Research On The Method Of Selecting Substation Location Based On The Win-lost Map And MCTS

Selection of substation location is one of the most essential tasks in power grid planning.Substations which act as a junction between upper and lower side of transmissions line take responsibilities of transmission cost evaluation,system efficiency and reliability.Also,it is tightly connected to power grid network.Meanwhile,Selection of substation location is multifactor optimization problem,which means all the factors need to be considered while processing this project.Under these circumstances,the thesis will sort out all the impact factors,quantize those elements and weight them for obtaining the most reasonable selection.Based on the Win-lost map,the Monte Carlo Tree Search algorithm is used for selecting substation location to prove the applicability of itself.The innovation of this paper lies in the establishment of Win-lost map and the application of Monte Carlo Tree Search.In the existing algorithm,the influence of environmental factors in selection of substation location is often fuzzy calculation,which causes the result is not scientific.The advantage of Win-lost map is that it quantifies the influence of various environmental factors,makes the results more scientific,and provides a basis for using Monte Carlo Tree Search.Monte Carlo Tree Search,which is derived from the intelligent go algorithm,provides a new idea for the selection of substation location.To train the algorithm,the AI go algorithm such as alpha Go algorithm combines the method of artificial neural network and Monte Carlo Tree Search,so as to finish intelligent chess.If a large number of substation location examples can be input into artificial neural network for training,neural network plus Monte Carlo Tree Search may also achieve the effect of artificial intelligent location.The research on the application of Monte Carlo Tree Search in selection of substation location take a small step about this topic.