; R o o t E n t r y F Tz C| C o m p O b j b W o r d D o c u m e n t ^ O b j e c t P o o l Tz C| Tz C| 4 @ = > ? @ A B C D E F G F Microsoft Word 6.0 Document MSWordDoc Word.Document.6 ; DMS-9527124 Trotter Advanced Scientific Computing Centers This award covers a broad research program in the area of computational optimization. The central thrust of the research addresses real-life, large-scale mixed-inter programming problems arising from important industrial application areas, such as manufacturing, telecommunications, and distribution. These problems are of vast importance towards the nation s present and future economic and technological development. The problems are also extremely intractable, both from the viewpoint of practical computation, as well as in terms of theoretical understanding, and they will only become larger and more difficult in future applications. The computational and methodological technology for solving integer programming models has made steady advances over the last thirty years, often following improvements in computing machinery as much as new theoretical insights. We are now at the starting point of a major quantum improvement in the technology. Not only are there several new and promising methodological venues available, but we are witnessin g a revolution in the computing world, in particular involving parallel computing and operating systems, with profound implications for science and engineering. It is almost certain that within a few years all computers will have multiple CPUs, and that all major scientific and engineering computing will be done on clusters of multiple-CPU machines. As supercomputing becomes redefined, a push to parallelize algorithms is starting now, and the end result will be software of extraordinary complexity by today s standards. However, it is fair to say that the field of parallel integer programming is, at best, in its infancy. The research seeks to propel computational integer programming into the realm of parallel computing, with the objective of effectively handling difficult real-world problems. The objective will require a multidisciplinary approach. Our work will seek to extend and develop solution methodologies for integer programming, in particular involving cutting planes, column generation, and branch-and-cut techniques. This will be integrated with computational theory techniques, especially the design of efficient data structures for parallel integer programming. Finally, there will be a large component of software engineering, necessary for the implementation of efficient parallel programs. This work combines the scientific and computational resources of three participating organizations, Columbia University, Cornell University, and the IBM T.J. Watson Research Center, to create a collaborative research environment to attain the goals outlined above. Among the many important advantages of such a close relationship between university computing research groups and that of a major hardware producer is the focus on practical ends provided by the industrial component. Oh +' 0 $ H l D h R:WWUSERTEMPLATENORMAL.DOT S u m m a r y I n f o r m a t i o n ( < DMS-9527124 Jocelyn J. Whitehurst Jocelyn J. Whitehurst @ " C| @ @ " C| @ F # Microsoft Word 6.0 1 ; e = e r ^ r l l l l l l l 1 r T I l l l l l l l l l 9 DMS-9527124 Trotter Advanced Scientific Computing Centers T
