[Paper Review] A Dynamic Resource Allocation Framework for Synchronizing Metaverse with IoT Service and Data

Yuexin Han, Dusit Niyato+3 more|arXiv (Cornell University)|Oct 31, 2021

Modular Robots and Swarm Intelligence17 citations

TL;DR

This paper proposes a hybrid evolutionary game-theoretic framework for dynamic resource allocation in Metaverse-IoT synchronization, where self-interested IoT device owners adaptively select virtual service providers (VSPs) based on payoff feedback. The framework achieves stable equilibrium states with faster convergence when using the Smith protocol, ensuring efficient and scalable data collection for digital twin applications.

ABSTRACT

Spurred by the severe restrictions on mobility due to the COVID-19 pandemic, there is currently intense interest in developing the Metaverse, to offer virtual services/business online. A key enabler of such virtual service is the digital twin, i.e., a digital replication of real-world entities in the Metaverse, e.g., city twin, avatars, etc. The real-world data collected by IoT devices and sensors are key for synchronizing the two worlds. In this paper, we consider the scenario in which a group of IoT devices are employed by the Metaverse platform to collect such data on behalf of virtual service providers (VSPs). Device owners, who are self-interested, dynamically select a VSP to maximize rewards. We adopt hybrid evolutionary dynamics, in which heterogeneous device owner populations can employ different revision protocols to update their strategies. Extensive simulations demonstrate that a hybrid protocol can lead to evolutionary stable states.

Motivation & Objective

• To address the challenge of efficient, scalable resource allocation for synchronizing real-world IoT data with the Metaverse in the absence of full rationality among device owners.
• To model the strategic selection of VSPs by heterogeneous IoT device owners as a multi-population evolutionary game with bounded rationality.
• To analyze the existence and stability of equilibrium strategies in a dynamic, self-adapting system where device owners update strategies based on payoff imitation.
• To evaluate the impact of different revision protocols—particularly the Smith protocol—on convergence speed and system equilibrium.
• To demonstrate the feasibility and robustness of the framework through extensive simulations under realistic IoT and Metaverse system parameters.

Proposed method

• Formulates a multi-population evolutionary game where device owners in different populations use distinct revision protocols (e.g., pairwise imitation, Smith protocol) to update strategies based on payoff comparisons.
• Models the utility of each device owner as a function of sensing data quality, energy cost, task reward, and route distance, with a general reward allocation scheme.
• Introduces a hybrid dynamics framework that combines multiple revision protocols across heterogeneous device owner populations to model realistic, bounded-rational decision-making.
• Employs a direction field visualization technique to analyze the stability of equilibrium states in a four-dimensional strategy space.
• Defines convergence time as the number of iterations until payoff differences across strategies in a population fall below a threshold (τ = 0.05), enabling quantitative evaluation of adaptation speed.
• Uses simulation with realistic parameters: UAV populations (Np ∈ [50,250]), sensing routes (Dm ∈ [1,1.8] km), rewards (Rm ∈ [1000,2000]), and energy cost (ζ = 0.001 $/Joule).

Experimental results

Research questions

• RQ1 Does a stable equilibrium exist in a multi-population, heterogeneous evolutionary game where IoT device owners self-select VSPs based on payoff feedback?
• RQ2 How does the adoption of different revision protocols—especially the Smith protocol—affect convergence speed to equilibrium?
• RQ3 What is the impact of varying the probability of using centralized payoff information (Smith protocol) on the stability and convergence of strategy distributions?
• RQ4 Can the system achieve a stationary strategy distribution where no device owner has incentive to unilaterally change its VSP selection?
• RQ5 How do differences in data quality and reward pools influence the long-term strategy distribution and payoff equilibrium across device populations?

Key findings

• An evolutionary stable equilibrium exists in the proposed hybrid evolutionary game framework, where no device owner has an incentive to unilaterally change its VSP selection strategy.
• The equilibrium is stable, as confirmed by direction field analysis showing that initial strategy distributions evolve toward equilibrium points regardless of initial conditions.
• Convergence time decreases significantly with higher adoption of the Smith protocol: increasing α³,¹ from 0 to 1 reduces convergence time, due to access to complete payoff information.
• Population-3 UAVs achieve higher average payoffs than populations 1 and 2 due to superior sensing data quality (b³ₘ ∈ [1,5]), influencing strategy distribution and equilibrium outcomes.
• Multiple equilibrium states exist, including cases where one strategy (e.g., selecting region 3) becomes extinct, indicating system resilience to strategy dominance.
• The system reaches convergence when payoff differences across strategies in a population fall below τ = 0.05, confirming the practicality of the convergence criterion in dynamic environments.