A discrete Teaching–Learning-Based Optimization algorithm to solve distribution system reconfiguration in presence of distributed generation

• A discrete Teaching–Learning-Based Optimization (DTLBO) algorithm is presented.
• Compared to other methods, DTLBO only requires population size and stopping criterion.
• DTLBO converge to global optimum and is faster than the other methods.
• Distribution system reconfiguration (DSR) problem is solved using DTLBO.
• Proposed method is comprehensive to be applied in smart grid and practical applications.

Reconfiguration is an important way to increase the power distribution systems efficiency. The problem of reconfiguration is a complicated combinatorial optimization problem with discrete decision variables. To solve such problem, a powerful optimization technique is required. This paper presents a discrete Teaching–Learning-Based Optimization (DTLBO) algorithm for solving the distribution system reconfiguration (DSR) problem. The objectives are power loss minimization and voltage profile improvement in presence of distributed generation (DG). The proposed method is applied to 33-bus and 69-bus test systems and a part of the distribution network of Ahvaz, a city in the south of Iran. A comparison between the proposed algorithm and other existing methods shows the effectiveness and capability of the proposed method to reach the global optimum and rapid convergence to the optimal solution.