1: \begin{abstract}

2: We consider a reconstruction problem of a reduced stable positive network system with the preservation of the original interconnection structure based on an $H^2$ optimal model reduction problem with constraints.

3: To this end, we define an important set using the Perron--Frobenius theory of nonnegative matrices such that

4: all elements of the set are stable and Metzler.

5: Using the projection onto the set,

6: we propose a cyclic projected gradient method to produce a better reduced model than an initial reduced model in the sense of the $H^2$ norm.

7: In the method,

8: we use Lipschitz constants of the gradients of our objective function to define

9: the step sizes without a line search method whose computational complexity is large.

10: Moreover, the existence of the Lipschitz constants guarantees the global convergence  of our proposed algorithm to a stationary point.

11:  The numerical experiments demonstrate that the proposed algorithm improves a given reduced model, and can be used for large-scale systems.

12: \end{abstract}

13: