A non-variational method of this type uses the Green’s function of an imaginary-time Schrödinger equation to evolve a set of “walkers” towards the ground state wavefunction. Hamiltonian dynamics are used to evolve analogues (15[1])of position and momentum (15[1])in parameter space in a method of this type implemented by Hoffman and Gelman’s NUTS algorithm. (15[1])Along with (15[1])simulated annealing, Donald and Stuart Geman developed a method of this type that is used when conditional (*) probabilities are easier to compute than joint probabilities, eliminates data from the burn-in period, and is named after Gibbs. (10[1])In a type (10[1])of these methods, (10[1])the likelihood ratio gives the probability that a step is accepted and added to a Markov chain. For 10 points, the Metropolis-Hastings (10[1])algorithm (10[1])exemplifies what class of methods that estimate an average by repeated sampling? ■END■ (10[1])