Designing Social Force: Control for Collective Behavior of Learning Individuals
Keiki Takadama, Katsunori Shimohara · The MIT Press eBooks · 2002
This paper explores methods for controlling the collective behavior of learning individuals (e.g., robots, agents) to achieve goals at the collective level while these individuals pursue their own goals at the individual level. Toward this enormous goal, we start by categorizing the four levels for controlling individuals in terms of both direct and indirect control, and then investigate the results at each control level to clarify the extent to which degree we should force the learning individuals to achieve goals at the collective level.