A PROJECTIVE METHOD FOR BIPARTITE NETWORKS AND APPLICATION TO THE MATRIX ESTIMATION AND TRANSPORTATION PROBLEMS
This paper deals with bipartite network problems where some objective function is to be optimized under flow conservation constraints at nodes, and, eventually, bounds on the link flows. Such problems are also interpreted as matrix optimization problems. In this study, a linear complexity projector was first constructed onto the feasible space defined by flow conservation constraints. Then, a variant of the scaled conjugate gradient method was developed, where the conjugate directions were all forced to lie in the feasible space; projections that were required at each iteration of this projected scaled conjugate gradient algorithm, are performed by the efficient projector developed earlier. Numerical results support superiority of this approach against classical conjugate gradient and LDL decomposition techniques for solving the problem. (A)
-
Corporate Authors:
CENTRE DE RECHERCHE SUR LES TRANSPORTS. UNIVERSITE DE MONTREAL
C.P. 6128, SUCCURSALE A
MONTREAL, QUEBEC Canada H3C 3J7 -
Authors:
- DRISSI-KAITOUNI, O
- Publication Date: 1991-4
Language
- English
Media Info
- Pagination: 22 p.
-
Serial:
- CENTRE DE RECHERCHE SUR LES TRANSPORTS PUBLICATION
- Issue Number: 766
- Publisher: Universite de Montreal
Subject/Index Terms
- TRT Terms: Calculation; Costs; Demand; Economics; Freight transportation; Location; Mathematical models; Methodology; Optimization; Origin and destination
- Uncontrolled Terms: Optimum
- ITRD Terms: 6464: Calculation; 224: Cost; 285: Demand (econ); 255: Economics; 741: Goods traffic; 9061: Location; 6473: Mathematical model; 9102: Method; 687: Origin destination traffic
- Subject Areas: Economics; Finance; Freight Transportation;
Filing Info
- Accession Number: 00726421
- Record Type: Publication
- Source Agency: Transportation Association of Canada (TAC)
- Files: ITRD, ATRI
- Created Date: Oct 28 1996 12:00AM