[Paper Review] Dynamic Spectrum Sharing Among Repeatedly Interacting Selfish Users With Imperfect Monitoring

Yuanzhang Xiao, Mihaela van der Schaar|arXiv (Cornell University)|Jan 16, 2012

Cognitive Radio Networks and Spectrum Sensing34 references4 citations

TL;DR

This paper proposes a deviation-proof dynamic spectrum sharing framework for selfish secondary users (SUs) in cognitive radio networks with imperfect monitoring. Using a repeated game model with bounded rationality, it designs distributed, time-division multiple access (TDMA)-compatible policies that achieve Pareto-optimal performance even under limited and noisy interference observations, outperforming constant-power policies in high-interference regimes.

ABSTRACT

We develop a novel design framework for dynamic distributed spectrum sharing among secondary users (SUs) who adjust their power levels to compete for spectrum opportunities while satisfying the interference temperature (IT) constraints imposed by primary users. The considered interaction among the SUs is characterized by the following three features. First, since the SUs are decentralized, they are selfish and aim to maximize their own long-term payoffs from utilizing the network rather than obeying the prescribed allocation of a centralized controller. Second, the SUs interact with each other repeatedly and they can coexist in the system for a long time. Third, the SUs have limited and imperfect monitoring ability: they only observe whether the IT constraints are violated, and their observation is imperfect due to the erroneous measurements. To capture these features, we model the interaction of the SUs as a repeated game with imperfect monitoring. We first characterize the set of Pareto optimal payoffs that can be achieved by deviation-proof spectrum sharing policies, which are policies that the selfish users find it in their interest to comply with. Next, for any given payoff in this set, we show how to construct a deviation-proof policy to achieve it. The constructed deviation-proof policy is amenable to distributed implementation, and allows users to transmit in a time-division multiple-access (TDMA) fashion. In the presence of strong multi-user interference, our policy outperforms existing spectrum sharing policies that dictate users to transmit at constant power levels simultaneously. Moreover, our policy can achieve Pareto optimality even when the SUs have limited and imperfect monitoring ability, as opposed to existing solutions based on repeated games, which require perfect monitoring abilities.

Motivation & Objective

• To design a distributed, deviation-proof spectrum sharing policy for selfish secondary users (SUs) who interact repeatedly and have limited, imperfect monitoring of interference levels.
• To ensure that SUs comply with the policy by making deviation unprofitable, even when they are self-interested and cannot observe exact power levels of others.
• To achieve Pareto-optimal performance in terms of user throughput under interference temperature (IT) constraints, despite imperfect monitoring and strong multi-user interference.
• To enable time-division multiple access (TDMA) operation through a policy that allows dynamic, time-varying power allocation, improving spectral efficiency over constant-power schemes.
• To construct a self-generating set of equilibrium payoffs that remains stable under imperfect monitoring and bounded rationality, using continuation payoff mechanisms.

Proposed method

• Models the interaction among SUs as a repeated game with imperfect monitoring, where users only observe whether IT constraints are violated, not exact interference levels.
• Introduces a continuation payoff mechanism that ensures incentive compatibility by linking future rewards to current compliance, using a punishment phase triggered by detected deviations.
• Derives a set of necessary and sufficient conditions for a payoff vector to be self-generating and equilibrium-achievable, based on discount factors and monitoring error probabilities.
• Constructs a deviation-proof policy via a decomposition of the target payoff into current action and continuation payoffs, using a user-specific index function to determine which user transmits at each stage.
• Uses a recursive algorithm to determine action profiles over time based on the current continuation payoff, ensuring that the policy remains within the self-generating set of equilibrium payoffs.
• Employs a key equation involving the ratio $ \frac{v_i}{\bar{v}_i} $ and monitoring error probabilities $ \rho(y_0|\mathbf{\tilde{p}}^i) $ to compute the minimum discount factor $ \underline{\delta}(\bm{\mu}) $ required for stability.

Experimental results

Research questions

• RQ1 Can a distributed spectrum sharing policy be designed such that selfish SUs have no incentive to deviate, even when they only observe imperfect signals about interference violations?
• RQ2 How can Pareto-optimal performance be achieved in a dynamic spectrum sharing system under imperfect monitoring and strong multi-user interference?
• RQ3 What is the minimal discount factor required for a repeated-game strategy to be incentive-compatible under imperfect monitoring and bounded rationality?
• RQ4 Can a time-division multiple access (TDMA)-like transmission strategy be embedded into a repeated-game framework to improve spectral efficiency over constant-power policies?
• RQ5 What conditions ensure that a set of equilibrium payoffs is self-generating under imperfect monitoring and user-specific continuation payoffs?

Key findings

• The proposed policy achieves Pareto-optimality in the feasible QoS region even with imperfect monitoring, unlike prior repeated-game approaches that require perfect monitoring.
• The constructed deviation-proof policy enables SUs to transmit in a TDMA-like fashion, significantly outperforming constant-power policies in high-interference scenarios with nonconvex feasible QoS regions.
• The minimum discount factor $ \underline{\delta}(\bm{\mu}) $ required for equilibrium is derived explicitly as $ \underline{\delta}(\bm{\mu}) = \frac{1}{1+z} $, where $ z $ depends on monitoring errors and channel gains.
• The self-generating set $ \mathcal{B}_{\bm{\mu}} $ of equilibrium payoffs is both necessary and sufficient for stability, ensuring that any payoff in the set can be sustained via a consistent strategy.
• Simulation results validate the analytical findings, showing significant performance gains in spectral efficiency and fairness compared to constant-power and existing repeated-game policies.
• The policy is amenable to distributed implementation, as each SU only needs to monitor local IT constraint violations and update its strategy based on a local index function.