The evolutionary algorithms and the simulated annealing are frequently used for the optimization. The evolutionary algorithms modeled their mechanism on the evolution of nature -“survival of the fittest”. That is, the evolutionary algorithms’ main concept is competition among individuals, while the simulated annealing has a different concept and a different history. The simulated annealing imitated annealing of an object. Proper annealing gives the object the crystalline one, which means the hardness.
The proposed algorithm used features of both the evolution algorithms and the simulated annealing-“competition and annealing”. The basic structure of the proposed algorithm follows the evolutionary algorithms, and at each generation stage of the offspring, a variable of temperature is used for annealing.
This proposed algorithm has two merits. The first merit can be found at the form of the selection. Using the temperature at the selection probability, the relation between the diversity and the crowdedness of the individuals can be controlled and this mechanism can avoid the pre-mature convergence. The second merit is that the proposed algorithm endows a competition ability to the parallel independent algorithm, such as parallel simulated annealing. This competition ability helps the algorithm to find more precise solution and to converge fast.
For testing the performance of the proposed algorithm, various test problems are used in the simulation. At these tests, the proposed algorithm shows better performance than either of the evolutionary algorithms and the simulated annealing.