A cost transformation based routing mechanism and its applications on constraint-routing problems

In this paper, we discuss a cost transformation based routing mechanism. The idea is to transform any traffic engineering (TE) routing problem to a known (e.g. shortest path routing) problem, such that a solution to the known problem is a sub-optimal solution to the original. Instead of relying on human to "guess" the best transformation function for each practical problem; we developed a numerical search mechanism to automatically discover the function. So far, we have tested this mechanism on four routing problems-among them are some classic constraint-routing problems and an emerging dual-path routing problem-and the results are comparable or better than those from some well-known heuristics. More importantly, these example applications showcase how routing solutions are achieved with minimum human intervention.