The optimal model reduction method for spatially distributed system based on simulated annealing algorithm

For partial differential equation description unknown spatially distributed systems, the number of local models determines the dimension of the model. So far, there is no mature method about how to obtain the optimal region division. Usually, the local region division is related with the location of sensors. It may affect the accuracy and computational complexityH of the modeling directly. This paper presents an optimal model reduction approach for spatially distributed systems based on simulated annealing algorithm. At first, the optimality criterion is presented. And then, the simulated annealing based iterative optimizing method is proposed to solve the optimal model reduction. The simulations demonstrated show the accuracy and efficiency of the proposed methodologies.