A learning automata based scheduling solution to the dynamic point coverage problem in wireless sensor networks

The dynamic point coverage problem in wireless sensor networks is to detect some moving target points in the area of the network using as few sensor nodes as possible. One way to deal with this problem is to schedule sensor nodes in such a way that a node is activated only at the times a target point is in its sensing region. In this paper we propose SALA, a scheduling algorithm based on learning automata, to deal with the problem of dynamic point coverage. In SALA each node in the network is equipped with a set of learning automata. The learning automata residing in each node try to learn the maximum sleep duration for the node in such a way that the detection rate of target points by the node does not degrade dramatically. This is done using the information obtained about the movement patterns of target points while passing throughout the sensing region of the nodes. We consider two types of target points; events and moving objects. Events are assumed to occur periodically or based on a Poisson distribution and moving objects are assumed to have a static movement path which is repeated periodically with a randomly selected velocity. In order to show the performance of SALA, some experiments have been conducted. The experimental results show that SALA outperforms the existing methods such as LEACH, GAF, PEAS and PW in terms of energy consumption.     

城市路网多路口路径动态实时选择方法

为了缓解城市交通拥堵问题,如何充分利用现有的道路资源进行有效的路线导航,一直是学者们关心的热点问题.现有的研究方法包括：优化交通灯信号周期以增大交通流量;对个别车辆的行驶路线进行优化;利用历史交通数据或者通过路网中心和车辆之间的主从式博弈进行路径导航等.然而,这些研究并没有考虑到微观行驶车辆的个性化交通需求以及多车辆彼此之间的路线选择冲突,对于城市路网中交通状况的动态不确定性也没有充分考虑.基于以上问题,提出了城市交通路网动态实时多路口路径选择模型DR²SM（dynamic and real-time route selection model in urban traffic networks）,结合车辆对前方可选路线的偏好以及可选路线的实时交通状况,并利用自适应学习算法SALA（self-adaptive learning algorithm）进行博弈,以使得各行驶车辆的动态路线选择策略达到Nash均衡.     

基于博弈论的内河港口作业车辆协同选路方法

针对以汽车运输为主且吞吐量较大的内河港口的交通拥堵问题,提出一种基于博弈论的内河港口作业车辆协同选路方法。首先,基于港口路网特征与车辆作业特点,将同时请求路径规划的作业车辆间的交互建模为不完全信息博弈,采用满足均衡（SE）的概念来分析该博弈。假设每个车辆对选路效用都有一个预期,当所有车辆都得到满足时博弈即达到均衡。然后,提出了一种车辆协同选路算法,算法中每个车辆首先按照贪心策略初始选路,之后将所有车辆按规则分组,组内车辆根据历史选路结果进行适应性学习并完成博弈。实验结果表明,当港区同时作业车辆数为286时,协同选路算法的车辆平均行驶时间分别比Dijkstra算法和自适应学习算法（SALA）少50.8%和16.3%,系统收益分别比Dijkstra算法和SALA提高51.7%和24.5%。所提算法能够有效减少车辆平均行驶时间,提高系统收益,更适用于内河港口车辆选路问题。