A method for control of discrete-time switched systems based on parameterized Lyapunov functions is presented. A polynomial representation of the switched system is formulated, and the method of moments is applied to obtain a relaxed model. The control problem is recast as an online receding horizon optimization problem where the Lyapunov function parameters and moment sequences are the decision variables. A novel procedure to synthesize the switching control signal from the recovered measure is described. The method is illustrated in simulation for the case of a pulse-width modulated multicellular converter, showing good reference tracking results.