An Iterative Deepening Genetic Algorithm for Scheduling of Direct Load Control

Abstract

A modified genetic algorithm (GA) called iterative deepening GA (IDGA) is proposed in this paper to optimize the scheduling of direct load control (DLC) strategies. The control strategy (or scheduling) arranged by the IDGA not only sheds the load so that the load required to be shed at each sampling interval is individually satisfied, but it also minimizes the shedding load so that the utility company's revenue loss due to DLC is minimized. The scheduling obtained by the proposed IDGA tends to level off the accumulated shedding time of each load group, avoiding customers' complaints about fairness of shedding time. IDGA is composed of a master GA and a sequence of slave GAs. As the master GA evaluates a status combination, it iteratively calls upon a slave GA at each of the following time steps, evaluating possible forward status combinations. With an iteratively deepening search scope, IDGA is able to find a satisfactory suboptimal scheduling.