A Community-Division Based Algorithm for Finding Relations Among Linear Constraints

Minghao Liu, Feifei Ma, Jun Yan

Abstract: Linear constraints are widely used in the modeling of many practical problems, and the solving technologies have important applications in satisfiability modulo theories, program analysis and verification. The efficiency of solving procedure could be improved by taking advantages of the relations among constraints. Traditional methods find relations through search, which do not take advantage of the structural characteristics and cost too much time. In this paper, a heuristic based on community division is proposed for finding relations among linear constraints. Firstly it builds a relation graph, which maps each constraint to a node. Then a division tool is employed to divide the nodes into several communities. At last, it tries to find relations among constraints in the same community through search. Experimental results show that the algorithm can effectively process large set of constraints, reduce time cost and find relations with higher quality.