An adaptive random search for short term generation scheduling with network constraints.

Area of Science:

• Electrical Engineering
• Operations Research

Background:

• Short-term generation scheduling is crucial for power system stability.
• Integrating network constraints (capacity, losses) complicates optimization.
• Existing methods may face computational challenges with complex models.

Purpose of the Study:

• To develop an efficient heuristic for short-term generation scheduling.
• To incorporate transmission network constraints into the scheduling model.
• To improve computational efficiency over traditional optimization solvers.

Main Methods:

• An adaptive random search heuristic based on Markov Chain Monte Carlo.
• Population-based approach with adaptive noise level control for exploration/exploitation.
• Integration of a local search mechanism to refine solutions.

Main Results:

• The proposed heuristic effectively schedules thermal units under network constraints.
• Demonstrated significant reduction in computational effort compared to a commercial solver.
• Validated performance and robustness across multiple test systems.

Conclusions:

• The adaptive random search offers a computationally efficient and robust solution for short-term generation scheduling.
• The method successfully handles network constraints, including capacity limits and line losses.
• This approach presents a viable alternative to complex commercial optimization solvers for power systems.