Optimal queries processing in a heterogeneous sensor network using multicommodity network flow method

In a wireless sensor network, it is a common practice to have in network partial data processing. As part of this, the mitigation of the demands issued by sink nodes in a heterogeneous resource constrained sensor network requires optimal resource sharing, which is a very challenging task. We find this task analogous to finding the optimal flow in a multi commodity flow network. The minimum cost multi commodity flow problem is an NP hard one. In our paper, we give an optimal cost algorithm to solve this problem where a special data structure is built, which we call Data Forwarding Tree. We compare our algorithm to a well known auction method where a commodity wins the bid based on the lowest cost path, and the other looser commodities are shipped using more expensive paths incurring more cost. An algorithmic analysis shows that our method is free of deadlock as well.