A Note on Building-block Identification by Simultaneity Matrix.pdf

A Note on Building-block Identification by Simultaneity Matrix Chatchawit Aporntewan and Prabhas Chongstitvatana Chulalongkorn University, Bangkok 10330, Thailand Chatchawit.A@student.chula.ac.th Prabhas.C@chula.ac.th Abstract. This paper presents a line of research in genetic algorithms (GAs), called building-block identification. The building blocks (BBs) are common structures inferred from a set of solutions. In simple GA, crossover operator plays an important role in mixing BBs. However, the crossover probably disrupts the BBs because the cut point is chosen at random. Therefore the BBs need to be identified explicitly so that the solutions are efficiently mixed. We propose an approach for identifying BBs. The approach is based on an observation of the compact GA. We observe that the vector elements that correspond to the bits in the same BB are simultaneously updated with a high probability. We formulate the observation to the simultaneity matrix – an `×` matrix of numbers constructed from a set of `-bit solutions. The matrix element in row i and column j is denoted by mij . The larger the mij is, the higher the dependency is between bit i and bit j. If mij is high, bit i and bit j should be passed together to prevent BB disruption. 1 Introduction This paper presents a line of research in genetic algorithms (GAs), called building- block identification. The GAs is a probabilistic search and optimization algo- rithm [7, 4]. The GAs begin with a random population – a set of solutions. A solution (or an individual) is represented by a fixed-length binary string. A solution is assigned a fitness value that indicates the quality of solution. The high-quality solutions are more likely to be selected to perform solution recom- bination. The crossover operator takes two solutions. Each solution splits in two pieces. Then, the four pieces of solutions are exchanged to reproduce two so- lutions. The population size is made constant by discarding some low-quality solutions