Multi-agent iterative learning control with communication topologies dynamically changing in two directions

This study aims to develop an iterative learning control (ILC) approach to solving finite-time output consensus problems of multi-agent systems. The communication topologies among agents are considered to dynamically change in two directions (along both time axis and iteration axis), for which a framework is presented to construct effective distributed protocols. It is shown that a protocol can be derived through ILC to enable multi-agent systems to accomplish the finite-time consensus, which moreover can possess an exponentially fast convergence speed. In particular, for any desired terminal output that is available to not all of but only a portion of agents, multi-agent systems can be guaranteed to achieve the finite-time consensus at the desired terminal output. Simulation tests are given to demonstrate the performance and effectiveness of the obtained consensus results.