Integer linear programming and branch & bound algorithm based approach for constructing MRST with high efficiency

Minimum rectilinear Steiner tree (MRST) construction is a fundamental problem in VLSI routing, combinational optimization and computational geometry. In this paper, a novel approach based on the combination of geometric analysis, integer linear programming and branch & bound algorithm is presented to construct MRST efficiently. Firstly, an integer linear programming model of MRST problem is built well versed in binding with special features of solution space. Then in order to depress the high computational complexity of constructing MRST, geometric analysis and the branch and bound algorithm are introduced to optimize search space. Experimental results demonstrated the capabilities of the proposed approach to build MRST with high accuracy and efficiency.