Identification of piecewise affine systems via mixed-integer programming. (Q1426258)

This paper deals with identification of hybrid dynamical systems, by focusing attention on hinging hyperplanes and Wiener piecewise affine autoregressive exogenous models. The regression space is partitioned into polyhedra with affine submodels for each polyhedron. The authors use algorithms based on mixed-integer linear or quadratic programming converging to a global optimum. A change detection approach is proposed for the special case where the estimation data only seldom switches between the different submodels. In this way it is possible to weigh between optimality and complexity. Numerical examples are treated.