Thanks Chen Lao Shi for introducing me to this email exchange; I am gland to know Professor Fang.
I have one quick comment: Frank Kelly of Cambridge University (who was just elected into US NAE this year as a foreign member) has been working on a stochastic approach to network for many years. In some sense, he is one of the 先行者. http://www.statslab.cam.ac.uk/~frank/ In 1979, he already published a book http://www.statslab.cam.ac.uk/~frank/rsn.html in which the issue you have emphasized constantly, the issue of detailed balance, has been extensively discussed. The Markov network approach (both as process and as Gibbs measure of interacting systems) is very much in the same spirit as your Darwinian dynamics. It puts a dynamics on the structure.
Best wishes to you all, Hong.
Fwd: comments on Prigogine type approach to nonequilibrium processes-II