Parallel Computation of Large-Scale Nonlinear Network Problems in the Social and Economic Sciences
Name
OR-221-90.pdf
Size
958.79 KB
Format
Adobe PDF
Checksum (MD5)
6b6b77da04d4a0fef709a7f7e233737a
Author(s)
Nagurney, Anna
Kim, Dae-Shik
Date Issued
July 1990
Publisher
Massachusetts Institute of Technology, Operations Research Center
Series/Report no.
Operations Research Center Working Paper;OR 221-90
Abstract
In this paper we focus on the parallel computation of large - scale equilibrium and optimization problems arising in the social and economic sciences. In particular, we consider problems which can be visualized and conceptualized as nonlinear network flow problems. The underlying network structure is then exploited in the development of parallel decomposition algorithms. We first consider market equilibrium problems, both dynamic and static, which are formulated as variational inequality problems, and for which we propose parallel decomposition algorithms by time period and by commodity, respectively. We then turn to the parallel computation of large-scale constrained matrix problems which are formulated as optimization problems and discuss the results of parallel decomposition by row/column.
MIT Department
Massachusetts Institute of Technology. Operations Research Center
Persistent DSpace Link
