Reinforcement learning is a significant part of machine learning algorithms, strategies and methods. It is useful especially when machine learning is necessary and no model is available. In case of automated electric installation systems, Artificial Intelligence can be integrated for autonomous actions of the installation minimizing user intervention. For this case, reinforcement-learning algorithms are best suited. In this paper, we first shortly describe the theory of reinforcement learning. We then focus on questions that arise when such an integration is needed, e.g. selection of an algorithm, the algorithm parameters, stability or convergence. In our research, we implement a customized reinforcement algorithm into automated installation system and we verify it in the experiments. We have made a series of experiments with parameters of the installation set. (Abstract).