Revisiting non-tree routing for maximum lifetime data gathering in wireless sensor networks

摘要

Wireless sensor networks usually adopt a tree structure for routing, where each node sends and forwards messages to its parent. However, lifetime maximization with tree routing structure is NP-hard, and all algorithms attempting to find the optimal solution run in exponential time unless P=NP. This paper revisits the problem of non-tree routing structure, where a node can send different messages to different neighbors. Though lifetime maximization with non-tree routing can be solved in polynomial time, the existing method transforms it into a series of maximum flow problems, which are either complicated or with high running time. This paper proposes an algorithm with O(mn) running time, where m is the number of edges and n is the number of nodes. The heart of the algorithm is a method to find a routing path from any node to the sink in O(m) time without disconnecting existing routing paths. The proposed algorithm is also suitable for distributed implementation. When a node fails, each influenced node can establish a new routing path in O(m) time. Simulations are conducted to compare the optimal lifetimes of tree structure and non-tree structure on random networks. The results verify the effectiveness of the proposed algorithm.