When Users Interfere with Protocols: Prospect Theory in Wireless Networks using Random Access and Data Pricing as an Example

Tianming Li, Narayan B. Mandayam

Abstract: Game theoretic models have found widespread use in the analysis and engineered system design of radio resource management algorithms for a wide variety of systems such as cellular, ad hoc and sensor networks. The fundamental principle behind such models and much of game theory has been the reliance on Expected Utility Theory (EUT). Motivated by the increasing amount of end-user control afforded in programmable radio devices, we envision a scenario where end-user actions essentially "interfere" with the underlying engineered system design. As an exemplary scenario, we consider in this paper wireless random access where players follow the precepts of Prospect Theory (PT), a theory developed by Kahneman and Tversky to explain real-life decision making that often deviates from the behavior expected under EUT. Specifically, we consider a game where selfish players adjust their transmission probabilities over a random access channel under throughput rewards, delay penalties and energy costs. By analyzing the Nash Equilibrium achieved, we prove under mild conditions that deviations from EUT results in degradation of system throughput while increasing delay and energy consumption. Finally, we consider a data pricing model and study the impact of end-user decision-making (i.e., players service choices) at the NE on wireless network performance.