A Simplex-Based Scalable Linear Constraint Solver for User Interface Applications

We propose a scalable algorithm called HiRise2 for incrementally solving soft linear constraints over real domains. It is based on a framework for soft constraints, known as constraint hierarchies, to allow effective modeling of user interface applications by using hierarchical preferences for constraints. HiRise2 introduces LU decompositions to improve the scalability of an incremental simplex method. Using this algorithm, we implemented a constraint solver. We also show the results of experiments on the performance of the solver.