A variable-barrier-function-based prescribed finite-time bounded-H∞ reinforcement learning optimal tracking control strategy without dependence on initial condition.