| Abstract: | Genetic algorithms (GA) are applicable to many kinds of difficult problems. When a population keeps enough diversity and similarity, GA can obtain good solutions quickly. However, because these often compete with each other, it is difficult to fulfill both of these conditions simultaneously. In this article, taking these into consideration, we propose a new GA for the floorplan design problem, and aimed at improving the efficiency of calculation, the maintenance of the solution’s population diversity, and reduction of the number of parameters. We applied it to two MCNC (originally established as the Microelectronics Center of North Carolina) benchmark problems. The experimental results showed that the proposed method performed better than the existing methods. This work was presented, in part, at the 8th International Symposium on Artificial Life and Robotics, Oita, Japan, January 24#x2013;26, 2003 |
