A Tree-based genetic algorithm for distributed database

The successful application of genetic algorithms in distributed database greatly depends on appropriate coding method for query optimization because the coding method of parameters can directly affect the construction of genetic operators and performance of the algorithms. Considering a combinatorial optimum with the restriction of both the position and the condition to include a lot of messages, this paper aims to devise a new coding method with tree structure based on the position and the value. Besides, the genetic operators, i.e. reproduction, crossover and mutation are designed for this coding. The improved crossover needs to be implemented by two steps and the improved mutation is composed of the value mutation and the position mutation. The proposed algorithm is used to implement query of distributed database, the experiment results showed that it is very effective for optimization.