Thermal Layout Optimization of Stacked Chip Based on Ant Colony Algorithm

Abstract

In this paper, an ant colony algorithm based on mutual information similarity is proposed to solve the stacked chips on the thermal layout optimization. In order to express the mutual information entropy between the optimal path and walking path, a new similarity impact factor is added in the algorithm operators of ant colony algorithm which increases the global search capability of original algorithm and moves the search speed in the solution space. Comparing with the traditional algorithm through experiment, the results show that the new algorithm improves some extent in solving performance and jumping out of the local minimum solution. The stacked chip temperature is significantly reduced when The new algorithm is applied to thermal layout optimization.