Metropolis Versus Simulated Annealing and the Black-Box-Complexity of Optimization Problems
Ingo Wegener · 2008
Many real-world optimization problems cannot be modeled by a well-described objective function to apply methods from mathematical optimization theory. Then randomized search heuristics are applied - often with good success. Although heuristical by nature, they are algorithms and can be analyzed like all randomized algorithms, at least in principle. Two fundamental results of this kind are presented to show how such a theory can be developed.