Computer Science · Computer Networks and Communications
A Parallel Algorithm for NMNF Problems with a Large Number of Capacity Constraints
See on the mapAbstract
In this paper, we propose a converting technique based method to solve nonlinear multi-commodity network flow (NMNF) problems with a large number of capacity constraints and discuss the associated implementation. We have combined this method with a successive quadratic programming (SQP) method and a parallel dual-type (PDt) method possessing decomposition effects. We have tested our method in solving a kind of lattice-type network system examples of NMNF problems. The simulation results show that the proposed algorithm is efficient for solving NMNF problems and successfully handles a large number of coupling capacity constraints. Furthermore, the computational efficiency of the proposed algorithm is more significant while the numbers of capacity constraints are increased.