Communication/computation tradeoffs in wireless sensor networks: Comparing node-level and network-level strategies

In a typical wireless sensor network, data collected from sensors to be conveyed at the base station requires an energy efficient solution due to the scant battery power of nodes in order to extend the network lifetime. The amount of this data usually depends on the amount of local processing performed on nodes. There may be more local processing than communication on a node and vice versa to attain energy efficiency. This analysis can be examined at network-level where a single algorithm is employed by all nodes in a network or at node-level which provides flexibility for different nodes to implement different algorithms. To guide designers in effectively using these tradeoffs to prolong network lifetime at node-level strategy, we develop a novel mixed integer programming (MIP) framework. We show that node-level strategy can extend network lifetime up to 22.50% than the case where a single algorithm is employed at network-level. We also develop a polynomial time heuristic algorithm in order to reduce the computational complexity of the proposed MIP model. Maximum network lifetime could be obtained approximately with an error less than 1.29% with this method in very short times compared with the proposed MIP model.