This paper presents an Iterative Learning Control (ILC) algorithm for iterative parameter update in a semi-active system. The ILC law is designed to minimize a cost function, for example, the mean squared tracking error. First, a parametrized lifted domain representation of a linear parameter-varying system is developed explicitly. Based on this lifted domain representation and a cost function, gradient-based laws for the parameter update from iteration to iteration are proposed. Stability, monotonicity, steady state error, and robustness properties of these algorithms are presented. Finally, an application of the proposed algorithm is illustrated through the simulation of a plastic blow molding system.