Unit Commitment

Unit commitment (UC) is the optimization problem of determining which generators should be switched on (committed) or off over a scheduling horizon, typically 24 hours to one week ahead, to meet forecast demand at minimum total cost while satisfying a wide set of technical and security constraints.

Key Aspects of Unit Commitment:

• Binary Decisions: Unlike economic dispatch, which allocates continuous power outputs among already-running units, unit commitment involves on/off (binary) decisions for each generator at each time period, making it a mixed-integer optimization problem that is inherently more complex.
• Startup and Shutdown Costs: Starting a thermal generator requires fuel, time, and wear on equipment. Startup costs can range from thousands to hundreds of thousands of euros depending on unit type and how long it has been offline (hot, warm, or cold start). These costs are central to the UC decision.
• Minimum Up/Down Times: Once started, a generator must typically run for a minimum number of hours (minimum up time) before it can be shut down, and vice versa (minimum down time). These constraints prevent excessive cycling that damages equipment.
• Solution Methods: Due to the combinatorial explosion of binary variables, UC is solved using techniques such as mixed-integer linear programming (MILP), Lagrangian relaxation, dynamic programming, or heuristic approaches. Modern solvers can handle systems with thousands of units over multi-day horizons.
• Day-Ahead Markets: In organized electricity markets, unit commitment forms the basis of the day-ahead market clearing process: generators submit cost and availability data, and the market operator solves a security-constrained UC to determine which units are committed and at what schedule for the following day.