A probabilistic multicommodity-flow solution to circuit clustering problems

Circuit clustering plays a fundamental role in hierarchical designs. identifying strongly connected components in the circuits can significantly reduce the complexity of the design and improve the performance of the design process. However, there hm not been a clear objective function for circuit clustering. In this paper, we present a new clustering metric based on the random graph model and the ratio cut(Wei891 concept, A probabilistic, multi-commodity flow based algorithm is proposed and tested under the new clustering metric. Experimental results show that this algorithm generates promising results with respect to the proposed metric. Extensions and directions for future work are also proposed.