Enhancing adaptive random testing through partitioning by edge and centre