多膜遗传算法及其应用

发布时间：2018-06-18 14:53

本文选题：膜计算 + P系统　；参考：《广西大学》2017年硕士论文

【摘要】：作为自然计算中最年轻的分支,膜计算(也称P系统)主要研究如何从细胞的结构和功能,以及生物的组织、器官的结构与功能中建立起来计算模型。大量的研究表明,膜计算模型具有图灵机的计算能力,研究膜计算模型可以在允许时间范围内解决计算难问题。膜计算应用研究中通常会采用P系统设计的优化方法用来解决实际应用中的一些问题,其中主要与遗传算法、粒子群算法、蚁群算法、模拟退火算法等结合,用来解决包括旅行商问题、背包问题、机器人路径规划问题等。但是到目前为止,大多数的研究没有考虑到膜与膜之间的通信交流。由于细胞与环境之间的浓度不同,细胞之间会被动的运输化学物质,本文将引用这种通信方式,并且将膜计算原理的三大特点应用到标准遗传算法中,从而设计出了一种新的优化算法——多膜遗传算法。本文所介绍的多膜遗传算法,对于三个单独的膜结构,以细胞之间的浓度通信为基础,将变群体规模遗传算(VPGA)法分别应用到三个膜系统中,并且将标准遗传算法(GA)中三种进化规则由串行改为并行。浓度通信机制可以使三个种群数目平均分配,变群体规模遗传算法可以使种群不至于收敛的太快,也不会浪费计算资源。这不但提高了搜索的质量还兼顾了快速寻优的效果。本文采用实数编码将新的算法程序在MATLAB中编译完成,并且通过几个经典函数进行验证,研究结合之后算法的有效性,并且与传统的遗传算法进行比较。最后,验证了多膜遗传算法的有效性,并将其应用到非最小相位系统校正装置的参数寻优中。在MATLAB环境下,通过结合SIMULINK仿真,实现了对该控制系统的参数优化,并且取得了理想的控制效果。
[Abstract]:As the youngest branch of natural computing, membrane computing (also called P system) mainly studies how to establish computational models from the structure and function of cells, as well as the structure and function of biological tissues and organs. A large number of studies have shown that the membrane computing model has the computing power of the Turing machine, and the membrane computing model can solve the difficult problem within the allowable time range. In the application research of membrane computing, the optimization method of P system design is usually used to solve some problems in practical application, which are mainly combined with genetic algorithm, particle swarm algorithm, ant colony algorithm, simulated annealing algorithm and so on. It is used to solve the traveling salesman problem, knapsack problem, robot path planning problem and so on. However, up to now, most studies have not considered the communication between membrane and membrane. Because of the different concentration between the cell and the environment, the cells can transport the chemical substances passively. In this paper, we will use this communication mode, and apply the three characteristics of the membrane computing principle to the standard genetic algorithm. Thus, a new optimization algorithm, multi-membrane genetic algorithm, is designed. In this paper, for three single membrane structures, the variable population size genetic algorithm (VPGA) is applied to the three membrane systems based on the concentration communication between cells. And the three evolutionary rules in the standard genetic algorithm (GA) are changed from serial to parallel. The concentration communication mechanism can make the three populations equally distributed, and the variable population size genetic algorithm can make the population not converge too fast, nor waste computing resources. This not only improves the quality of search, but also gives consideration to the effect of fast optimization. In this paper, the new algorithm program is compiled in MATLAB by real number coding, and verified by several classical functions. The validity of the algorithm is studied and compared with the traditional genetic algorithm. Finally, the effectiveness of the multi-membrane genetic algorithm is verified and applied to the parameter optimization of the non-minimum phase system calibration device. In MATLAB environment, the parameters of the control system are optimized by Simulink simulation, and the ideal control effect is obtained.
【学位授予单位】：广西大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18

【参考文献】