Search On The Cultural Algorithm With Dual Knowledge And Its Application

Evolutionary algorithms have been applied to all kinds of fields successfully, because it doesn’t depend on professional knowledge when solving optimization problem. Cultural algorithm is proposed to simulate human being behaviors, this algorithm includes population space and belief space, individual evolution in population space is guided by the knowledge in belief space. The algorithms that meet requirements all can embed cultural algorithms’ double structures, and make up the inadequacy for itself. While cultural algorithm also has some disadvantages: it can only simulate simply, have not adaptability; population space have not regionalism, individuals are influenced by the some knowledge.In this dissertation, to solve the problem that different individuals get the same experience when they face to the same knowledge, propose the dual knowledge. The dual knowledge is determined by situational knowledge, normative knowledge and the current individual. Dual knowledge means particular evolutionary experience after individual is affected by the knowledge. So dual knowledge conducts individual evolution, the direction of different individual is decided by the individual dual knowledge, and it can improve convergence rates and apply to high dimensions problems successfully. The main contents in this paper are as following:(1) Research on the cultural algorithm in cellular space structure, and introduce the belief space, situation knowledge, normative knowledge and their influence strategies. Cultural algorithm, cellular genetic algorithm and cultural algorithm under cellular space structure are all tested by some functions; the results show the cultural algorithm under cellular space structure is better than others, especially the convergence performance.(2) By improving the cultural algorithm under cellular space structure, the dual knowledge is proposed. The dual knowledge is determined by the current individual, situational knowledge, and normative knowledge. Offspring will close to the direction of dual knowledge and situational knowledge slowly, finally retain the optimal individual by competing. And from this put forward a new algorithm,the cultural algorithm with dual knowledge. Simultaneously, the experiments test that the proportions of knowledge affect in process evolution of the cultural algorithm with dual knowledge, it can prove the effectiveness of dual knowledge, and the dual knowledge improves the ability of jumping out of the local optimal solution.(3) This part is that improved cultural algorithm applies to unmanned aerial vehicle path planning. In the framework of cultural algorithm, extract knowledge, put some planning methods together, it can make up some disadvantages. The simulation shows that cultural algorithm based on path planning can track target and avoid threats rapidly in dynamic environments.