Thermal generation planning strategy facilitating units decomposition by particle swarm optimization and multi-stage dynamic programming

This paper presents a methodology for determining optimal generation schedule for thermal units by using a combined approach of particle swarm optimization (PSO) and dynamic programming (DP). At first, the units are decomposed into several hours based on the forecasted load demand. Then unit commitment (UC) problem is solved for each single hour using binary version of PSO. While solving hourly UC, the decoupled constraints such as system power balance, system spinning reserve are considered. A number of better solutions are stored for each hour which will be applied generating the multi stage graph for merging procedure. Then the decomposed hourly units' schedules are merged to produce the final solution. For merging procedure, this method applies multi-stage dynamic programming approach. Coupling constraints such as ramp rate, minimum up/down time constraint are integrated with that DP approach. The simulation results show the effectiveness of this algorithm by comparing the outcome with several established methods.