Incentive mechanisms for smartphone collaboration in data acquisition and distributed computing
Incentive mechanisms for smartphone collaboration in data acquisition and distributed computing
复制标题
DOI：
10.1109/infcom.2012.6195541
复制
发表时间：
2012-03
期刊：
影响因子：
--
通讯作者：
Lingjie Duan;T. Kubo;Kohei Sugiyama;Jianwei Huang;T. Hasegawa;J. Walrand
中科院分区：
文献类型：
--
作者：
Lingjie Duan;T. Kubo;Kohei Sugiyama;Jianwei Huang;T. Hasegawa;J. Walrand
本文分析并比较了客户端的不同激励机制，以激励智能手机用户在数据采集和分布式计算应用程序上的协作。从大量用户获取数据对于构建丰富的数据库和支持新兴的基于位置的服务至关重要。我们提出了一种基于奖励的协作机制，客户宣布总奖励由协作者共享，如果有足够多的用户愿意协作，则协作成功。我们表明，如果客户知道用户的协作成本，那么他可以选择通过提供少量的总奖励来仅涉及成本最低的用户。然而，如果客户不知道用户的私人成本信息，那么他需要提供更大的总奖励来吸引足够的合作者。用户将受益于在数据采集之前了解其成本。分布式计算旨在以分布式且廉价的方式解决计算密集型问题。我们研究客户如何通过为不同的用户类型指定不同的任务奖励组合来设计最佳合约。在完整信息下，我们表明只要客户对该类型的偏好超过相应的成本，客户就会涉及该类型的用户。在这种情况下，所有合作者都获得零回报。但如果客户端不知道用户的私人成本信息，他会保守地瞄准一小部分成本较小的高效用户。他必须给合作者最大的好处，而合作者的回报会提高他的计算效率。
This paper analyzes and compares different incentive mechanisms for a client to motivate the collaboration of smartphone users on both data acquisition and distributed computing applications. Data acquisition from a large number of users is essential to build a rich database and support emerging location-based services. We propose a reward-based collaboration mechanism, where the client announces a total reward to be shared among collaborators, and the collaboration is successful if there are enough users willing to collaborate. We show that if the client knows the users' collaboration costs, then he can choose to involve only users with the lowest costs by offering a small total reward. However, if the client does not know users' private cost information, then he needs to offer a larger total reward to attract enough collaborators. Users will benefit from knowing their costs before the data acquisition. Distributed computing aims to solve computational intensive problems in a distributed and inexpensive fashion. We study how the client can design an optimal contract by specifying different task-reward combinations for different user types. Under complete information, we show that the client will involve a user type as long as the client's preference for that type outweighs the corresponding cost. All collaborators achieve a zero payoff in this case. But if the client does not know users' private cost information, he will conservatively target at a smaller group of efficient users with small costs. He has to give most benefits to the collaborators, and a collaborator's payoff increases in his computing efficiency.