A distribution network reconfiguration algorithm based on Multi-agent system and minimum spanning tree led by particle swarm optimization

作者：Li, Chuan-Jian^*; Liu, Qian-Jin

来源：Power System Protection and Control, 2011, 39(6): 24-28+33.

• Convergence characteristics
• Distribution network reconfiguration
• Global optimal solutions
• Information transmission
• Learning mechanism
• Minimum spanning trees
• Power System
• Survival-of-the-Fittest

A novel distribution network reconfiguration algorithm is proposed, which is based on minimum spanning tree algorithm led by particle swarm optimization proposed on the basis of minimum spanning tree, and combines with collaborative learning between particles, self-learning of excellent particles, survival of the fittest and the features of multi-agent system. The proposed method enhances information transmission in the process, and effectively solves the problems of the blindness of the minimum spanning tree, uncertainty of the value of edge, bionic significance ambiguity of improved particle swarm optimization and so on. In the Multi-agent system, the approach improves the interaction between particles and finds the global optimal solution quickly by competitive and cooperative operation. The convergence speed is further improved through collaborative learning and self-learning. The proposed method is evaluated on a typical example of PG&E 69 nodes distribution system. The result shows that the method has good computation efficiency, good convergence characteristics and high-quality solutions, etc.

• 出版日期2011-3-16
• 单位华南理工大学