【作者】 任晓娜；

【导师】 张大方； 王泽平；

【作者基本信息】 湖南大学 ， 软件工程， 2010， 硕士

【摘要】 DNA计算是以DNA分子和生物酶等为载体,以生化反应作为“信息处理工具的”一种崭新的计算模式。由于DNA计算所具有的巨并行性、海量存储以及低能耗等优点,DNA计算显示了巨大的生命力。但生化反应中由于各种外界因素以及编码本身的缺陷,会存在非特异性杂交,这种随机性为后续计算进程带来了很大的不确定性,严重的可能会导致生化反应不可控,进而造成计算失败,有效的编码设计能够提高DNA计算的可靠性。本文重点研究了DNA编码问题,研究重点主要集中在如何设计出可以减少不合法反应的DNA编码序列集合。介绍了DNA编码的相关约束条件,并进行数学建模,研究现有流行的智能进化算法。在此基础上提出两种DNA编码的优化算法。本文的主要工作包括：1.分析DNA计算中编码序列设计的影响因素、总结了现有研究主要考虑的六个数学约束条件：H-measure约束、连续性约束、相似度约束、发夹结构约束、解链温度、GC含量。建立一套基于以上六个约束条件的多目标评价体系的DNA编码序列组合优化模型,设计出了适应度函数Fitness的具体实现。2.提出一种基于离散粒子群的DNA编码优化方法(简称DPSO)。根据DNA编码问题离散量的特点,对DPSO算法粒子的位置、速度等量及运算规则进行了重新定义,满足了六种约束条件,引入基于权重的适应度函数来评价DNA序列集合,并设计出了DPSO算法的具体实现。经实验验证,DPSO算法大大降低了计算复杂度,并且参数设置和调整非常方便。3.提出一种基于文化粒子群(cultural-based PSO, CBPSO)的DNA编码序列优化方法。种群空间利用了DPSO的快速演化能力,通过接收精英个体送入信念空间。同时利用文化算法模型中的遗传操作,对个体选择,交叉,变异等,影响种群空间的粒子来增加DPSO群体多样性和全局搜索能力。并设计出CBPSO的具体实现,经实验验证得到了较高质量的DNA编码序列,验证算法的有效性。4.用Java实现了DNA计算编码设计系统。系统架构采取了五层框架,分别是表现层,算法层,约束层,对象层,数据库层,层和层之间松散耦合。其框架是灵活的,可扩展的。经实验证明,GPSDS可以帮助验证不同的DNA编码优化算法模型。更多还原

【Abstract】 DNA computing is a brand news computing mode which is based on DNA molecular and biological enzyme and uses biochemical reaction as its information processing tool. By nature of the advantages of great parallelity, huge storage ability and low power costs, DNA computing reveals its glorious prosperity. As a branch of DNA computing, DNA coding means designing the initial database of DNA computing. Designing of effective DNA code will remarkably improve the reliability of DNA computing.The paper discusses DNA coding and the main focus is on the decreasing of illegal code sequences. To achieve the DNA coding, the paper briefly introduces the thermodynamics and biology restrictions. And then, mathematic model is designed for the restrictions respectively. Moreover, some evolution algorithms are researched. Based on the research mentioned above, two optimal approaches for DNA coding are designed. The structure of the paper is as follows:1. The main restrictions which have influences on the DNA codes designing are studied. Six mathematic restrictions are discussed respectively. There are H-measure, Continuity, Similarity, Hairpin, Tm as well as GC content. Based on six restrictions mentioned above, an optimal model for DNA coding with a multiple aims evaluation system is established. In the paper, the evaluation system is realized as Fitness function.2. An optimal algorithm for DNA coding named discrete particle swarm optimization (DPSO) is researched. First, the location and velocity updating rule is derived according to the DNA coding restrictions and the characteristic of discrete value. And then, in order to evaluate the performance of DNA consequence set, a fitness function based on the six restrictions as well as their respective power is designed. The DPSO algorithm is realized in chapter four. And in contrast with traditional genetic algorithm, the DPSO algorithm remarkably decreases the computing complexity, also, it’s very convenient to set and adjust the parameters.3. A more optimal algorithm for DNA coding named cultural-based PSO (CBPSO) is designed. The population space adopts the DPSO and makes use of its rapid evolvement attribute, and sends elite individual to the belief space. Meanwhile, the belief space adopts the genetic operations which is including the selection, crossover and variation operation. The belief space will influences the population space to guide the population space and increase the diversity of the population space as well as the global search ability. Through the realization of the CBPSO algorithm, DNA code sequence with higher quality is derived.4. A General-Purpose Sequence Design System is designed and realized using JAVA language. The system structure contains five layers:behavior layer, algorithm layer, restriction layer, object layer and database layer, and those layers are loose coupled with each other. The good traits contribute to an agile, extendable and general-purposed system. Through the realization of the GPSDS system, different optimal model for DNA coding can be implemented, and the respective DNA code sequence acquired.更多还原