New distribution network planning algorithm based on multilayer Hopfield neural network
Gao Weixin, Luo Xian-jue · Journal of systems engineering · 2006
This paper analysis the distribution network planning model,and introduces Hopfield neural network to plan distribution network.We present multiplayer Hopfield neural network model for planning a radical distribution network.The energy function and the regular of choosing parameters are also given in this paper.Each layer of neural network decide a power line for one node,the energy function can optimize all power lines for all nodes simultaneously.A new algorithm based on the multilayer Hopfield neural network is put forward.By minimizing the value of the energy function,the algorithm can obtain the optimal or nearly optimal solution for the topology of radical distribution network.The new algorithm doesn't need coding of city streets and normalizing of data,so the program is easier to be realized.An example shows that the algorithm is practical.