Abstract Due to the limitation of active power exchange between dynamic voltage restorers (DVRs) in an interline DVR (IDVR), selection of the appropriate voltage for each DVR in an IDVR is very important. In previous studies, an optimization problem by selecting the “minimization of the sum of rated apparent power of DVRs” as the objective function was proposed to select the rated voltage of DVRs in an IDVR. In that problem, feeders load is assumed to be constant and equal to the rated values, while load variations in feeders affect active power exchange between the DVRs and their injected voltages. Therefore, load variations in feeders, being the main contribution of the current research, must be considered in the mentioned optimization problem. As the number of time intervals in a load curve is big, the solution space will be very large and thus, searching the entire solution space is not possible. Therefore, it is essential to use smart optimization methods For this purpose, genetic algorithm (GA) is used. By proposing a convenient strategy based on a GA for considering load variations in feeders, the sum of the rating of DVRs is minimized. In order to prove the ability of the proposed method, various examples of load variations in feeders (2, 3, and 6 time intervals) are provided. Then, using the data obtained from an actual industrial area, optimal values are determined for exiting DVRs in an IDVR structure in the case of a 24-time interval variation of feeders load.