Differential annealing for global optimization

This paper propose a hybrid stochastic approach called differential annealing algorithm. The algorithm integrated the advantages of differential evolution and simulated annealing. It can be considered as a swarm-based simulated annealing with differential operator or differential evolution with the Boltzmann-type selection operator. The proposed algorithm is tested on benchmark functions, along with simulated annealing and differential evolution. Results show that differential annealing outperforms the comparative group under the same amount of function evaluations.