Abstract
Vehicular Metaverses are developed to enhance the modern automotive industry with an immersive and safe experience among connected vehicles and roadside infrastructures, e.g., RoadSide Units (RSUs). For seamless synchronization with virtual spaces, Vehicle Twins (VTs) are constructed as digital representations of physical entities. However, resource-intensive VTs updating and high mobility of vehicles require intensive computation, communication, and storage resources, especially for their migration among RSUs with limited coverages. To address these issues, we propose an attribute-aware auction-based mechanism to optimize resource allocation during VTs migration by considering both price and non-monetary attributes, e.g., location and reputation. In this mechanism, we propose a two-stage matching for vehicular users and Metaverse service providers in multi-attribute resource markets. First, the resource attributes matching algorithm obtains the resource attributes perfect matching, namely, buyers and sellers can participate in a double Dutch auction (DDA). Then, we train a DDA auctioneer using a generative pre-trained transformer (GPT)-based deep reinforcement learning (DRL) algorithm to adjust the auction clocks efficiently during the auction process. Simulation results show that our proposed GPT-based DRL auction schemes have better performance than others.
Introduction
In a Vehicular Metaverse, vehicles rely on Vehicle Twins (VTs)—resource-intensive digital models that sync physical and virtual spaces for applications like AR navigation and 3D entertainment. Because Roadside Units (RSUs) have limited coverage and vehicles move quickly, these VTs must frequently "migrate" from one RSU to another to maintain a seamless experience.
Allocating resources for these migrations is a complex challenge. Standard auctions usually focus solely on price, but for a metaverse to feel immersive, factors like the RSU’s physical distance (latency) and its historical reliability (reputation) are equally critical. Furthermore, traditional auction mechanisms are often slow or computationally expensive, making them ill-suited for the split-second decisions required in high-speed vehicular environments.
The MADDA Framework
To address these issues, this paper introduces a two-stage Multi-Attribute Double Dutch Auction (MADDA) mechanism. This framework optimizes resource allocation by balancing monetary bids with critical non-monetary attributes.
Stage 1: Resource-Attribute Matching
The system uses the Kuhn-Munkres (KM) algorithm to create a weighted bipartite graph. This ensures that vehicles (buyers) are only matched with RSUs (sellers) that meet their specific needs for CPU, storage, and bandwidth, while also factoring in location and reputation.
Stage 2: Winner Determination and Pricing
Once matched, the parties enter a Double Dutch Auction. Here, the "buyer clock" starts high and ticks down, while the "seller clock" starts low and ticks up. The auction concludes when the prices meet, establishing a clearing price.
GPT-based DRL Optimization
A key innovation is treating the auction rounds as a sequence modeling problem. By applying a Generative Pre-trained Transformer (GPT)-based DRL agent, the system learns to adjust the auction clocks dynamically. This allows the auction to reach a clearing price much faster and more efficiently than standard reinforcement learning algorithms like Soft Actor-Critic (SAC).
Key Results and Insights
The framework was tested in a simulation environment involving 50 to 200 Vehicular Users (VUs) and RSUs. Key findings include:
- Higher Social Welfare: The GPT-based DRL scheme achieved a social welfare score approximately 9.2% higher than SAC and 22.8% higher than random methods.
- Efficiency: The GPT approach significantly reduced "information exchange costs," reaching deals in fewer steps—critical for low-latency vehicular environments.
- Reputation Freshness: The model includes a time-decay factor for reputation. This ensures that vehicles are not migrated to infrastructure that was reliable in the past but has recently become unstable, resulting in a 42% to 51% improvement in rewards compared to baseline methods.
- Economic Robustness: The authors verified that the mechanism maintains Individual Rationality (no participant loses money) and Incentive Compatibility (truthful bidding is the optimal strategy).
Conclusion
By treating auctions as sequences and incorporating non-monetary attributes, this approach provides a scalable solution for resource management in the Vehicular Metaverse. Future work will focus on making the GPT-based DRL more robust to real-world network jitter and exploring predictive migration patterns to further reduce latency.
Summary updated