An optimal modeling method based on differential evolution algorithm is suggested for complex nonlinear system
that can be denoted as piecewise affine model(PWA). The method combines partitioning the operating regimes with building
the local models on the basis of input-output data, the PWA identification is transformed to a mixed integer quadratic
program(MIQP). Then the number of model and the parameters of the models are optimized by using the chaotic adaptive
differential evolution algorithm simultaneously, and the boundary of hyper-plane is estimated by support vector machine.
Simulation results show that the proposed method employs the optimal sub-model to approximate the nonlinear system
precisely.