Graph-based parallel distributed genetic programming model
LU Qiang1,2,ZHONG Wei2,WANG Zhiguang2
1. Beijing Key Laboratory of Earth Prospecting and Information Technology, China University of Petroleum, Beijing 102249, China
2. College of Geophysics and Information Engineering, China University of Petroleum, Beijing 102249, China
Abstract:Since Genetic Programming (GP) is of natural parallelism, the parallel and distributed GP model was developed, including master-slave model, island model and grid model. However, the realizing algorithm of these distributed models is complex and they cannot be reused. It is difficult to achieve the scale computation of GP quickly based on different topologies. Due to these shortcomings, the authors presented the graph-based parallel distributed GP model which realizes formal description of the various operations of GP, and could support the distributed parallel computation of GP for different topologies. It is easy to achieve the master-slave model, island model and grid model of GP through experimental test. The new model is stable, efficient and easy to realize.