Ideally, one would like a method for determining the annealing schedule based on the behavior of the specific system. We can obtain some guiding principles by considering the statistical mechanics of the problem.
One good rule of thumb is that we would like to minimize the variation from equilibrium. During the annealing, we can measure the energy (cost) and specific heat (variance of the energy) at each temperature value. These values can be used to help determine the annealing schedule.
Note that since we are constantly
changing T, it is not easy to obtain equilibrium values of E and 
for
each temperature.