• A knowledge-based genetic algorithm for unit commitment

      Aldridge, C.J.; McKee, S.; McDonald, J.R.; Galloway, S.J.; Dahal, Keshav P.; Bradley, M.E.; Macqueen, J.F. (2001)
      A genetic algorithm (GA) augmented with knowledge-based methods has been developed for solving the unit commitment economic dispatch problem. The GA evolves a population of binary strings which represent commitment schedules. The initial population of schedules is chosen using a method based on elicited scheduling knowledge. A fast rule-based dispatch method is then used to evaluate candidate solutions. The knowledge-based genetic algorithm is applied to a test system of ten thermal units over 24-hour time intervals, including minimum on/off times and ramp rates, and achieves lower cost solutions than Lagrangian relaxation in comparable computational time.