Abstract:
For solving global optimization problems, a new algorithm,
which is called Fractal Algorithm, is presented. Feasible
region is partitioned by fractal combining with golden
section. Bad region is deleted gradually and finally optimal
solution is remained. The full advantage of the local
fine structure of fractal and the quick convergence of
golden section method were taken. Hence it is high efficient
and high speedy. The algorithm has following character:
strong adaptability, adapting to a class of complex function.
It needs only that the object function has one order
derivative. The minimum can be found at any precision at
which a computer can work. On the other hand, this method
needs so little memory that it almost can be implemented on
any personal computer and its efficiency is not almost influenced.
The proof of convergence of the algorithm is given.
The illustrations show the algorithm is effective.
