This method comes from statistical physics. Basically you want to generate a Boltzmann-distribution of states in the function of "energy". By defining the Marko

This method comes from statistical physics. Basically you want to generate a Boltzmann-distribution of states in the function of "energy". By defining the Markov chain in a certain way you can achieve this. This is the Metropolis-Hasting algorithm[1]. I believe it's somewhat hard to use this algorithm for optimum finding. You have to tune the temperature and it's works best if you have a good initial temperature and you need to cool the system down slowly (so you have to choose a reasonable timescale for thermalization). Some physical insight could help with both. However this algorithm is used to actually simulate physical systems on finite temperature where you do not need dynamics. You only simulate the Boltzmann-distribution and you measure interesting mean values and correlations. Mostly used for describing phase transitions. [1] http://en.wikipedia.org/wiki/Metropolis%E2%80%93Hastings_algorithm http://en.wikipedia.org/wiki/Metropolis%E2%80%93Hastings_alg...