Dictionary based estimation of distribution algorithms

This paper proposes a new algorithm in the field of estimation of distribution algorithms. The proposed algorithm combines a data compression algorithm to extract a model into the dictionary. This dictionary is used as a part of the generator to generate the better next generation of population. The proposed method is tested and compared with two well-known genetic algorithms, QHFC and BOA Both algorithms can solve difficult problems and are claimed to scale well. Our algorithm compared well with them. We also report the behaviour of using dictionary as the model for EDAs.