Advanced Smoothing Techniques For Penalty-Based Optimization

Abstract

This manuscript presents a novel approach for enhancing the effectiveness of objective penalty functions involving inequality constraints. The introduction of a penalty function that lacks smoothness is addressed by applying a novel smoothing technique to ensure its smoothness. The paper analyses the error estimates for both the original and smoothed problems. To verify practicality of the proposed approach, some numerical cases are solved.