On the Performance of Nash Equilibria for Data Preservation in Base Station-less Sensor Networks
On the Performance of Nash Equilibria for Data Preservation in Base Station-less Sensor Networks
复制标题
DOI：
10.1109/mass58611.2023.00038
复制
发表时间：
2023-09
期刊：
影响因子：
--
通讯作者：
Giovanni Rivera;Yutian Chen;Bin Tang
中科院分区：
文献类型：
--
作者：
Giovanni Rivera;Yutian Chen;Bin Tang
无基站传感器网络（BSNs）是指部署在具有挑战性的环境（例如水下勘探）中的新兴传感应用。由于在这种环境下安装基站是不可行的，所以BSN中产生的数据在被上传机会收集之前会被保存在网络中。这个过程在BSN中称为数据保存。考虑到传感器节点的智能性和自私自利性，本文研究了基于纳什均衡的bsn数据保存方法。我们设计了一套数据保存算法，检查它们是否达到了网元，并使用现有的（即无政府状态的价格和稳定的价格）和我们自己设计的指标（即效率损失率）严格分析网元的性能。我们发现基于最小成本流的算法产生的网元在数据保存过程中以最小的能量消耗达到社会最优。我们证明了我们直接设计的贪心算法的NE达到了3的无政府价格。另一方面，我们证明了贪婪算法总是存在的（尽管不直接），实现了社会最优的NE。最后，我们进行了大量的仿真来研究各种数据保存网在不同网络参数下的性能，并验证了我们的理论结果。
Base station-less sensor networks (BSNs) refer to emerging sensing applications deployed in challenging environments (e.g., underwater exploration). As installing a base station in such an environment is not feasible, data generated in the BSN will be preserved in the network before being collected by the uploading opportunities. This process is called data preservation in the BSN. Considering that sensor nodes are intelligent and selfish, this paper studies the Nash Equilibrium (NE) for data preservation in BSNs. We design a suite of data preservation algorithms, examine whether they achieve NEs, and rigorously analyze the performance of the NEs using existing (i.e., price of anarchy and price of stability) and our own designed metrics (i.e., rate of efficiency loss). We find that a minimum cost flow-based algorithm produces a NE that achieves the social optimal with minimum energy consumption in the data preservation process. We show the NE from one of our straightforwardly designed greedy algorithms achieves a price of anarchy of 3. On the other hand, we prove that a greedy algorithm always exists (although non-straightforward), achieving the socially optimal NE. Finally, we conduct extensive simulations to investigate the performances of various data preservation NEs and validate our theoretical results under different network parameters.