A network decomposition/aggregation procedure for a class of multicommodity transportation problems

An effective method of solving a practical class of multicommodity transportation problems is presented. The heuristic procedure combines theoretical results concerning multicommodity network structures and recent results on aggregation theory. Computational implementation has shown that problems with thousands of variables can be solved to within a few percent of optimality in a matter of seconds. Only widely available network codes are required to implement the procedure, thus making it attractive for smaller computers.