中国声学学会声学媒体与信息分会2015年学术会议征稿通知.doc

服务网络QoE测试节点鲁棒选址问题研究 宁颖丹1 杨文国1,2 高随祥1,2 (1 中国科学院大学 数学科学学院 北京 100049; 2 中国科学院大数据挖掘与知识管理重点实验室) 摘要：QoE测量网选点问题是指在服务网络中部署少量点模拟用户来测试不同网站的服务效果，选择尽可能少的点代表用户，同时又能准确反映网络中用户获取不同服务的质量情况。选址问题通常是NP难问题，影响覆盖的不确定因素大大增加了QoE测量网选点的难度。建立了QoE测量网选点问题的鲁棒选址集覆盖模型，设计了以最小化选取测试点为目标，求解鲁棒集覆盖问题的贪婪算法。算法对备选点集的元素进行处理，再应用贪婪步骤求解。仿真案例的测试结果表明：所提出的鲁棒选址集覆盖模型和贪婪算法对解决QoE测量网选点问题的可行性和有效性。 关键词：设施选址，集覆盖，鲁棒优化，贪婪算法 Robust Facility Location Problem of Test Points in Service Networks for Measuring QoE NING Yingdan 1, YANG Wenguo 1,2，GAO Suixiang 1,2 (1 School of Mathematical Sciences University of Chinese Academy of Sciences, Beijing 100049, China 2 Key Laboratory of Big Data Mining and Knowledge Management，Chinese Academy of Sciences) Abstract: Test points selecting in service networks for measuring QoE is to select as few points as possible to test the QoE of different website and reflect the qualities of all users’ different services under different scenarios. This is a special kind of location problem, which belongs to NP hard and the uncertainty factors in Test points selecting in service networks for measuring QoE increase the difficulty dramatically. To solve this problem, a robust set covering facility location model is established based on different coverage scenarios. The minimum number of test points to cover all users could be obtained by a heuristic algorithm based on the greedy procedure. The main idea of the algorithm is that process the elements in the sets of points before applying greedy procedure and remove unnecessary sets in last step. Experimental results demonstrate that both robust set covering facility location model and the algorithm proposed in this paper are efficient for solving the problem. Keywords: Facility Location, Set Covering, Robust Optimization, Greedy Algorithm 1 引言 近几年来，随着互联网的普及和高速宽带基础网络设施的建设，更多用户能够通过PC 机快速地访问互联网资源，更方便地获得包括高品质声音、影像等多媒体信息在内的各种数字化信息。面对用户对高品质音、视频需求的进一步增长，利用现有的网络资源为用户提供更好的宽带体验是亟待解决的问题。 QoE（Quality of Experience）[1]是指网络用户感觉到的网络“质量”或“性能”或“舒适度”，是以用户体验认可程度为标准的网络服