| Title |
A Study on a New Function Optimization Method Using Probabilistic Tabu Search Strategy |
| Authors |
김형수(Kim, Hyung-Su) ; 황기현(Hwang, Gi-Hyun) ; 박준호(Park, June-Ho) |
| Keywords |
Probabilistic Tabu Search ; Belief Space ; Function Optimization |
| Abstract |
In this paper, we propose a probabilistic tabu search strategy for function optimization. It is composed of two procedures, one is Basic search procedure that plays a role in local search, and the other is Restarting procedure that enables to diversify search region. In basic search procedure, we use Belief space and Near region to create neighbors. Belief space is made of high-rank neighbors to effectively restrict searching space, so it can improve searching time and local or global searching capability. When a solution is converged in a local area, Restarting procedure works to search other regions. In this time, we use Probabilistic Tabu Strategy(PTS) to adjust parameters such as a reducing rate, initial searching region etc., which makes enhance the performance of searching ability in various problems. In order to show the usefulness of the proposed method, the PTS is applied to the minimization problems such as De Jong functions, Ackley function, and Griewank functions etc., the results are compared with those of GA or EP. |