Multi-agent temporal-difference learning with linear function approximation: Weak convergence under time-varying network topologies | IEEE Conference Publication | IEEE Xplore