Evolutionary techniques, a sensitivity-based approach for handling discrete variables in reactive power planning

The paper proposes an application of an optimization evolutionary technique, improved Differential Evolution, to reactive power planning. This problem could be formulated mathematically as a nonlinear, non smooth, mixed-integer, multi-objective optimization problem. The main objective deals with the minimization of the real power losses that result in a better reactive power dispatch and improving the voltage profile across the system and keeping the control variables in their operational limits. Here we use an approach to deal with discrete variables, Mixed Integer Hybrid Differential Evolution (MIHDE), a modified Differential Evolution algorithm. The proposed approach has been tested in a modified version of two electric networks, the Nordic 32, 60 bus and IEEE 118 bus system.