Distribution system reconfiguration using genetic algorithm based on connected graphs

• Original genetic algorithm based on connected graphs and comparative tests.
• Implementation of the selection operator by using the ecological niche method in order to detect a variety of optimum points.
• Implementation of the crossover operator where the number of cut points is equal with cyclomatic number − 1.
• Implementation of the mutation operator which satisfies the first condition, for a configuration, in order to be a tree.
• Implementation of the inversion operator which makes branch-exchanges (for not connected configurations).

The reconfiguration of a power distribution system, by exchanging the functional links between its elements, represents one of the most important procedures to improve the performance in the operation of a distribution system. The optimization through reconfiguration (or optimal reconfiguration) of a power distribution system is not a new problem but still represents a difficult one and is specified in some strategies for smart grids. The paper addresses this problem as a single objective one, where, as main criterion, active power losses have been chosen. There is no unique acceptance regarding which approach (based on heuristics, meta-heuristics, genetic algorithms, etc.) is the most proper to use in order to solve this problem. The most important thing is how the specific knowledge of the problem domain is used and modeled in the implementation. Moreover, the indices which can quantify the quality of a reconfiguration method consist in runtime and the obtained solution and these indices can be confirmed only by experimental results. In order to solve the problem (to search and find the optimal solution) an original genetic algorithm based on connected graphs is proposed. Comparative tests performed on some test systems demonstrated the accuracy and the promptness of the proposed algorithm.