First-order, networked control models of swarming silkworm moths

Social insects have long served as inspiration to the multi-agent community. In this paper, we take the opposite approach and see if tools from decentralized, networked control can be used to predict observed, biological behaviors. In particular, we study the silkworm moth, the Bombyx Mori, and we model these moths as first-order networks in which the male- male interactions are defined through a proximity graph. The male-female interactions are given by a broadcast protocol in which the females that are releasing pheromones are visible to all the males. Using barrier certificates, the resulting, switched network is analyzed and it is shown that the males are attracted to and trapped in a region defined by the female moths, as is the case in actual silkworm moths as well.