Given a graph G(V,E) that represents the routing resources such as channels and regions and a netlist NLthat consists of multiple nets, where each net is a subset of vertices in V, the goal of multi-net routing is to construct routing trees of all nets in NLsuch that the capacity constraint specified in each edge in Eis satisfied. The objective is to minimize the total wirelength, routing congestion, longest source-sink path length, etc. This chapter presents sample problems related to the following works:
Steiner min-max tree algorithm [Chiang et al., 1990]
Multi-commodity flow routing algorithm [Shragowitz and Keel, 1987]
Iterative deletion algorithm [Cong and Preas, 1988]
Yoshimura and Kuh algorithm [Yoshimura and Kuh, 1982]
The first router is a sequential router, where the nets are routed one-by-one. The others route all nets simultaneously. The first three routers are global routers, whereas the last one is a detailed router.
KeywordsInteger Linear Programming Minimum Span Tree Edge Weight Incoming Edge Short Path Tree
Unable to display preview. Download preview PDF.