Title :
Communication issues in parallel and distributed optimization algorithms
Author :
Bertsekas, Dimitri P.
Author_Institution :
Lab. for Inf. & Decision Syst., MIT, Cambridge, MA, USA
Abstract :
Summary form only given, as follows. Massive speedups through parallel optimization methods will not likely be achieved without efficient communication between processors. Three related issues are discussed: optimization algorithms that can take advantage of special communication-efficient hardware such as processor arrays; optimal routing algorithms for linear algebra calculations in message-passing architectures; and convergence theory of asynchronous algorithms
Keywords :
convergence of numerical methods; optimisation; parallel algorithms; asynchronous algorithms; communication-efficient hardware; convergence theory; distributed optimization algorithms; linear algebra calculations; message-passing architectures; parallel algorithms; processor arrays; routing algorithms; Books; Computer architecture; Concurrent computing; Convergence; Distributed computing; Hardware; Linear algebra; Optimization methods; Routing;
Conference_Titel :
Decision and Control, 1988., Proceedings of the 27th IEEE Conference on
Conference_Location :
Austin, TX
DOI :
10.1109/CDC.1988.194564
