From: Smoothing approximation to the lower order exact penalty function for inequality constrained optimization