Nondifferentiable optimization via smooth approximation: General analytical approach (Q1207839)

The authors present a method for nondifferentiable optimization, based on smoothing techniques by convolution. It is shown that these techniques preserve such useful properties of the original function as convexity and continuous differentiability. It is demonstrated that the smoothed functions are convenient for implementation on computers. Applications of these smoothing techniques to optimization of min-max problems are discussed.