Energy Minimization for Participatory Federated Learning in IoT Analyzed via Game Theory
Authors:
Alessandro Buratto,
Elia Guerra,
Marco Miozzo,
Paolo Dini,
Leonardo Badia
Abstract:
The Internet of Things requires intelligent decision making in many scenarios. To this end, resources available at the individual nodes for sensing or computing, or both, can be leveraged. This results in approaches known as participatory sensing and federated learning, respectively. We investigate the simultaneous implementation of both, through a distributed approach based on empowering local no…
▽ More
The Internet of Things requires intelligent decision making in many scenarios. To this end, resources available at the individual nodes for sensing or computing, or both, can be leveraged. This results in approaches known as participatory sensing and federated learning, respectively. We investigate the simultaneous implementation of both, through a distributed approach based on empowering local nodes with game theoretic decision making. A global objective of energy minimization is combined with the individual node's optimization of local expenditure for sensing and transmitting data over multiple learning rounds. We present extensive evaluations of this technique, based on both a theoretical framework and experiments in a simulated network scenario with real data. Such a distributed approach can reach a desired level of accuracy for federated learning without a centralized supervision of the data collector. However, depending on the weight attributed to the local costs of the single node, it may also result in a significantly high Price of Anarchy (from 1.28 onwards). Thus, we argue for the need of incentive mechanisms, possibly based on Age of Information of the single nodes.
△ Less
Submitted 27 March, 2025;
originally announced March 2025.
Static and Repeated Cooperative Games for the Optimization of the AoI in IoT Networks
Authors:
David Emanuele Corrado Raphael Catania,
Alessandro Buratto,
Giovanni Perin
Abstract:
Wireless sensing and the internet of things (IoT) are nowadays pervasive in 5G and beyond networks, and they are expected to play a crucial role in 6G. However, a centralized optimization of a distributed system is not always possible and cost-efficient. In this paper, we analyze a setting in which two sensors collaboratively update a common server seeking to minimize the age of information (AoI)…
▽ More
Wireless sensing and the internet of things (IoT) are nowadays pervasive in 5G and beyond networks, and they are expected to play a crucial role in 6G. However, a centralized optimization of a distributed system is not always possible and cost-efficient. In this paper, we analyze a setting in which two sensors collaboratively update a common server seeking to minimize the age of information (AoI) of the latest sample of a common physical process. We consider a distributed and uncoordinated setting where each sensor lacks information about whether the other decides to update the server. This strategic setting is modeled through game theory (GT) and two games are defined: i) a static game of complete information with an incentive mechanism for cooperation, and ii) a repeated game over a finite horizon where the static game is played at each stage. We perform a mathematical analysis of the static game finding three Nash Equilibria (NEs) in pure strategies and one in mixed strategies. A numerical simulation of the repeated game is also presented and novel and valuable insight into the setting is given thanks to the definition of a new metric, the price of delayed updates (PoDU), which shows that the decentralized solution provides results close to the centralized optimum.
△ Less
Submitted 12 May, 2025; v1 submitted 27 March, 2025;
originally announced March 2025.
