We propose a new method for network reconstruction by the stationary
distribution data of Markov chains on this network. Our method has the merits
that: the data we need are much few than most method and need not defer to the
time order, and we do not need the input data. We define some criterions to
measure the efficacy and the simulation results on several networks, including
computer-generated networks and real networks, indicate our method works well.
The method consist of two procedures, fist, reconstruct degree sequence,
second, reconstruct the network(or edges). And we test the efficacy of each