A multi-objective memetic algorithm for relay node placement in wireless sensor network

This paper proposes a multi-objective memetic algorithm, namely Mem-NSGA-II, to optimize the relay node placement problem. The network lifetime and the number of relay nodes are two objectives to be optimized. In Mem-NSGA-II, three new local search (LS) operations are designed and incorporated into the fast non-dominated genetic algorithm II (NSGA-II). The first LS operation inserts relay nodes into solutions for extending the network lifetime. The second LS operation aims to reduce the number of relay nodes to save cost. The third LS operation fine-tunes positions of the relay nodes for finding better solutions. The performance of Mem-NSGA-II is compared with a deterministic two-phased method and NSGA-II. Simulation results on five networks reveal that Mem-NSGA-II yields much better performance than the two algorithms.