In this paper, a new demand-side management problem of networked smart grid is formulated and solved based on evolutionary
game theory. The objective is to minimize the overall cost of the smart grid, where individual communities can switch
between grid power and local power according to strategies of their neighbors. The distinctive feature of the proposed formulation
is that, a small portion of the communities are cooperative, while others pursue their own benefits. This formulation can
be categorized as control networked evolutionary game, which can be solved systematically by using semi-tensor product. A
new binary optimal control algorithmis applied to optimize transient performances of the networked evolutionary game.