Iterative Learning Control of Stochastic Multi-Agent Systems with Variable Reference Trajectory and Topology