Wireless Sensor Network Lifespan Optimization with Simple, Rotated, Order and Modified Partially Matched Crossover Genetic Algorithms

This paper proposes a genetic algorithm (GA) with a benchmarking study for optimizing the wireless sensor network (WSN) lifespan. Four crossover operators combined with four mutation operators were developed to enhance the GA efficiency, and thus the performance of the lifespan optimization algorithm. The traditional one-point crossover operator referred as the "simple crossover", is used to evaluate the modified "partially matched" and the "order" crossovers. Also, a new crossover operator referred as "rotated" crossover is also proposed and evaluated. Different combinations of these crossover operators with the one-point and two-points deterministic and random mutations are used to optimize the WSNs lifespan. The algorithms were coded in C programming language and applied to different instances of WSNs initial configurations. The optimization software tool developed based on the combinatorial operators allows selecting the best solution among 16 through a smart decision making. For illustration, among all the investigated algorithms, the modified partially matched crossover associated with the random two-points mutation has shown the best performances on the studied instance, due to its capability to reach closer to the optimal solution.