Authors: Efstratios Vamvourellis, Ilias Theodoropoulos
Organization: Eight Bells LTD
In the constantly evolving world of 5G and Beyond-Radio Access Networks (B-RAN), dynamic pricing techniques are crucial for ensuring competitive fairness and network profitability. The Smart Pricing Module (SPM) is an innovative AI-driven NANCY subsystem that calculates the best pricing policy for network providers. The SPM, based on auction theory and reinforcement learning, allows suppliers to dynamically modify pricing while maximizing efficiency and profit.
How the Smart Pricing Module Works
The SPM is built around an AI-driven decision-making framework that uses modern reinforcement learning (RL) techniques. The RL-based approach allows AI agents to adapt to changes in network traffic and provider behaviors. It is designed to continuously gather and analyze critical data from the NANCY Marketplace about providers operating in a specific region. This data includes the minimum and maximum pricing ranges for services, provider IDs, and network congestion indicators for each provider.
The SPM employs this input data to simulate a multi-round blind reverse auction. In this auction-based system, each provider is represented by an AI agent who places bids to deliver services. Providers compete on price, and the agent with the most attractive bid wins the auction. This approach promotes competition, resulting in fair and dynamic pricing models that reflect real-time network demands and market conditions.
The design of the SPM depends on auction theory since it uses a blind auction mechanism to calculate fair and profitable pricing results by means of a simulated bidding process. Providers submit bids without knowing their competitors’ offers, ensuring an unbiased pricing environment. Unlike static pricing models that maintain fixed rates, the SPM promotes flexibility by dynamically adjusting to real-time network conditions.
Technical architecture and deployment
The Smart Pricing Module is designed with modern development tools and frameworks to ensure scalability and efficiency. The primary technical features are:
- Implementation with Python, which is ideal for AI-driven applications due to its extensive ecosystem of machine learning libraries and tools.
- Development of a REST API, for communicating with the NANCY Marketplace as a Blockchain oracle.
Use of production-grade AI libraries, for reinforcement learning, allowing agents to make data-driven decisions in auction settings.
Benefits of the Smart Pricing Module
The Smart Pricing Module provides considerable benefits for network operators, suppliers, and end-users in the B-RAN context.
- Continuous processing of incoming data, for keeping pricing strategies updated and consistent with network conditions.
- Auction-based pricing ensures provider profitability and competitiveness, preventing market monopolization.
- AI-driven reinforcement learning enables efficient and intelligent decision-making based on data.
Conclusion
For the NANCY 5G network, the Smart Pricing Module represents a significant step in dynamic pricing and AI-powered decision-making. Combining reinforcement learning with auction theory, the SPM gives vendors a flexible, competitive, efficient pricing solution.
If you are interested, deliverable D4.5 will offer a more thorough technical breakdown of the Smart Pricing Module.