Distribution Network Reconfiguration based on Opposition Learning Genetic Algorithm

The distribution network reconfiguration is a complex multi-objective optimization problem, where the typical objectives include to minimize the power loss and to ensure the voltage quality and load balancing. A synthetic evaluation index based on Analytic Hierarchy Process (AHP) was used to convert the multiple objectives into single objective. Then an opposition-based learning genetic algorithm (OBLGA) was proposed to solve the distribution network reconfiguration problem. Taking into account the characteristics of “closed-loop design and open-loop operation” of the distribution network, the decimal coding of sequence number of the breaker or switches on the loop was adopted by the OBLGA, which will decrease the chromosome length and reduce the infeasible solutions. Furthermore, the OBLGA could improve the population diversity by opposition learning, i.e., to create the opposite population whose individuals are opposite to the individuals in current population, and to merge the opposition population and current population and select the individuals with better fitness into the next generation. In order to get the opposite point for the distribution network reconfiguration, the sequence number of switch on one loop was projected to the resistance range, whose upper and lower bounds were the sum of resistance of this loop and half of the sum. The test case of 69-bus system showed that the searching efficiency was improved by the OBLGA in comparison with the traditional GA due to the improved population diversity.