基于可信工人路径规划的声望检验算法

针对工人自选择任务模式下无法确定路径规划算法工人可信度的问题,提出了基于可信工人路径规划的声望检验算法。首先,根据欧氏距离与区域限制对任务进行k均值聚类,缩小可信工人候选任务集,从而降低路径规划算法的时间复杂度;其次,提出了利用权值k对可信工人路径优化的算法,该算法综合考虑可检验工人数量、距离成本和截止时间计算任务权值,基于任务权值最大化寻找最佳插入位置来建立和更新可信工人的执行路径;最后,通过采用抽查检验的方式对可信工人与普通工人提交数据对比的结果来建立工人声望模型,该模型使用了可信性、不可信性和不确定性三个参数的期望值来描述工人的可靠性。在模拟数据集和gMission真实数据集上对声望检验算法进行实验,结果表明提出的声望校验算法数据采集质量提升13%,路径规划算法在保证抽查质量的基础上,可信工人的旅行成本降低11%。

考虑工人培养的移动群智感知任务分配机制

移动群智感知(MCS)通过大量感知工人的移动性和工人随身携带的感知设备来收集数据,是一种新的大规模数据感知范式。现有大量研究致力于解决移动群智感知中的任务分配问题,使感知数据质量得以提高,但忽略了缺乏优质工人的感知任务,导致任务完成质量降低。为了解决上述问题,对于缺乏优质工人的感知任务,该文关注将经验不足的工人培养为优质工人,并令其执行这些感知任务,实现工人的长期复用,提高感知数据质量和长期平台效用。具体来说,该文考虑了缺乏优质工人的感知任务所需的能力和工人的能力类型,并据此应用稳定匹配算法选择待培养工人,提出一种基于能力聚合和半马尔可夫预测的多阶段工人选择培养(MWSD)算法。结果表明,相比基于区块链的非确定团队协作(BNTC)算法,该文所提算法能够有效将缺乏优质工人的感知任务的数据质量提高24%,长期平台效用提高17%。

移动众包中基于多约束工人择优的激励机制研究

随着移动众包的快速发展,市面上的众包平台如雨后春笋般出现,它们发布任务并利用人群的力量来执行任务、收集数据。此时,移动众包中有效的激励机制变得十分重要。然而现有的激励机制只片面地考虑工人的信誉度、所在位置和执行时间等,这使得众包平台在有限的预算或其他约束的情况下选定优质工人并分配多个任务变得困难。针对以上问题,文中提出了一种基于多约束工人择优的激励机制(Multi-constrained Worker Selection Incentive Mechanism, MSIM),该模型依赖于两个相关算法:一是基于改进逆向拍卖的工人择优算法,该算法综合考虑工人信誉度、地理位置、任务完成度、结果质量等多个重要约束来选择最优的工人执行任务;二是评估和奖惩算法,该算法对任务执行结果和工人信誉度进行评估,从而制定对工人的奖励与惩罚规则。实验结果表明,MSIM可以选出优秀的工人,并提高任务执行结果的可信度和工人信誉度,是一种良好的激励机制。

信息年龄敏感的稀疏移动群智感知机制

稀疏移动群智感知是一种新兴的模式,它从感知区域的子集收集数据,然后推理其他区域的数据.然而,在实际应用中,工人不足或分布不均的情况广泛存在.因此,在有限的预算下,必须优先选择相对更重要的工人收集数据.此外,许多稀疏移动群智感知应用对数据的时效性要求较高.因此本文将考虑数据的新鲜度,并使用信息年龄作为新鲜度指标.为了解决这些挑战,本文提出了一种轻量级年龄敏感的数据感知和推理框架.该框架旨在预算约束下,选择合适的工人收集数据,并通过准确捕捉感知数据时空关系进行数据推理,以优化信息年龄和推理的准确性.由于预算和工人有限,可能会导致数据量较少的情况.因此,本文还提出了精简数据推理模型的方法,以提高推理效率.通过广泛的实验进一步论证了该框架在实际应用中的优越性.

A self region based real-valued negative selection algorithm

Point-wise negative selection algorithms,which generate their detector sets based on point of self data,have lower training efficiency and detection rate.To solve this problem,a self region based real-valued negative selection algorithm is presented.In this new approach,the continuous self region is defined by the collection of self data,the partial training takes place at the training stage according to both the radius of self region and the cosine distance between gravity of the self region and detector candidate,and variable detectors in the self region are deployed.The algorithm is tested using the triangle shape of self region in the 2-D complement space and KDD CUP 1999 data set.Results show that,more information can be provided when the training self points are used together as a whole,and compared with the point-wise negative selection algorithm,the new approach can improve the training efficiency of system and the detection rate significantly.

Feature selection for chemical process fault diagnosis by artificial immune systems

With the Industry 4.0 era coming, modern chemical plants will be gradually transformed into smart factories, which sets higher requirements for fault detection and diagnosis(FDD) to enhance operation safety intelligence. In a typical chemical process, there are hundreds of process variables. Feature selection is a key to the efficiency and effectiveness of FDD. Even though artificial immune system has advantages in adaptation and independency on a large number of fault samples, antibody library construction used to be based on experience. It is not only time consuming, but also lack of scientific foundation in fault feature selection, which may deteriorate the FDD performance of the AIS. In this paper, a fault antibody feature selection optimization(FAFSO) algorithm is proposed based on genetic algorithm to optimize the fault antibody features and the antibody libraries' thresholds simultaneously. The performance of the proposed FAFSO algorithms is illustrated through the Tennessee Eastman benchmark problem.

Effect of an escape vent in accordion-shaped traps on the catch and size of Asian paddle crabs Charybdis japonica in an artificial reef area

Accordion-shaped traps are widely used in China to catch the Asian paddle crab C harybdis japonica but traps of conventional design often catch juvenile crabs. A new type of accordion-shaped trap with an escape vent(L×W=4.3 cm×3.0 cm) was designed and a comparative study between the newly designed and conventional traps was performed in the artifi cial reef area of Zhuwang, Laizhou Bay, China from June to August 2012. The mean catch per unit effort(CPUE) of undersized crabs was signifi cantly lower in the vented traps than in the conventional traps(paired t-test, n =30, P <0.001), while the CPUE of marketable crabs was signifi cantly higher in the vented traps(paired t-test, n =30, P <0.001). The mean size of crabs(carapace length) caught in the vented traps was signifi cantly larger than in conventional traps(paired t-test, n =29, P <0.001). The ratio of undersized crabs was 35.05%±2.57% in conventional traps and 12.53%±0.69% in vented traps(signifi cantly lower, paired t-test, n =29, P <0.001). Therefore, a 4.3 cm×3.0 cm escape vent was considered appropriate for C. japonica fi shing in the artifi cial reef area. This fi nding will assist the development of more sustainable and effi cient crab fi shing methods using accordion-shaped traps.

Application of FAHP and Artificial Neural Network on Clothing Plant Location

Clothing manufacturers＇ direct investment and joint ventures in developing regions have seen to grow rapidly in the past few decades. Non-optimized selection can contribute to adverse effects affecting the performance of the plants on aspects of productivity, manufacturing and logistics cost. Selection of proper plant location is thus crucial. The conventional approaches to sites location are based on the factors and their weights. However, determining the weight of each factor is very difficult and time consuming. While the situation is changed, all the work must be redone again. This study aims to develop a decision-making system on clothing plant location for Hoog Kong clothing manufacturer. The proposed system utilizes artificial neural network to study the relationship between the factors and the suitability index of candidate sites. Firstly, the factors are stratified using the fuzzy analytical hierarchy process （FAHP） by review the related references and interviewing the experts. Secondly, the corresponding data are collected from the experts by questionnaire and the related government publication. Finally, the feedforward neural network with error backpropagation（EBP） learning algorithm is trained and applied to make decision. The results show that the proposed system performs well and has the characteristic of adaptability and plasticity.

A matting method based on color distance and differential distance

A new matting algorithm based on color distance and differential distance is proposed to deal with the problem that many matting methods perform poorly with complex natural images.The proposed method combines local sampling with global sampling to select foreground and background pairs for unknown pixels and then a new cost function is constructed based on color distance and differential distance to further optimize the selected sample pairs.Finally,a quadratic objective function is used based on matte Laplacian coming from KNN matting which is added with texture feature.Through experiments on various test images,it is confirmed that the results obtained by the proposed method are more accurate than those obtained by traditional methods.The four-error-metrics comparison on benchmark dataset among several algorithms also proves the effectiveness of the proposed method.

Research on a randomized real-valued negative selection algorithm

A real-valued negative selection algorithm with good mathematical foundation is presented to solve some of the drawbacks of previous approach. Specifically, it can produce a good estimate of the optimal number of detectors needed to cover the non-self space, and the maximization of the non-self coverage is done through an optimization algorithm with proven convergence properties. Experiments are performed to validate the assumptions made while designing the algorithm and to evaluate its performance.

Online Adaptation of Game AI with Evolutionary Learning

Since the beginning of computer games era, artificial intelligence (AI) has been a standard feature of games. The current emphasis in computer game AI is improving the quality of opponent AI. Our research question reads: How can unsupervised online learning be incorporated in Computer Role Playing Game(CRPG) to improve the strategy of the opponent AI? Our goal is to use online evolutionary learning to design strategies that can defeat the opponent. So we apply a novel technique called dynamic scripting that realizes online adaptation of scripted opponent AI and report on experiments performed in a simulated CRPG to assess the adaptive performance obtained with the technique.

Method for Selecting the Optimal Solution for the Design of the Gear Train

Some researchers in mechanical engineering have developed systems for the design of gear transmission boxes, but almost no calculation methods exist for widespread synthesis. In this article, we outline methods for automatic determination of toothed helical gear trains and the selection criteria for the optimal choice of gear trains. In this work two methods were applied. As first design to use an expert system for the design and then optimize the design that is why we used Kappa PC and Catia for CAD.

基于信誉模型的众包质量控制算法

针对目前众包平台会产生大量恶意工人以及较少考虑激励工人多次提供可信服务的问题,提出一种基于信誉模型的众包质量控制算法——信誉期望最大化(reputation expectation maximum,Rep-EM)算法。根据可信因子和惩罚因子建立信誉模型;基于工人信誉值和对任务的熟悉度提出一种工人选择机制;将工人匹配度作为权重赋予相应的工人并使用多数投票方法进行初始值选取,解决期望最大化(expectation maximum,EM)算法对初始值敏感和收敛困难的问题,避免算法陷入局部最优,提高评估结果的准确率;利用公开的众包数据集Adult2和Duck对Rep-EM算法和本研究提出的机制进行验证。试验结果表明,Rep-EM算法在评估准确率和运行时间方面有很大的提升,也从任务完成率和平均数据质量验证了本研究提出的工人选择机制的有效性。