A soft approach for hard continuous optimization

This paper is to introduce a soft approach for solving continuous optimization models where seeking an optimal solution is theoretically or practically impossible. We first review methods for solving continuous optimization models, and argue that only a few optimization models with some good structure are solved. To solve a larger class of optimization problems, we suggest a soft approach by softening the goal in solving a model, and propose a two-stage process for implementing the soft approach. Furthermore, we offer an algorithm for solving optimization models with a convex feasible set, and verify the validity of the soft approach with numerical experiments.