Embedding the priority list into tabu search for unit commitment

This paper presents a new tabu search based method for the unit commitment in power systems. The formulation of the unit commitment may be described as nonlinear mixed integer programming. However, it is hard to optimize a problem with discrete and continuous variables in a large-scale system at the same time. In this paper, the problem is decomposed into two phases. One handles on-off conditions of generators with tabu search (TS) while the other determines output variables of generators using the equal lambda method. TS provides better solutions through the neighborhood search with the adaptive memory. However, TS is inclined to increase the solution candidates in the neighborhood in a large system. This paper proposes an efficient method that reduces the solution candidates of TS with the priority list of units. The expensive or cheap generators are fixed to speed up the neighborhood search. The effectiveness of the proposed method is demonstrated in 10-unit and 54-unit systems