An Integrated CP/OR Method for Optimal Control of Modular Hybrid Systems

Paper in proceeding, 2014

This paper concerns the optimal control of modular hybrid systems synchronized by shared variables. Instead of synchronizing the discrete dynamics of the system into one global mode before optimization, Constraint Programming (CP) is used to model the discrete dynamics of each modular system separately. Integrated in the CP solver are also classic Operations Research (OR) models in the form of Nonlinear Programs (NLPs) approximating the continuous dynamics of the system. Using CP considerably decreases the number of NLPs which must be solved, compared to that of using a traditional mixed integer nonlinear programming approach.