Incorporating the stretching function technique into simulated annealing timely to accelerate global optimization convergence

Stretching function technique is one recently proposed approach to make use of already obtained information to accelerate convergence and improve success rate of an algorithm for global optimization (GOP). Simulated annealing algorithm (SA) is a well known stochastic method for GOP. Incorporating the "Stretching" technique into SA timely to form an effective and robust algorithm for GOP is very important to users. Two typically different manners to add "Stretching" technique in SA were selected and investigated in detail. Comparisons were reported to assess the actual performance of the two different selections. Numerical results of test on 8 benchmark problems demonstrate that combining "Stretching" technique and SA reasonably can greatly accelerate the convergence and significantly improve the success rate of finding a global minimum.