A Penalty Approach for Generalized Nash Equilibrium Problem