A Bipopulation-Based Evolutionary Algorithm for Solving Full Area Coverage Problems

The full area coverage sensor deployment problem is a challenging issue in wireless sensor networks. We focus on sensor deployment strategies that aim to acquire a full-coverage state with a minimum number of sensors in a predetermined target region that includes non-penetrable obstacles. This paper presents an efficient bipopulation-based evolutionary full area coverage (BEFAC) algorithm that involves a bipopulation structure composed of a full- and partial-coverage populations. Fitness functions, stochastic unary recombination operators, and selection procedures between the two populations are well designed. Through applying the proposed BEFAC, a full-coverage state is acquired with a minimum number of deployed sensors in the target region, which has non-penetrable obstacles, and the algorithm avoids getting caught in local minima. The performance results reveal that BEFAC outperforms the conventional deployment methods in terms of the number of deployed sensors and the number of required fitness evaluations.