Parameter Estimation for nonlinear sloppy model

Mathematical models are very powerful tool to understand real system. For technical reasons, some parameters in models can’t be measured directly, and they have to be identified with the help of optimization algorithm. Due to the nonlinear and sloppy property of models, some algorithms fail to find a good solution even if it takes a long time. In this paper, Bregman regularization homotopy algorithm with geodesic acceleration is proposed to solve the problem. Here, Bregman regularization term is used to suppress ill-posed in inverse problem. Then homotopy plays the role of adjusting or controlling optimization convergence. And geodesic path points out the accelerate direction of the optimization process. At last polymorphnuclear leukocyte metabolic network is used to check the effectiveness of the proposed method.