Local open-loop manipulation of multi-agent networks

The energy required for manipulation of a double-integrator network model via local external actuation is examined. Some previous results on the controllability Gramian, which specifies the required energy, are briefly reviewed. Three new results are then developed. First, the energy required for manipulation along the synchronization manifold over an arbitrary horizon is characterized. Second, several scalar measures that give a global indication of a network's manipulability are analyzed. Based on these measures, we study design of the DIN to prevent or facilitate manipulation.