Dynamic Simulated Annealing with Adaptive Neighborhood Using Hidden Markov Model

Abstract

The Simulated Annealing (SA) is a stochastic local search algorithm. Its efficiency involves the adaptation of the neighbourhood structure. In this paper, we integrate Hidden Markov Model (HMM) in SA to dynamically adapt the neighbourhood structure of the simulated annealing at each iteration. HMM has proven its ability to predict the optimal behavior of the neighbourhood function based on the search history. An experiments were performed on many benchmark functions and compared with others SA variants.