In this paper, we present a framework for non-linear control of a swarm of agents based on the artificial potential characterized by attractive and repulsive properties. In this context, the swarm is able to reach a configuration and to maintain it, while migrating as a group and avoiding collisions among agents. Therefore, the behaviors of the swarm system proposed in this study are group migration and configuration, including collision avoidance capabilities. Different potentials expressions are evaluated and one proposed, in order to determine how quickly the swarm converges to a desired direction and velocity, and how robust the swarm is against collisions among the agents. Furthermore we provide two metrics that estimate which potential is the best one in a certain scenario. One quantifies how quickly the swarm converges to the given velocity, and the second evaluates how robust the potential is against collisions. Numerical simulations are included for verification purposes.