A self-adaptive very fast simulated annealing based on Hidden Markov model

The simulated annealing (SA) is amongst the well-known algorithms for stochastic optimization. Unfortunately, its major weakness is the slow rate of convergence, leading to a large period of poor improvement towards a global optimum. In this paper, we present a self-adaptive approach to enhance the SA performance during the run using Hidden Markov Model (HMM). Experiments have been performed on many benchmark functions.