With the development of the Internet, network service requirements exceed the capabilities of the network, which causes a great influence to the application of Quality of Service (QoS). As the scale of new generation of the Internet becoming larger, it is difficult to know the global information of the network. How to solve the anycast routing with QoS guaranty on the condition of just being aware of the local information has been a hot researching point in recent years. Most of the current anycast algorithms are not distributed and not suitable for the super large-scale distributed network. A novel approach using distributed Multiple Ant Colonies algorithm with positive interaction is presented in this paper, in which ants adjust individual behavior via cooperation to make colony behavior intelligent and only need local information to find optimal routing. The theoretical analysis and simulation results show that the innovative ant algorithm is a desirable routing algorithm for the anycast routing problem with multiple QoS constraints, which offers significant performance gains across a wide range of operational scenarios.

基于环的简单扩展性，Petersen图的短直径与超立方体互联网络中节点的高可连接性相结合，提出了一种新型互联网络RHP(n)(Ringed Hypercube Connected Petersen)，并对其特性进行了研究。证明了RHP(n)网络不但具有正则性以及良好的可扩展性，同时还具有比Qn、HP(n)网络更短的直径和更小的构造开销。另外，还基于RHP(n)网络分别给出了其上的单播和广播路由算法，证明了其通信效率分别为n-1和n-1。

n 维超立方体结构的多处理机系统在并行与分布式处理中具有良好的性能。随着多处理机系统规模的增大，系统出现链路与节点故障的概率也随之增大，因此设计容错性更强的路由算法对n 维超立方体结构的多处理机系统具有重要意义。针对系统中存在链路故障的情况，提出了用于记录最优通路的安全通路向量(safety path vectors，简称SPVs)概念，并给出了建立SPVs 及其容错路由算法。其中SPVs 的赋值可以通过n-1轮邻节点之间的信息交换来完成，且算法中各节点的存储开销仅为n bits，因此，SPVs 是安全向量(SVs)与扩展安全向量(ESVs)的一种扩展，具有比SVs 和ESVs 更好的记录最优通路的能力。另外，与基于最优通路矩阵(optimal path matrices，简称OPMs)及扩展最优通路矩阵(extended optimal path matrices，简称EOPMs)的容错路由算法相比，SPVs 呈指数级地降低了算法的存储开销，且能够记录OPMs 和EOPMs 所不能记录到的最优通路信息。理论分析和仿真实验验证了SPVs 的上述性能。

无线传感器网络有限的资源，如能量、通信带宽等，难以适应网络中大量数据的传输，需要在网络内部对原始监测数据进行压缩或聚合处理.设计了一个适合小波变换的环模型。针对任意支撑长度的小波函数，给出了一种基于环模型的分布式时空小波数据压缩算法.该算法将传感器网络中的数据抽象为一个矩阵，将时间相关性与空间相关性映射为该矩阵的小波列变换与行变换，以同时挖掘传感器网络中数据的时间和空间相关性。从能量消耗和网络延时两方面定性地分析了算法的性能.理论分析和实验结果表明，该算法能够有效地去除传感数据中存在的时间和空间相关性,降低网络能耗和延时。

由于传感器网络自身的一些特殊需求，加上其节点资源非常有限和网络特征的多样性，其安全问题在很大程度上不同于传统网络。给出了传感器网络的安全体系结构，力图对该领域的研究进行整体的描述。讨论了3个方面的问题：1) 所采用的密码体制，如SKE，MAC，PKC等；2) 密钥管理中的多种密钥建立算法； 3) 路由、能耗和容错等多因素相关的综合考虑。最后，简要介绍了其他一些安全问题以及该领域今后的研究热点。

蚁群算法是一种新型的模拟进化算法，重点始于组合优化问题的求解。 作者运用该算法优化PID控制参数,但在基本蚁群算法中，存在收敛速度较慢，易出现停滞，以及全局搜索能力较低的缺陷。论文提出了一种具有遗传因子的自适应蚁群算法最优PID 控制参数的方法，设计出参数优化图。该方法克服了基本蚁群算法的不足，能够满意地实现PID控制参数优化。仿真结果与Z-N法、遗传算法、基本蚁群算法相比较，优化效果明显得到改善。 实验表明，该方法对于控制其他对象和过程也具有应用价值。

提出了一种基于决策图贝叶斯网络的强度Pareto进化算法，该算法把贝叶斯概率模型结合到多目标进化算法中，通过构造和学习网络来替代传统进化算法中的交叉重组和变异等遗传操作，避免对大量参数的人工设置和重要构造块的破坏。求解多目标背包问题的仿真结果表明，所提算法可以快速收敛到较好的Pareto前沿，有很强的鲁棒性。

This Paper investigates the TCP (Transmission Control Protocol) friendliness of multicast videoconferencing system s. Through the analysis and simulation experiments it is shown that the slow response to network state changes and the fixed rate adjustment Process lead to TCP unfriendliness in the bandwidth sharing. Therefore, this Paper proposes a new TC P friendly flow control algorithm called TFR C-IV S flow control algorithm for the current best-effort Internet. TFRC-IVS (TCP-Friendly Rate Control- INR. IA Videoconferencing System) algorithmutilizes TCP friendly control function derived from complex TC P model to calculate TC P friendly sending rate. Simulation results show that TFRJC-IV S flow control algorithm improves the smoothness of transmission rates and converges quickly to the stable sending rate. In addition, the TCP friendly control function in TFRC-IV S flow control algorithm ensures the TCP friendliness of video flows and fair bandwidth allocation with TC P flows, which the traditional static rate adjustment algorithm lacks.

The extension of reinforcement learning (RL) to large state space has inevitably encountered the problem of the curse of dimensionality. Improving the learning efficiency of the agent is much more important to the practical application of RL. Consider learning to optimally solve Markov decision problems in a particular domain, if the domain has particular characteristics that are attributable to each state, the agent might be able to take advantage of these features to direct the future learning. This paper firstly defines the local state feature, then a state feature function is used to generate the local state features of a state. Also a weight function is introduced to adjust current policy to the actions worth exploring. Based on the above, an improved SARSA algorithm, Feature-SARSA, is proposed. We validate our new algorithm by experiment on a complex domain, named Sokoban. The results show that the new algorithm has better performance.

Bayesian spam filters, in general, compute probability estimations for tokens either without considering the email areas of occurrences except the body or treating the same token occurred in different areas as different tokens. However, in reality the same token occurring in different areas are inter-related and the relation too could play role in the classification. In this paper we incorporated this novel idea, co-relating multi-area information by co-weighting them and obtaining more effective combined integrated probability estimations for tokens. The new approach is compared with individual area-wise estimations and traditional separate estimations in all areas, and the experimental results with three public corpora showed significant improvement, stability, robustness and consistency in the spam filtering with the proposed estimation.