Multiple bit encoding-based search algorithms

For a given real-world problem, we do not know a priori which representation suits this problem. Schnier and Yao showed the benefit of using the multiple real-coded evolutionary algorithm. This paper discusses the multiple bit encoding-based (standard binary and reflected Gray code) search algorithms. The population-based genetic algorithm and single individual-based random bit climber algorithm are used to show the efficacy of the multiple encoding. Each algorithm has three versions which are based on the standard binary encoding, standard reflected Gray code and multiple encoding scheme respectively. The experiments on two search algorithms show the robust and better performance than the single encoding scheme and also show empirically how multiple representations can benefit search as much as a good search operator could.