This study examines selected components of the optimization function, used to evaluate the optimal kinematic structure of a robot for a given task. Automated generation of the kinematic structure is based on scalable drive modules of the joints and modules of the carrying arms with a check on the permissible torque of the drive and bending moment of the carrier element. An optimization algorithm is used to generate variations of kinematic structures, the base requirement of the fitness function is the ability to traverse a given trajectory with a defined orientation of the tool. The suitability of a given kinematic structure is evaluated further by a set of evaluation functions such as a check for spatial collisions, energy consumption, minimization of total weight, minimization of degrees of freedom for a given task and several other criteria. Two of these criteria – evaluation of the total weight of a robotic arm with drives in joints and evaluation of power consumption for a defined handling task are examined here.