RNA遗传算法及在桥式吊车中的应用研究

发布时间：2019-04-03 14:11

【摘要】：遗传算法是一类简单易用、鲁棒性强、具备并行能力且扩展空间大的智能优化算法。它模拟了生物进化"优胜劣汰、适者生存"的自然法则,而且由于遗传算法对待优化问题没有连续性或可微的要求,被广泛应用于科研与工程领域。RNA遗传算法(RNA-GA)是受RNA分子操作启发的一类新型遗传算法,本文在前人的研究基础上,针对RNA遗传算法的改进以及在桥式吊车中的应用问题进行了较为深入的探讨。本文的主要研究内容如下:(1)简要综述了遗传算法的发展历程与研究现状,以及桥式吊车的建模与控制方面的研究状况。(2)提出了一种受鱼群行为启发的RNA遗传算法(fsRNA-GA)。在fsRNA-GA中采用了所设计的一种邻域搜索操作,该操作通过定义适应度函数的拥挤度因子来衡量个体与个体之间的拥挤程度,以便在局部空间内发现更有潜力的个体。针对待优化参数较多的情况,使用矩阵编码的方式代替RNA-GA中的链式编码。通过对一些典型测试函数的寻优实验,表明了该算法对低、高维无约束优化问题寻优的良好性能。将所提出的fsRNA-GA用于优化径向基函数(RBF)神经网络基函数中心,利用桥式吊车实验平台采集的数据训练神经网络,建立桥式吊车位置和摆角的RBF神经网络模型,仿真实验表明了所建立神经网络模型的有效性。(3)提出了一种高位密码子选择操作RNA遗传算法(csRNA-GA)。该算法通过选择操作增加了精英基因在进化过程中进入下一代的概率,丰富了种群的多样性。通过对一些典型测试函数的寻优实验,结果表明所提算法具有更好的全局搜索能力和平均收敛精度。针对桥式吊车系统PID控制器参数整定问题,将csRNA-GA算法用于优化桥式吊车系统位置和摆角双PID控制器参数,仿真实验表明优化得到的PID控制器参数能够实现二维桥式吊车快速、无超调、摆动幅度小的控制效果。
[Abstract]:Genetic algorithm is a kind of intelligent optimization algorithm, which is simple and easy to use, strong robustness, parallel ability and large expansion space. It simulates the natural law of biological evolution, "survival of the fittest," and because genetic algorithms have no continuity or differentiability in their treatment of optimization problems, RNA genetic algorithm (RNA-GA) is a new type of genetic algorithm inspired by RNA molecular operations. The improvement of RNA genetic algorithm and its application in bridge crane are discussed in this paper. The main contents of this paper are as follows: (1) the development and research status of genetic algorithms are summarized briefly. And the research status of modeling and control of bridge crane. (2) A new RNA genetic algorithm (fsRNA-GA), which is inspired by fish herd behavior, is proposed. A neighborhood search operation designed in fsRNA-GA is used to measure the degree of crowding among individuals by defining the crowding factor of fitness function in order to find more potential individuals in local space. In view of the large number of parameters to be optimized, matrix coding is used instead of chain coding in RNA-GA. Experiments on some typical test functions show that the algorithm has good performance for low and high dimensional unconstrained optimization problems. The proposed fsRNA-GA is used to optimize the radial basis function (RBF) neural network basis function center. The neural network is trained by the data collected from the bridge crane experiment platform, and the RBF neural network model of the bridge crane position and swing angle is established. The simulation results show the effectiveness of the proposed neural network model. (3) A high-order codon selection operation RNA genetic algorithm (csRNA-GA) is proposed. This algorithm increases the probability of elite genes entering the next generation in the process of evolution, and enriches the diversity of the population. Through the optimization experiments of some typical test functions, the results show that the proposed algorithm has better global search ability and average convergence accuracy. Aiming at the parameter setting problem of PID controller in bridge crane system, the csRNA-GA algorithm is used to optimize the parameters of double PID controller of bridge crane system. The simulation results show that the optimized parameters of the PID controller can realize the control effect of the two-dimensional bridge crane with fast speed, no overshoot and small swing amplitude.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18

【参考文献】