A Study Of RRNA Inspired DNA Genetic Algorithm

The DNA genetic algorithm is based on the framework of genetic algorithm, which introduces the coding method and the operating mechanism of DNA computing. It improves the performance of genetic algorithm. The research and application of DNA genetic algorithm have a huge potential. Aiming at the DNA genetic algorithm, the mutation operators and mutation probability are studied, and then we take application object for background to test the proposed algorithms. The main contents of this thesis are as follows:(1) Based on the structure and framework of DNA genetic algorithm and its genes'characteristics, structure, mutation and the transfer information of ribosomal RNA, a RDGA Algorithm (ribosomal RNA inspired DNA Genetic Algorithm) is presented. This algorithm uses the crossover operator and the anticodon inversion operator of the DNA genetic algorithm, meanwhile adopts three new mutation operators---the Pre-cut operator, Self-mutation operator, Environmental guide operator. Tested by 4 typical test functions and compared with standard genetic algorithm (SGA) and RNA genetic algorithm (RNA -GA),the results show that RDGA algorithm can obtain better performance, especially in convergence speed and precision.(2) Taking the RDGA algorithm to solve the design problem of high order controllers with multiple performances of the ship diesel. Compared with that of FST-MC, GA, the results show that the RDGA is of good performance.(3) Considering the effect of mutation probability, the fuzzy RDGA (FRDGA) is proposed. This algorithm use a fuzzy controller to turn the mutation probability. in every generation. So the performance of RDGA is improved. Based on the experiments of 4 typical test functions, the comparison results show that the FRDGA has certain improvement in the convergence and precision.(4) The FRDGA algorithm is applied for the gasoline blending scheduling problem.The solutions of 24 hour and 72 hour gasoline blending problem are got. Comparison with PSO and GA show that the FRDGA algorithm is more effective.