Abstract
This paper dealt with the optimal reconfiguration of a power distribution network where a genetic algorithm (GA) based on graph theory was considered. This paper addressed this issue as a multi-criteria one with low investment cost and reduced active power losses objectives. The topological and voltage stability constraints were taken into account in this study, and each objective was firstly optimized individually. The Pareto-optimality principle for the criteria trade-off was then presented. In order to calculate the active power losses, a new load-flow approach was implemented. The genetic algorithm was tested on IEEE 10-bus system. Results show that the proposed load-flow algorithm was an efficient tool to obtain an accurate calculus for the system power losses. Efficiency and capability of the GA to achieve good solutions in a conspicuously short time are also discussed, and all the simulation results explained that the two objectives mentioned above are conflicting with each other.