Fast optimization by Demon Algorithms
Ian Wood, Tom Downs · QUT ePrints (Queensland University of Technology) · 1998
We introduce four new general optimization algorithms based on the `demon' algorithm from statistical physics and the simulated annealing (SA) optimization method. These algorithms use a computationally simpler acceptance function, but can use any SA annealing schedule or move generation function. Computation per trial is significantly reduced. The algorithms are tested on traveling salesman problems including Grotschel's 442-city problem and the results are comparable to those produced using SA. Applications to the Boltzmann machine are considered.