Local Search For Maximizing Satisfiability In Qualitative Spatial And Temporal Constraint Networks

We focus on the recently introduced problem of maximizing the number of satisfied constraints in a qualitative constraint network (QCN), called the MAX-QCN problem. We present a particular local search method for solving the MAX-QCN problem of a given QCN, which involves first obtaining a partial scenario S of that QCN and then exploring neighboring scenarios that are obtained by disconnecting a variable of S and repositioning it appropriately. The experimentation that we have conducted shows the interest of our approach for maximizing satisfiability in qualitative spatial and temporal constraint networks.