Soy: An Efficient MILP Solver for Piecewise-Affine Systems. Wu, H., Wu, M., Sadigh, D., & Barrett, C. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS '23), pages 6281–6288, October, 2023. IEEE. Detroit, MI, USA
Soy: An Efficient MILP Solver for Piecewise-Affine Systems [link]Paper  doi  abstract   bibtex   1 download  
Piecewise-affine (PWA) systems are widely used for modeling and control of robotics problems including modeling contact dynamics. A common approach is to encode the control problem of the PWA system as a Mixed-Integer Convex Program (MICP), which can be solved by general-purpose off-the-shelf MICP solvers. To mitigate the scalability challenge of solving these MICP problems, existing work focuses on devising efficient and strong formulations of the problems, while less effort has been spent on exploiting their specific structure to develop specialized solvers. The latter is the theme of our work. We focus on efficiently handling one-hot constraints, which are particularly relevant when encoding PWA dynamics. We have implemented our techniques in a tool, Soy, which organically integrates logical reasoning, arithmetic reasoning, and stochastic local search. For a set of PWA control benchmarks, Soy solves more problems, faster, than two state-of-the-art MICP solvers.
@inproceedings{WWS+23,
  url       = "https://doi.org/10.1109/IROS55552.2023.10342011",
  author    = "Wu, Haoze and Wu, Min and Sadigh, Dorsa and Barrett, Clark",
  title     = "Soy: An Efficient {MILP} Solver for Piecewise-Affine Systems",
  booktitle = "2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS '23)",
  publisher = "IEEE",
  pages     = "6281--6288",
  month     = oct,
  year      = 2023,
  doi       = "10.1109/IROS55552.2023.10342011",
  note      = "Detroit, MI, USA",
  category  = "Conference Publications",
  abstract  = "Piecewise-affine (PWA) systems are widely used for modeling and
                  control of robotics problems including modeling contact
                  dynamics. A common approach is to encode the control problem
                  of the PWA system as a Mixed-Integer Convex Program (MICP),
                  which can be solved by general-purpose off-the-shelf MICP
                  solvers. To mitigate the scalability challenge of solving
                  these MICP problems, existing work focuses on devising
                  efficient and strong formulations of the problems, while less
                  effort has been spent on exploiting their specific structure
                  to develop specialized solvers. The latter is the theme of
                  our work. We focus on efficiently handling one-hot
                  constraints, which are particularly relevant when encoding
                  PWA dynamics. We have implemented our techniques in a tool,
                  Soy, which organically integrates logical reasoning,
                  arithmetic reasoning, and stochastic local search. For a set
                  of PWA control benchmarks, Soy solves more problems, faster,
                  than two state-of-the-art MICP solvers.",
}

Downloads: 1