A multi-agent system, conformed by differential mobile robots, can execute many tasks. An example is the positioning of mobile robots in order to represent a 2D formation on the ground that represents a geometric shape. In this application, the mobile robots move to a position assigned by an algorithm. This algorithm assigns the 2D position that each mobile robot must achieve. This paper presents a methodology that defines seven criteria necessaries to compute a good 2D position assignation. Those criteria are based on the unweighted Euclidean distances between the robot's location and the desired position. They are also evaluated in order to select the criterion that allows all robots in the multi-agent system to move the shortest distance when performing a geometric representation. This means a reduction in the length of the path required by each mobile robot to complete the assigned task.