Introduction: The Power Control Dilemma
In dense wireless environments, users compete for limited spectrum. To maintain a target Quality of Service (QoS), often defined by a Signal-to-Interference Ratio (SIR), devices may increase their transmission power. This creates a negative externality: higher power from one user increases interference for others, prompting them to also increase power. In overloaded networks, this can lead to a "power spiral" where transmission levels escalate indefinitely without any user achieving their QoS target.
While algorithms like Foschini-Miljanic can find stable power levels if they exist, they do not address the fundamental question of admission control: which users should be allowed to transmit in the first place? Traditional methods are often arbitrary. This research proposes a game-theoretic framework where users self-select based on pricing incentives.
Methodology: Incentive-Compatible Pricing
The authors model the network as a non-cooperative game. Each user $i$ aims to maximize a utility function $U_i = V_i(\gamma_i) - \text{Cost}_i$, where $V_i$ is the value derived from achieving a target SIR $\bar{\gamma}_i$. Three pricing schemes are evaluated to manage network load:
- Fixed Admission Pricing: A flat fee $C$ is charged to any participating user.
- Power-Based Pricing: Users pay $c \cdot p_i$, proportional to their transmission power. This penalizes "loud" transmitters.
- Interference-Based Pricing: Users pay $c \cdot g_i$, where $g_i$ represents their channel gain, serving as a proxy for the potential interference they cause to the network.
A Nash Equilibrium is reached when no user can improve their utility by unilaterally changing their decision to join or their transmission power.
Simulation and Evaluation
The schemes were tested in a simulated environment consisting of 25 transmitter-receiver pairs randomly distributed on a 1000x1000 grid. The propagation model assumed a $1/d^4$ power drop-off. User valuations were tied to the information-theoretic capacity of their links, $\hat{v}_i = \ln(1 + \bar{\gamma}_i)$.
Key Results: Efficiency and Stability
The study reveals that pricing acts as a "control knob" for network stability. When prices are too low, the system remains overloaded and fails to converge. As the cost parameter increases, users with low valuations or high interference footprints voluntarily drop out, allowing the system to reach a stable, Pareto-optimal state.
Interference-based pricing demonstrated superior performance over fixed pricing. By accounting for the "cost" of interference, it incentivizes a more efficient set of users to participate, leading to higher aggregate network value. Notably, power-based pricing exhibited non-monotonic behavior, where increasing costs could occasionally allow new users to join due to a lowered interference floor.
Conclusion and Future Directions
This game-theoretic approach shifts admission control from centralized exclusion to decentralized self-selection. While effective, the current model assumes binary utility (all-or-nothing QoS). Future work could explore "soft" QoS targets and the use of auctions to dynamically discover optimal prices in environments with asymmetric information. Additionally, proving the convergence of best-response dynamics in power-based pricing remains an open mathematical challenge.
Summary updated