Unification of filled function and tunnelling function in global optimization (Q871668)

The authors consider unconstrained global optimization problems on finite dimensional spaces, where the objective function is continuously differentiable coercive function with bounded gradients and has a finite number of minimal function value. In this paper, two auxiliary functions for global optimization are proposed. These two functions possess all characters of tunneling functions and filled functions under certain general assumptions. Result of numerical experiments shows that such two auxiliary functions are effective.