绝对值距离Steiner最小树问题是在集成电路布线等领域应用广泛的属于NP难的经典组合优化问题,由于该问题的搜索空间与元胞自动机的结构相似,设计了求解绝对值距离Steiner最小树问题的改选的元胞蚂蚁算法。经大量数据实验表明,该算法要比最小生成树平均改选15%,优于多数已有的基于最小生成树的近似算法,验征了算法的实用性。
The rectilinear Steiner minimum tree problem is a classical NP-complete problem of combinatorial optimization which has been widely used in the layout of integrated circuit etc.This paper designed an improved cellular automata ant algorithm based on the discovery of the similar structure between the cellular automata and the search area of the rectilinear Steiner minimum tree.The computational results showed that this algorithm can improve the total length about 15% than that of the minimum spanning tree,so the effectiveness of the algorithm has been validated.