Computing along the axon

A special form of spiking neural P systems, called axon P systems, corresponding to the activity of Ranvier nodes of neuron axon, is considered and a class of SN-like P systems where the computation is done along the axon is introduced and their language generative power is investigated.     

基于膜计算模型的数独游戏基本解法

为了更好地求解数独问题,提出了一种新的求解方法,利用一个具有抑制催化和膜溶解规则以及进化规则的优先级的膜系统来进行求解;结果表明,对于一个数独问题,只要其所有部分解都至少包含一个具有唯一解的单元格,方法都是有效的;如果数独问题可以利用此策略求解,则膜系统在计算的最后一步将问题的解编码并返回物质YES,否则,膜系统可以检测出数独问题不符合上述特征,返回物质NO,计算停止;方法求解策略与人类求解数独问题的思考过程非常类似,并且给出的是数独问题的统一解,即与数独问题的维度和提示数无关。     

P systems with array objects and array rewriting rules

Array P systems were introduced by Paun Gh. which is linking the two areas of membrane computing and picture grammars. Puzzle grammars were introduced by us for generating connected picture arrays in the two-dimensional plane, motivated by the problem of tiling the plane. On the other hand, incorporating into arrays the developmental type of generation used in the well-known biologically motivated L systems, Siromoney and Siromoney proposed a very general rectangular array generating model, called extended controlled tabled L array system (ECTLAS). In this paper we introduce two variations of the array P system, called BPG array P system and parallel array P system. The former has in the regions array objects and basic puzzle grammar rules (BPG), which are a specific kind of puzzle grammar rules. In the latter, the regions have rectangular array objects and tables of context-free rules. We examine these two types of P systems for their array generative power.     

P systems with array objects and array rewriting rules

Array P systems were introduced by Paun Gh. which is linking the two areas of membrane computing and picture grammars. Puzzle grammars were introduced by us for generating connected picture arrays in the two-dimensional plane, motivated by the problem of tiling the plane. On the other hand, incorporating into arrays the developmental type of generation used in the well-known biologically motivated L systems, Siromoney and Siromoney proposed a very general rectangular array generating model, called extended controlled tabled L array system (ECTLAS). In this paper we introduce two variations of the array P system, called BPG array P system and parallel array P system. The former has in the regions array objects and basic puzzle grammar rules (BPG), which are a specific kind of puzzle grammar rules. In the latter, the regions have rectangular array objects and tables of context-free rules. We examine these two types of P systems for their array generative power.     

移动计算环境中应用程序感知适应性的设计

移动设备和无线网络的发展使得移动数据存取越发重要，而移动计算应用程序必须感知环境的变动来更好地利用有限的资源提供相关的扩展功能。提出能够用来开发环境敏感适应性应用程序的技术和原型，并对低层的运行系统进行了部分实现。在管理环境变化和将环境变化报告给应用程序的新体系结构中，环境变化被抽象成类型事件，事件、传递政策、适应性策略和适应性反应被组织成可扩展类型层次结构，从而将环境参数管理的复杂性限定在底层运行系统中，简化了应用程序的设计。该体系结构不但能够用于建造新的移动性感知应用程序，也可以很方便地将移动感知适应性结合到现存的应用程序中去，实验结果证明了系统的有效性。     

基于CUDA并行平台实现核膜系统求解可满足性问题的研究

膜系统能以空间换时间的方式在多项式时间内求解NP难问题,作为膜系统中的一个新类型,核膜系统不仅包含了不同类型膜系统的特征,而且能提供一种统一性的描述．因此基于核膜系统构建系统模型,并以此求解一个典型的NP难问题：可满足性问题．为了加速对核膜系统的仿真过程,程序将在CUDA并行方式下仿真实现,通过与传统的串行方式比较,使用CUDA能减少95％的仿真时间,这也表明使用CUDA并行方式非常适合仿真核膜系统．     

Solving the factorization problem with P systems

P systems have been used many times to face with computationally difficult problems, such as NP-complete decision problems and NP-hard optimization problems. In this paper we focus our attention on another computationally intractable problem: factorization. In particular, we first propose a simple method to encode binary numbers using multisets. Then, we describe three families of P systems: the first two allow to add and to multiply two binary encoded numbers, respectively, and the third solves the factorization problem.     

Solving the factorization problem with P systems

P systems have been used many times to face with computationally difficult problems, such as NP-complete decision problems and NP-hard optimization problems. In this paper we focus our attention on another computationally intractable problem: factorization. In particular, we first propose a simple method to encode binary numbers using multisets. Then, we describe three families of P systems: the first two allow to add and to multiply two binary encoded numbers, respectively, and the third solves the factorization problem.     

软计算及其新一代模糊控制器

介绍了软计算的概念及构成软计算的三个组成部分：模糊逻辑、神经网络和概率推理,并讨论了以软计算为基础的新一代模糊控制器．     

粒度计算在人工神经网络中的应用

介绍了粒度计算的基本思想和理论基础,详细论证了基于词计算理论的一种粒向量空间,研究了基于这种粒向量空间的人工神经网络模型.该模型可以在一定程度上提高人工神经网络的时效性、知识表达的可理解性,并能增强人工神经网络的功能.     

普适计算对我国IT行业的挑战

普适计算是正在发展的一种新的计算模式，硬件技术、软件技术、通讯技术、网络技术等是普适计算发展和普及的关键技术．我国在普适计算时代面临着巨大机遇和挑战，所以应积极研发具有自主知识产权的普适计算技术，加大开发普适计算市场的力度．     

对分布式计算、网格运算和云计算分析

云计算是分布式计算、网格计算的发展,或者说是这些计算机科学概念的商业实现。本文针对分布式计算、网格计算和云计算从定义、工作原理、各自特性之间的关联与不同进行了分析研究。     

基于MPI的投影法数控编程算法

提出了基于消息传递的刀位轨迹并行计算方法.该方法首先将串行计算任务合理的划分成若干并行子任务,然后动态的分配到各处理机上,能充分利用网络中空闲的计算资源,提高计算速度.分别应用于投影法和自适应投影法编程,试验结果表明该方法具有令人满意的并行加速比.     

不可测过程参数的智能化软计算方法及应用

为解决控制系统中存在的时变过程数学建模困难,有些过程参数难以在线实时检测的问题,研究基于过程神经元网络的不可测过程参数软计算方法。分析正交基函数展开的学习算法收敛速度慢的问题,将BP网络的改进算法引入到过程神经元网络的训练中,增加基函数展开系数的规一化处理、权函数动量项调整项,以及学习率自适应调整方法,改进算法明显消除了误差收敛曲线振荡和网络收敛速度慢的问题。该方法在污水处理过程出水水质BOD预测中网络训练速度快,取得较好的预测结果,是一种不可测过程参数智能化软计算的有效方法。     

面向云端融合计算的防御性数据销毁机制

云端融合计算将云计算和对等计算有机地融合,充分挖掘网络中心集群服务器端和网络边缘终端所蕴含的各种可利用的资源。引入了动态性比较强的网络边缘节点之后,云端融合计算环境下的数据安全保护和远程数据销毁更加困难,特别是恶意的异地主体可能对被托管的数据实施非法篡改等攻击行为。为此,本文提出了基于移动Agent的防御性数据销毁机制,利用移动Agent对云端数据进行检测,将待上传密文数据的HASH值拼接在密文后面,并定时对数据进行双重检测,有效区分云端数据的正常修改和非法篡改。通过预测相关数据受攻击的可能性,对已被非法篡改或有被非法篡改可能性的数据进行有效处理。设计并构建了基于移动Agent的云端数据防御性数据销毁原型系统。实验结果表明,基于移动Agent的防御性数据销毁机制在检测准确性、安全性、时间开销等方面具有良好的性能表现。      

基于紧致编码的高阶神经网络光计算

提出一种能压缩三阶神经网络的互连张量规模近一半的紧致编码方法。设计两种光电混和系统-非相干多重成象系统和非相干相关系统,并用它实现了三阶互连的二维神经网络,构成图象平方关联存贮器。在光学实现中,借助于一种多重矩阵排列方式,作者用一块二维掩模板表示六维的互连张量,用二维的发光二极管(LED)阵列表示四维的输入张量,实验结果与数值模拟的结果吻合。     

基于异种机互联网络的并行神经计算环境的实现和性能分析

给出了在多台微机与SUN工作站的互连网上实现的一个并行计算环境 :HCPC(Heteroge neousComputersParallelComputing)系统 ,并通过对其性能的分析和在ART1神经网络上的模拟实现 ,验证了HCPC系统的功能     

不饱和度的计算方法及通用公式的推导

介绍了几种常见不饱和度的计算方法并扩展了相应烃法,推导了一个适合包括多元有机复合物在内的所有有机物的不饱和度通用计算公式．     

云计算研究综述

介绍了云计算产生的背景和云计算的定义,辨析了云计算与其他计算形式的区别,阐述了云计算的特点及其体系结构,展望了云计算的发展前景。     

A membrane evolutionary algorithm for DNA sequence design in DNA computing

DNA sequence design has a crucial role in successful DNA computation,which has been proved to be an NP-hard(non-deterministic polynomial-time hard) problem.In this paper,a membrane evolutionary algorithm is proposed for the DNA sequence design problem.The results of computer experiments are reported,in which the new algorithm is validated and out-performs certain known evolutionary algorithms for the DNA sequence design problem.