Joint Bidding on Intraday and Frequency Containment Reserve Markets

Battle Count: 58

Highly relevant to quantitative trading in electricity markets. The paper addresses multi-market portfolio allocation, continuous order book trading, and ML-based strategy selection - all core quantitative trading concepts. The rolling intrinsic approach is analogous to delta-hedging in options markets. The classification-based strategy selection parallels regime detection in trading. The work demonstrates practical algorithmic trading with sub-2-minute daily execution time. However, it is domain-specific to electricity/energy markets rather than traditional financial instruments.

High complexity due to: (1) MILP formulation for the intrinsic problem with order book data, (2) rolling re-optimization every minute with state tracking, (3) SoC constraint management across FCR commitments, (4) feature engineering with ~300 features including interaction terms and correlation filtering, (5) anchored walk-forward validation with hyperparameter tuning, (6) strategy pool selection via a separate MILP, (7) handling of energy drift from FCR activations, (8) rounding to minimum bid sizes, and (9) integration of all components into a coherent daily trading pipeline. Requires expertise in optimization, ML, and electricity market mechanics.

The paper provides detailed algorithm descriptions (Algorithm 1 for RI), market rules, feature engineering steps, and hyperparameter search spaces. Historical data sources are publicly available (EPEX Spot, Regelleistung.net, Bundesnetzagentur). However, no code repository is provided, and the exact feature engineering pipeline, MILP formulation details for strategy pool selection, and specific XGBoost hyperparameter configurations are described but not fully reproducible without implementation. The backtest period (2023-2024) and battery specifications (10MW/10MWh) are clearly stated.