Wireless Sensor Networks (WSNs) are applied in many time-critical applications, e.g., industrial automation and smart grid. This highlights the importance of Delay Constrained Relay Node Placement (DCRNP) problem that builds a path fulfilling a specified delay constraint between each sensor and the sink by using a minimum number of relays. Due to the NP-hardness of the DCRNP problem, in this paper, a polynomial time Subtree-and-Mergence-based Algorithm (SMA) is proposed to approximately solve the DCRNP problem. First, a shortest path tree rooted at the sink and connecting all sensors is built to check the feasibility of the DCRNP problem. If the DCRNP problem is feasible, then the paths of this tree are progressively merged at some relays, which are not limited to those relays lying in the originally the originally built shortest path tree, to save deployed relays while maintaining the obedience of delay constraints. With the repetition of this mergence, the number of deployed relays is gradually reduced. Furthermore, the approximation ratio and the time complexity of the proposed SMA are elaborately analyzed. Finally, extensive simulations are conducted to demonstrate the effectiveness of this work. Simulation results show that SMA can significantly save deployed relays comparing with existing algorithms.

This work was published on Mobile Networks and Applications,2017:1-13. titled Delay Constrained Relay Node Placement in Wireless Sensor Networks: A Subtree-and-Mergence-based Approach.