Simulated Annealing Algorithms for Continuous Global Optimization: Convergence Conditions

In this paper, simulated annealing algorithms for continuous global optimization are considered. After a review of recent convergence results from the literature, a class of algorithms is presented for which strong convergence results can be proved without introducing assumptions which are too restrictive. The main idea of the paper is that of relating both the temperature value and the support dimension of the next candidate point, so that they are small at points with function value close to the current record and bounded away from zero otherwise.