扩散过程的元胞自动机模拟

以元胞自动机为研究方法,通过编写Matlab程序,以演化示意图的形式,模拟了圆域内的扩散过程,给出了累积演化率曲线.同时,用微积分的方法,解析了该扩散过程,并结合实例与元胞自动机方法对比.研究证明,元胞自动机可以生动形象地模拟扩散过程,且与微积分得到的结论一致.     

Cellular automaton-based algorithms for the dispersion of mobile wireless sensor networks

Due to the advent of sensor technology and its applications, mobile wireless sensor networks (MWSNs) have gained a significant amount of research interest. In a typical MWSN, sensors can move within the network. We develop a set of probabilistic and deterministic cellular automaton (CA)-based algorithms for motion planning problems in MWSNs. First, we consider a scenario where a group of sensors are deployed and they need to disperse in order to maximise the area covered by the network. In this variant of the problem we do not explicitly consider that the sensors should maintain the connectivity of the network while they move. Second, we consider a scenario where the sensors are initially randomly distributed and they need to disperse autonomously to both maximise the coverage of the network and maintain its connectivity. We carry out extensive simulations of both deterministic and randomised variants of the algorithms. For the first variant of the problem we compare our algorithms with one previous algorithm and find that our algorithm yields better network coverage than the earlier algorithm. We also find that probabilistic algorithms have better overall performance for the second variant. CA algorithms rely only on local information about the network and, hence, they can be used in practice for MWSN problems. On the other hand, locality of the algorithm implies that maintaining connectivity becomes a non-trivial problem.     

细胞自动机：简单的规则，复杂的行为

细胞自动机是一种时间、空间与状态都离散的数学模型。回顾了初等细胞自动机模型的应用及其符号动力学刻画的相关文献,在大量统计性质和计算机模拟基础上,着重分析具有鲁棒Bernoulli移位特征的细胞自动机规则的一些符号动力学性质,揭示了细胞自动机的简单规则中蕴含着复杂、混沌的非线性动力学特性。这些结果丰富了细胞自动机的理论基础,也将促进符号动力系统的理论和应用的研究。     

元胞驻留视频群体状态预报方法

视频场景人群状态预报是防止公共场所安全事故的一种重要措施．为此提出了一种视频场景群体状态预报模型．该模型使用格内驻留元胞自动机(inner-grid parking cellular automata, IPCA)模型预报视频中群体状态．首先,通过光流法进行运动跟踪作为IPCA的输入特征对场景进行建模,并根据特征点的运动状态自适应地调整其在元胞格内的驻留时间以提高预报精度;其次,通过分析IPCA模型的预测结果判断群体微团间的相互作用得到视频帧的状态信息,从而达到对场景中人群状态进行预报的目的;最后,利用当前场景的状态信息对该场景的预报输出进行反馈校正,使模型能够及时地反映场景中群体状态的变化,该模型不依赖于具体的视频帧,以微团结构的碰撞作为异常检测依据．实验结果表明,与传统的检测方法相比,该预报模型能够提前预报视频场景中的异常状态,并具有较好的准确性．     

基于遗传元胞机的多态性作业车间布局建模

针对多态性作业车间布局问题,根据车间物流关系和二维空间约束条件,构建车间元胞机框架模型,并且引入遗传算法优化模型当中的自组织演化规则,最终建立多态性作业车间布局遗传元胞机模型.采用系统布置设计方法求取布局方案的初始解,在保证初始解有效性的基础上提高算法的寻优速度和精度,以物料搬运费用最小、车间面积利用率和设备平均利用率最大为模型多目标函数.通过实例验证了遗传元胞机模型求解多态性作业车间布局的可行性与实用性.     

基于元胞自动机的交通流分析

结合元胞自动机的理论,对车辆运行建立了交通流数学模型.通过C++软件描述了运行车辆除超车之外靠右行规则下,交通流量随换道概率变化的情况;通过Matlab软件进行仿真模拟,揭示了交通密度在右行规则控制下所形成的交通流的规律性.结果表明:在靠右行驶规则下,适当的转向概率和车速有利于交通流量的提升.     

基于曲率驱动机制的晶粒生长元胞自动机模型

结合晶粒生长的统计分忻理论和概率性转变觇则，建立了基于曲率驱动机制的晶粒正常生长的二维元胞自动模型，该模型的一个显著特点是晶粒生长速率和晶界曲率(1／RC-1／R)之间满足线性关系，而传统的元胞自动机模型则难以满足这种线性关系模拟结果比较全面地反映了等温条件下晶粒正常生长的各种现象：平均晶粒面积与时间的关系，晶粒尺寸、晶粒边数的分布规律及其时间不变性，晶粒边数与尺寸的关系，以及晶粒生长速率与晶粒尺寸和曲率的关系等．模拟结果与晶粒生长的动力学理论预测相符合。     

Construction of living cellular automata using the Physarum plasmodium

The plasmodium of Physarum polycephalum is a unicellular and multinuclear giant amoeba that has an amorphous cell body. To clearly observe how the plasmodium makes decisions in its motile and exploratory behaviours, we developed a new experimental system to pseudo-discretize the motility of the organism. In our experimental space that has agar surfaces arranged in a two-dimensional lattice, the continuous and omnidirectional movement of the plasmodium was limited to the stepwise one, and the direction of the locomotion was also limited to four neighbours. In such an experimental system, a cellular automata-like system was constructed using the living cell. We further analysed the exploratory behaviours of the plasmodium by duplicating the experimental results in the simulation models of cellular automata. As a result, it was revealed that the behaviours of the plasmodium are not reproduced by only local state transition rules; and for the reproduction, a kind of historical rule setting is needed.     

基于GPU的元胞自动机熔岩流动模拟

为解决基于元胞自动机进行熔岩流动模拟的计算效率问题,提出一种应用在元胞自动机上的GPU并行计算方法。将元胞自动机中每一个方形网格映射到GPU的一个逻辑计算单元上,通过并行计算,提高模拟的效率,解决传统串行计算方法的不足,使模拟达到实时性。模拟结果表明,在元胞自动机的物理模型理论基础上,用GPU并行计算进行加速,在模拟效果和时间效率上均取得了良好的效果。     

非确定型离散事件系统双模拟控制的实现

近年来,双模拟等价关系与离散事件系统监控理论相结合的研究引起了国内外许多学者的广泛关注.本文针对作者在前期工作中提出的非确定型离散事件系统的双模拟控制机制,进一步探讨其实现问题.利用投影映射对系统规范说明语言进行等价类划分,构造了一棵计算树,得到了一个判断规范说明是否具有基于模拟关系可观性的多项式算法,证明了双模拟控制机制是多项式时间算法可实现的.同时,通过对控制器配备具有存储和判断功能的模拟关系识别器,阐述了这种双模拟控制机制是物理可实现的.