DID YOU KNOW? DTIC has over 3.5 million final reports on DoD funded research, development, test, and evaluation activities available to our registered users. Click HERE
to register or log in.
Parallel Computing in Optimization.
COLORADO UNIV AT BOULDER DEPT OF COMPUTER SCIENCE
Pagination or Media Count:
One of the major developments in computing in recent years has been the introduction of a variety of parallel computers, and the development of algorithms that effectively utilize their capabilities. Very little of this parallel algorithm development, however, has been in numerical optimization. Nevertheless, significant opportunities exist for the utilization of parallelism in optimization, especially on computers that support independent concurrent processes. This paper first gives a very brief survey of parallel architectures and general characteristics of parallel algorithms. Next we indicate what we see as the leading opportunities for the utilization of parallelism in optimization. Then we survey the small amount of existing research in parallel optimization most of this has been conducted at The Hatfield Polytechnic. Finally we discuss some recently initiated research at the University of Colorado concerned with solving optimization problems by parallel algorithms suitable for implementation on a local area network of computers we focus on a new parallel algorithm for global optimization. Additional keywords Computer architecture, and Multiprocessors.
APPROVED FOR PUBLIC RELEASE