基于群智感知的移动互联网服务质量评估与分析研究

发布时间：2018-06-17 14:39

本文选题：移动互联网 + 服务质量　；参考：《北京邮电大学》2017年硕士论文

【摘要】：移动互联网是传统移动通信网络与互联网融合的产物,能够为用户提供更具灵活性,随时随地的网络服务。随着移动互联网市场规模的不断增长,激烈的市场竞争也使运营商意识到提高移动互联网服务质量,才是生存和盈利的关键。因此,移动互联网的服务质量渐渐受到了越来越广泛的关注。如何准确的评估移动互联网服务质量成了亟需解决的问题。同时,随着无线移动终端设备的爆炸式普及以及众包(crowdsourcing)模式的日趋成熟,两者结合产生了群智感知(crowdsensing)。群智感知的优点在于能够利用群体的力量,克服地理因素,时间因素的限制,进行大规模,高质量的数据收集任务。因而,逐渐被广泛应用在日常生活的各个方面。本文围绕基于群智感知的移动互联网服务质量评估展开研究。首先,提出了一种基于用户加权迭代的移动互联网服务质量评估方法。该方法考虑到用户受个人因素影响导致的差异性,以及不同评估场景的需要,分别赋予用户及各类型指标不同权重。本文从仿真模拟和实测数据两个角度验证其准确性。通过仿真模拟,得到用户加权迭代算法的平均误差距离比多数投票算法小44.9%,用户权重平均误差距离比多数投票算法小16.7%。基于LTE网络中用户侧实测数据,区分不同基站场景,分别对比算法准确性。其中,在交通主干道和居民区场景中,用户加权迭代算法的误差率比多数投票算法小30%以上。其次,提出一种基于价格策略的群智感知任务平台激励机制。从群智感知任务平台角度出发,将感知任务发布者与感知任务平台之间的交互过程利用斯坦科尔伯格博弈模型进行建模。进而通过迭代学习算法求解感知任务平台在纳什均衡下的最优价格策略,并利用仿真模拟评估该算法性能。综上,本文通过理论分析及实验为基于群智感知的移动互联网服务质量评估提供了理论和技术支撑。
[Abstract]:Mobile Internet is the combination of traditional mobile communication network and Internet, which can provide users with more flexible, anytime, anywhere network services. With the continuous growth of the mobile Internet market, the fierce market competition also makes operators realize that improving the quality of mobile Internet service is the key to survival and profitability. Therefore, the quality of service of the mobile Internet has been paid more and more attention. How to accurately evaluate the quality of mobile Internet service has become an urgent problem. At the same time, with the explosive popularity of wireless mobile terminal devices and the increasing maturity of crowd-sourcing mode, the combination of the two has produced crowdsensing. The advantage of group intelligence perception lies in its ability to use the power of the group, overcome the limitation of geographical factors and time factors, and carry out large-scale and high quality data collection tasks. As a result, it has gradually been widely used in all aspects of daily life. This paper focuses on the quality of service evaluation of mobile Internet based on group intelligence perception. Firstly, a user-weighted iterative method for mobile Internet QoS evaluation is proposed. Taking into account the difference caused by individual factors and the needs of different evaluation scenarios, the method gives users and different types of indicators different weights. This paper verifies its accuracy from the point of view of simulation and measured data. Through simulation, the average error distance of user weighted iterative algorithm is 44.9 less than that of majority voting algorithm, and the average error distance of user weight is 16.7. Based on the measured data of user side in LTE network, different base station scenarios are distinguished and the accuracy of the algorithm is compared. The error rate of the user-weighted iterative algorithm is more than 30% less than that of the majority voting algorithm in the traffic main roads and residential area scenarios. Secondly, a kind of incentive mechanism based on price strategy is proposed. From the perspective of group intelligence perception task platform, the interaction process between perceptual task publisher and perceptual task platform is modeled by using Steinkelberg game model. Then the iterative learning algorithm is used to solve the optimal price strategy of the perceptual task platform under Nash equilibrium, and the performance of the algorithm is evaluated by simulation. In summary, this paper provides theoretical and technical support for mobile Internet QoS evaluation based on group intelligence perception through theoretical analysis and experiments.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5;TP393.0

【相似文献】