TL;DR
This paper proposes an "energy group buying with load sharing" strategy for mobile network operators (MNOs) to reduce their energy costs in the context of hybrid electricity markets and 5G networks. It introduces novel approaches for both fully cooperative and self-interested MNOs, demonstrating significant energy cost savings through collaborative purchasing and efficient resource utilization.
What Problem Does This Paper Address?
The rapid growth of mobile data traffic, driven by 5G technology, is leading to substantial increases in energy consumption and operational costs for mobile network operators (MNOs). Simultaneously, the electricity grid is evolving towards smart grids with dynamic pricing, creating a hybrid electricity market where day-ahead energy is cheaper than real-time energy. MNOs face the challenge of optimizing their energy purchases to minimize costs while meeting fluctuating traffic demands. Furthermore, cellular base stations (BSs) are often underutilized during off-peak hours, leading to energy inefficiency. This paper addresses these challenges by proposing collaborative strategies for MNOs to jointly optimize their energy procurement and wireless traffic management.
Methods
The paper proposes a two-pronged approach:
- Energy Group Buying with Load Sharing: This core concept involves two MNOs acting as a single entity to:
- Energy Group Buying: Jointly commit to day-ahead energy purchases and engage in real-time energy trading, leveraging the lower prices of the day-ahead market.
- Wireless Load Sharing: Allow paired base stations (BSs) from different MNOs to share wireless traffic. This enables lightly loaded BSs to be put into sleep mode, saving energy.
- Optimization Framework: A two-stage stochastic programming approach is used to find the optimal day-ahead commitments and real-time trading/load sharing strategies.
- Repeated Nash Bargaining Scheme (for self-interested MNOs): When MNOs are not fully cooperative, this scheme facilitates negotiation and cost sharing. It involves:
- Day-ahead Bargaining: Negotiating aggregated energy commitments and how to share them, considering future real-time cooperation benefits.
- Real-time Bargaining: Negotiating real-time energy trading, load sharing, and inter-MNO monetary payments to fairly distribute cost reductions. This is modeled as a Nash bargaining problem to ensure Pareto optimality and fairness.
Datasets
The paper uses simulated data based on:
- Practical day-ahead energy prices: Data from the PJM day-ahead energy market.
- Normalized wireless traffic profiles: Measurement results from practical cellular BSs.
- Power consumption parameters: Values for BS energy consumption (transmission, non-transmission, and sleep modes) are cited from existing literature.
The authors generate wireless traffic and prediction errors based on these profiles and parameters for simulation purposes.
Key Insights and Ideas
- Collaboration is Key: The most significant energy cost savings are achieved when MNOs collaborate, both in energy purchasing and wireless load sharing.
- Hybrid Market Opportunity: The dynamic pricing in hybrid electricity markets presents a clear opportunity for MNOs to reduce costs by strategically committing to day-ahead purchases.
- Load Sharing for Sleep Mode: Enabling BSs to share traffic is crucial for putting underutilized BSs into sleep mode, a major source of energy savings, especially during off-peak hours.
- Fairness in Self-Interest: For competing MNOs, a well-designed bargaining mechanism (like Nash bargaining) is essential to incentivize cooperation by ensuring a fair distribution of the achieved cost reductions.
- Day-Ahead vs. Real-Time Trade-offs: The optimal strategy involves balancing day-ahead commitments with the flexibility of real-time adjustments, considering prediction errors for both traffic and energy prices.
Key Results
- Significant Cost Reductions:
- Full Cooperation: Achieved an average of 22.81% energy cost reduction in symmetric traffic scenarios and 31.72% in asymmetric traffic scenarios compared to the non-cooperative benchmark.
- Repeated Nash Bargaining: Achieved comparable cost reductions to full cooperation (e.g., 22.69% and 22.92% for MNO 1 and MNO 2 respectively in symmetric traffic) while ensuring fairness.
- Pareto Optimality: The proposed schemes, particularly the Nash bargaining approach, are shown to achieve Pareto-optimal energy costs, meaning no MNO can improve its cost without making another MNO worse off.
- Impact of Traffic Patterns: Asymmetric traffic patterns can lead to larger overall cost reductions through cooperation, as the lightly loaded MNO can significantly benefit from sleep modes.
Limitations and Future Work
- Prediction Errors: The models account for prediction errors in energy prices and wireless traffic, but the accuracy of these predictions remains a factor.
- Simplified Models: The paper uses simplified models for wireless traffic and energy consumption. More complex and realistic models could be explored.
- Inter-MNO Payment Complexity: While monetary payments are proposed for self-interested MNOs, the practical implementation and negotiation of these payments can be complex.
- Extension to More MNOs: The current work focuses on two MNOs. Extending the framework to scenarios with more than two MNOs, potentially involving group bargaining, is a significant area for future research.
- Heterogeneous Networks (HetNets): The authors suggest that the proposed energy group buying and load sharing can be extended to heterogeneous networks with different types of BSs (e.g., macro, pico, femto), which presents an interesting avenue for future work, especially regarding inter-type BS load sharing.
Summary updated