线程池技术应用研究

目前几乎所有操作系统都支持多线程的工作方式,以提高系统的处理能力.本文研究了采用线程池技术可以提高服务器程序性能,并通过一个实例分析了线程池技术中线程池尺寸和任务数对服务器程序的性能影响.     

Learning With A Friendly Interactive Robot For Service Tasks in hospital environments

Our work addresses the problem of mobile robots control and programming. We present FIRST, a Friendly Interactive Robot for Service Tasks, designed to carry heavy loads in hospitals. A learning method is used to teach it the set of all the trajectories it has to follow. During the learning phase, a human operator teleoperates the robot while its sensors are activated. So the trajectory is built according to these data. The teleoperation is easy and accurate thanks to several locomotion modes of the robot, specially a crab mode. A Learning Computer helps the operator in any phase of the learning mode. As the missions feasibility must be assumed, each trajectory will be checked before used. Furthermore, any learned trajectory can be linked to a trajectory generated with the help of another method, provided that all files are of the same kind. Eventually, while performing the mission, the robot compares the measurements provided by the localization system and the learned points in order to ensure an accurate trajectory tracking.     

A genetic based disk scheduling method to decrease makespan and missed tasks

Disk scheduling is an operating system process to service disk requests. It has an important role in QOS guarantee of soft real-time environments such as video-on-demand and multimedia servers. Since now, some disk scheduling algorithms have been proposed to schedule disk requests in an optimized manner. Most of these methods try to minimize makespan by decreasing the number of disk head seeks as one of the slowest operations in modern computers and crucial for system performance because it usually takes some milli-seconds. In this paper, we propose a new disk scheduling method based on genetic algorithm that considers makespan and number of missed tasks simultaneously. In the proposed method, a new coding scheme is presented which employs simple GA procedures such as crossover and mutation and a penalty function in fitness. To get the best performance of the proposed method, its parameters such as number of chromosomes in initial population, mutation, and crossover probabilities, etc have been adjusted by applying it on some sample problems. The algorithm has been tested on several problems and its results were compared with well-known related methods. Experimental results showed that the proposed method worked very well and excelled most related works in terms of miss ratio and average seeks.     

Reliability-aware scheduling strategy for heterogeneous distributed computing systems

Heterogeneous computing systems are promising computing platforms, since single parallel architecture based systems may not be sufficient to exploit the available parallelism with the running applications. In some cases, heterogeneous distributed computing (HDC) systems can achieve higher performance with lower cost than single-machine supersystems. However, in HDC systems, processors and networks are not failure free and any kind of failure may be critical to the running applications. One way of dealing with such failures is to employ a reliable scheduling algorithm. Unfortunately, most existing scheduling algorithms for precedence constrained tasks in HDC systems do not adequately consider reliability requirements of inter-dependent tasks. In this paper, we design a reliability-driven scheduling architecture that can effectively measure system reliability, based on an optimal reliability communication path search algorithm, and then we introduce reliability priority rank (RRank) to estimate the task’s priority by considering reliability overheads. Furthermore, based on directed acyclic graph (DAG) we propose a reliability-aware scheduling algorithm for precedence constrained tasks, which can achieve high quality of reliability for applications. The comparison studies, based on both randomly generated graphs and the graphs of some real applications, show that our scheduling algorithm outperforms the existing scheduling algorithms in terms of makespan, scheduling length ratio, and reliability. At the same time, the improvement gained by our algorithm increases as the data communication among tasks increases.     

面向小样本库的全局Gabor滤波人脸识别

虽然约束条件下的人脸识别已取得很大进展,但开放环境下的人脸识别仍有很多问题需要解决;其次,人脸识别实用系统往往难以获得待识别人的多个样本。针对上述两点,研究小样本库情况下的人脸识别问题,提出了一种新的人脸特征提取方法:对人脸图像进行多尺度多方向Gabor滤波,用特定尺度/方向滤波后的均值和方差作为新的人脸特征进行识别。基于JAFFE和ORL数据库进行"小样本训练,大样本测试"验证,实验结果及后续分析比对证明了方法的有效性。     

The effectiveness of synchronous computer-mediated communication for solving hidden-profile problems: Further empirical evidence

We experimentally compared the effectiveness of face-to-face (FTF) and synchronous computer-mediated communication when using a chat tool in solving hidden-profile business problems. In such problems, information critical to its solution is dispersed among team members and they must share it to solve the problem. Unlike prior research using hidden-profile tasks, our study used a real-world business-oriented task, established real rather than ad hoc teams, and imposed a time constraint on them. Hypotheses derived from media richness theory were found to be supported, with the results revealing that computer-mediated teams using the chat tool were less successful in exchanging and processing information than FTF teams and were thus less successful at solving the hidden-profile problem. The results also showed that, when operating under a time constraint, FTF was preferred over computer-mediated communication due to the relative immediacy of feedback and multiplicity of cues available in the FTF setting, as media richness theory predicted.     

Dynamic scheduling of maintenance tasks in the petroleum industry: A reinforcement approach

Petroleum industry production systems are highly automatized. Maintenance of such systems is vital, not only to maintain production efficiency but also to insure minimal safety levels. Maintenance task scheduling is difficult since some tasks are already identified because they must be done repeatedly, and other tasks need to be identified dynamically. In this paper, we present a multi-agent approach for the dynamic maintenance task scheduling for a petroleum industry production system. Agents simultaneously insure effective maintenance scheduling and the continuous improvement of the solution quality by means of reinforcement learning, using the SARSA algorithm. Reinforcement learning allows the agents to adapt, learning the best behaviors for their various roles without reducing the performance or reactivity. To demonstrate the innovation of our approach, we include a computer simulation of our model and the results of experimentation applying our model to an Algerian petroleum refinery.     

基于Fast R-CNN的车辆目标检测

目的 在传统车辆目标检测问题中,需要针对不同图像场景选择适合的特征。为此提出一种基于快速区域卷积神经网络（Fast R-CNN）的场景图像车辆目标发现方法,避免传统车辆目标检测问题中需要设计手工特征的问题。方法 该方法基于深度学习卷积神经网络思想。首先使用待检测车辆图像定义视觉任务。利用选择性搜索算法获得样本图像的候选区域,将候选区域坐标与视觉任务示例图像一起输入网络学习。示例图像经过深度卷积神经网络中的卷积层,池化层计算,最终得到深度卷积特征。在输入时没有规定示例图像的规格,此时得到的卷积特征规格不定。然后,基于Fast R-CNN网络结构,通过感兴趣区域池化层规格化特征,最后将特征输入不同的全连接分支,并行回归计算特征分类,以及检测框坐标值。经过多次迭代训练,最后得到与指定视觉任务强相关的目标检测模型,具有训练好的权重参数。在新的场景图像中,可以通过该目标检测模型检测给定类型的车辆目标。结果 首先确定视觉任务包含公交车,小汽车两类,背景场景是城市道路。利用与视觉任务强相关的测试样本集对目标检测模型进行测试,实验表明,当测试样本场景与视觉任务相关度越高,且样本中车辆目标的形变越小,得到的车辆目标检测模型对车辆目标检测具有良好的检测效果。结论 本文提出的车辆目标检测方法,利用卷积神经网络提取卷积特征代替传统手工特征提取过程,通过Fast R-CNN对由示例图像组成定义的视觉任务训练得到了效果良好的车辆目标检测模型。该模型可以对与视觉任务强相关新场景图像进行效果良好的车辆目标检测。本文结合深度学习卷积神经网络思想,利用卷积特征替代传统手工特征,避免了传统检测问题中特征选择问题。深层卷积特征具有更好的表达能力。基于Fast R-CNN网络,最终通过多次迭代训练得到车辆检测模型。该检测模型对本文规定的视觉任务有良好的检测效果。本文为解决车辆目标检测问题提供了更加泛化和简洁的解决思路。     

A QoS-sensitive task assignment algorithm for mobile crowdsensing

Recently, mobile crowdsensing has attracted many researchers’ attention due to the pervasiveness of smart phones. In this paper, we study the QoS-sensitive Task Assignment (QSTA) problem for mobile crowdsensing, which involves variable tasks and flexible rewards. Each user can conduct one or more tasks, and the number of tasks assigned to each user is flexible. The reward paid to the user is related to the assigned tasks. To ensure the sensing QoS of tasks, each task might be assigned to multiple users. We prove the QSTA problem to be NP-hard, and propose a greedy algorithm to solve this problem. Moreover, we also analyze the approximation ratio of the proposed algorithm. Additionally, we demonstrate the significant performance of the algorithm through simulations     

云制造环境下大规模产品定制技术研究

结合云制造及大规模产品定制的特点,提出了一种云制造环境下大规模产品定 制的组织模式以及相配套的需求转换模型。需求转换模型涵盖了完整的需求分解、重组和再 分配过程,对其中的基于产品结构树的需求分解、基于相似性聚类的制造批次重组都进行了 详细阐述,并构建了需求再分配的非线性规划模型,从而可以利用最优化方法实现制造需求 寻优分配。最后,以轮式装载机产品定制为示例,论述了需求转换的详细过程,验证了方法 的有效性和可行性。