Global optimization in problems with uncertainties: the gamma algorithm. (Q1416293)

A new concept of global optimality is proposed for an uncertain function viewed as unspecified point values of set-valued functions. For the class of interval-valued functions, the optimal interval value is defined leading to double optimization with respect to upper/lower closures of such functions. A monadic iterative algorithm is then developed yielding approximate solutions up to the precision specified in advance. A notion of piecewise continuous function \(f: \mathbb{R}^n\to\mathbb{R}\) is defined and the method is generalized for an uncertain function over a bounded set \(X\subset\mathbb{R}^n\) defined by a closed set-valued function with piecewise Lipschitz continuous upper and lower functions.