[Paper Review] Strategic Contention Resolution with Limited Feedback

George Christodoulou, Martin Gairing+3 more|arXiv (Cornell University)|Jun 21, 2016

Network Traffic and Congestion Control4 references3 citations

TL;DR

This paper studies equilibrium contention resolution in multi-access channels under acknowledgment-based feedback, where users only learn if their own transmissions succeed. It presents a two-player equilibrium protocol with finite expected latency and proves impossibility results for age-based and backoff protocols with finite expected time in larger groups, while introducing a novel age-based protocol with linear completion time with high probability despite infinite expected latency.

ABSTRACT

In this paper, we study contention resolution protocols from a game-theoretic perspective. We focus on \emph{acknowledgment-based} protocols, where a user gets feedback from the channel only when she attempts transmission. In this case she will learn whether her transmission was successful or not. Users that do not transmit will not receive any feedback. We are interested in equilibrium protocols, where no player has an incentive to deviate. The limited feedback makes the design of equilibrium protocols a hard task as best response policies usually have to be modeled as Partially Observable Markov Decision Processes, which are hard to analyze. Nevertheless, we show how to circumvent this for the case of two players and present an equilibrium protocol. For many players, we give impossibility results for a large class of acknowledgment-based protocols, namely \emph{age-based} and \emph{backoff} protocols with finite expected finishing time. Finally, we provide an age-based equilibrium protocol, which has infinite expected finishing time, but every player finishes in linear time with high probability.

Motivation & Objective

• To design incentive-compatible contention resolution protocols under limited acknowledgment-based feedback, where users only learn the outcome of their own transmissions.
• To analyze whether equilibrium protocols can be constructed for multiple players when feedback is restricted to success/failure of one's own transmissions.
• To investigate the limitations of age-based and backoff protocols in achieving equilibrium with finite expected finishing time under strategic user behavior.
• To identify conditions under which equilibrium protocols exist, particularly focusing on two-player and many-player scenarios.
• To provide a novel age-based protocol that ensures linear completion time with high probability, even if expected latency is infinite.

Proposed method

• Models the contention resolution problem as a non-cooperative stochastic game where each user acts selfishly to minimize expected transmission time.
• Uses Markov decision processes with partial observability to analyze best response strategies under limited feedback, focusing on equilibrium conditions.
• Applies a structural lemma (Lemma 4.1) showing that any consistent transmission sequence in an equilibrium protocol must preserve expected latency when deviating to a modified protocol.
• Employs contradiction-based proofs to show that no equilibrium can exist for age-based or backoff protocols with finite expected latency when more than two players are involved.
• Constructs a specific two-player equilibrium protocol with $ p_1 = \frac{2}{3} $, $ p_2 = 1 $, and proves its optimality via expected latency comparison.
• Introduces a new age-based protocol with infinite expected latency but linear completion time with high probability, using a carefully designed sequence of transmission probabilities.

Experimental results

Research questions

• RQ1 Can an equilibrium protocol be designed for two users under acknowledgment-based feedback with finite expected latency?
• RQ2 Are age-based or backoff protocols with finite expected finishing time implementable as equilibria in systems with more than two players?
• RQ3 What structural constraints must an equilibrium protocol satisfy under limited feedback, particularly regarding consistent transmission sequences?
• RQ4 Can a protocol be constructed that ensures linear completion time with high probability, even if expected latency is infinite?
• RQ5 What are the necessary and sufficient conditions for a protocol to be in equilibrium under acknowledgment-based feedback with strategic users?

Key findings

• For two players, an equilibrium protocol exists with $ p_1 = \frac{2}{3} $, $ p_2 = 1 $, achieving finite expected latency and optimal performance.
• Any equilibrium protocol for two players must have $ p_1 = \frac{2}{3} $ and $ p_2 = 1 $; any deviation in these values leads to a profitable deviation for at least one player.
• For $ n \geq 3 $, no age-based or backoff protocol with finite expected finishing time can be in equilibrium, due to structural contradictions in expected latency under unilateral deviation.
• An age-based equilibrium protocol exists with infinite expected latency but ensures that every player finishes in linear time with high probability.
• The proof relies on a key lemma showing that consistent transmission sequences in equilibrium must preserve expected latency under protocol modifications, enabling contradiction-based analysis.
• The impossibility result holds for all protocols where $ p_2 < 1 $, as such configurations lead to a contradiction when comparing expected latencies of alternative deviation strategies.