstations, our algorithm solves within minutes instances of the problem that would require one billion binary variables in a Quadratic Assignment formulation.en
dc.titleOptimal and Asymptotically Optimal Equi-partition of Rectangular Domains via Stripe Decompositionen
dc.typeTechnical Reporten


Files in this item

Thumbnail

This item appears in the following Collection(s)

  • Math Prog Technical Reports
    Math Prog Technical Reports Archive for the Department of Computer Sciences at the University of Wisconsin-Madison

Show simple item record