With the developments and applications of wireless communications, more and more applications require advanced radio transmission technologies (RTT) to reach the goal of high power and spectrum efficiencies as well as flexibility and adaptation to multiple scenarios such as mobile broadband, ultra reliable communications, the internet of things, etc. Recently, intelligent optimization and self-learning algorithms have been widely studied as potential solutions. Among the latter, the vey promising evolution algorithms aim to find the optimal operating point of complex non-continuous cost functions using biologically inspired techniques such as genetic algorithms and particle swarm optimization. Self-learning algorithms are lighted up with the success of machine learning in the artificial intelligence field. Given the strong requirements expected from RTT and the fruitful achievements in evolution and self-learning algorithms (ESLA), it is foreseen that applying ESLA to RTT may solve some of the most daunting challenges in wireless communications.