<< Chapter < Page | Chapter >> Page > |
A heat bath algorithm is simply a procedure of updating the vertices one at a time based on the conditional probabilities given by . All vertices must be chosen infinitely often so that there is no bias towards any assignments, . The following equation provided by Propp and Wilson [link] indicates how will be updated:
where is time, and is a random real number in chosen according to the uniform distribution. All 's are chosen independently of each other [link] .
We may order sets of configurations by the following definition: iff , where . Two configurations are incomparable if one is not greater than the other. We also say that a probability distribution, , is monotonic if and only if
or equivalently, ). If the probability distribution is monotonic (or “attractive"), then the partial ordering of and is preserved after the heat bath algorithm is applied. To see why this is true, note that if the inequality in formula [link] holds, then must be greater than or equal to ). The proof is as follows:
If we choose the same for both and , it is impossible for and to occur simultaneously. Therefore after any amount of time, , whatever configuration becomes after updates must remain greater than or equal to whatever configuration becomes after updates. This preservation of partial ordering leads to the theorem proved by Propp and Wilson:
Theorem: If one runs the heat bath for an attractive spin system under the coupling-from-the-past protocol, one will generate states of the system that are exactly governed by the target distribution, [link] .
Expanding the Monotone Monte Carlo Algorithm to YD's is fairly simple, if we treat each YD as a configuration set, , in a spin system. The unit squares are vertices in the vertex set of all in the a b box, where spins up if is in the YD and spins down otherwise. The configurations are selected according to the probability distribution
where denotes how many vertices in spin up. The vertices are updated according to the conditional probabilities given by . There are three equations to update by: 1) if is improbable, , ; 2) if is improbable , ; 3) if both and have positive probability, update according to the equation
where is time, and is a random real number in chosen according to the uniform distribution where all 's are chosen independently of each other.
It is easy to demonstrate using the three equations that for all possible , and , formula [link] holds given . Essentially, we have 5 possible cases:
1.) If is chosen to be a box in such that is not a valid YD, then is not a valid YD either, since . Therefore and . Formula [link] becomes .
2.) If is chosen to be a box such that both and are YD's, but is not a valid YD, then
Notification Switch
Would you like to follow the 'The art of the pfug' conversation and receive update notifications?