基于图论的改进遗传算法在配网重构中的应用

Application of the improved genetic algorithm based on graph theory in distribution network reconfiguration

针对遗传算法求解配电网络重构优化时,随机编码操作将产生大量不可行解的缺点,设计了基于图论的改进遗传算法进行重构优化。根据配电网络与图论中树结构的相似性,确定了配电网络重构优化实质上是在初步连接图的基础上寻找最优的生成树或某组树,并引入环路的概念,基于破圈法和环路的性质进行编码操作,在产生初始解和变异操作时生成的解均为可行解,交叉操作时满足树的基本条件,避免或大大减少了不可行解的产生,从而提高了算法的计算效率。分别对IEEE33和IEEE69节点配电系统进行了重构优化,两个优化算例说明了该方法的有效性。

When genetic algorithm is used to optimize the distribution network reconfiguration,if coding is operated randomly,large quantities of infeasible solutions will be generated.Aiming at the disadvantage,improved genetic algorithm based on graph theory is designed to optimize network reconfiguration.Based on the similarity between distribution network and tree structure in the graph theory,t he essential of distribution network reconfiguration optimization is determined,that is to search for optimal spanning tree or tree team based on preliminary connection diagram,i ntroduce the concept of loop,c ode based on circle breaking and loop,m ake the solutions feasible in the operations of initial solutions creating and mutation,s atisfy the essential condition of tree crossover in the operation,t hus the number of infeasible solutions is reduced,a nd the computational efficiency of the method is enhanced.IEEE33 and IEEE69 bus distribution systems are reconstructed and optimized respectively,and the two examples show that the algorithm is efficient.