“…The analytical results are illustrated by carrying out Monte-Carlo simulations, which were executed over 10 5 independent trials. Unless stated otherwise, the noise power at all nodes are set as σ 2 = 0 dBw, the path-loss exponent m = 2.7, η = 0.8 [35] For illustration purposes, we consider a 2D system topology where the source, the users and the eavesdropper are located at (x ds , y ds ) = (0, 0), (x du1 , y du1 ) = (0, 2), (x du2 , y du2 ) = (1.5, 0), (x dM , y dM ) = (1, 0), and (x de , y de ) = (0, 1), respectively, and the relays are located at (x dRe1 , y dRe1 ) = (1, 0.5), (x dRe2 , y dRe2 ) = (1.5, 0.5), (x dRe3 , y dRe3 ) = (1, 1.5) and (x dRe4 , y dRe4 ) = (1.5, 1.5). Due to the symmetry between the strongest user and the eavesdropper paths, this topology can explain the benefit of the relay selection schemes.…”