Intelligent Optimization Of Reservoir Flood Control Dispatch Decision-Making Scheme Model Studies

China’s precipitation time and space distribution is uneven,flood disasters are frequent,reservoir is all flood control projects on the flood control ability and regulation effect of the strongest flood control project,the use of reservoir flood dispatch,can greatly reduce flood disasters.The use of appropriate scheduling decision-making methods can greatly improve the efficiency of flood control decision-making and give more reasonable decision-making recommendation schemes,which is of great research significance.The knowledge graph is used to store,mine and display data such as complex water conditions and other data used in the scheduling decision-making process,as well as data calculation results added in the middle of the process,which can improve the scheduling and decision-making rate faster.This paper takes Shanxi Reservoir as an example to construct the reservoir knowledge map,and on this basis,the optimal scheduling and intelligent decision-making of the reservoir are carried out,and the scheduling and decision-making results are stored and displayed in the knowledge graph to provide reference for future scheduling decisions,and the main research contents and results are as follows:(1)Build a knowledge graph of Shanxi ReservoirBased on the relevant information obtained from the reservoir,the pattern layer is established,and the data knowledge is extracted on this basis,and the NEO4 J database is used for storage to provide data support for subsequent flood control optimization scheduling and decision-making.(2)The optimal scheduling model of reservoir flood control was constructedIn order to better utilize Shanxi Reservoir for flood control,four typical floods were selected,with the smallest flood control capacity of the largest reservoir occupied and the minimum maximum overflow of important flood control sections downstream,constrained by the relationship between the lowest water level,water balance,water level reservoir capacity,etc.,and punished by the amplitude of the outflow flow and the water level at the end of the reservoir,and the genetic algorithm was used to solve the model,and the weight of the target was changed in equal steps,and the optimal solution under21 different target weights was generated for each flood to form a scheme set.(3)An intelligent optimization model of reservoir flood control optimal scheduling scheme was constructedOn the basis of the knowledge graph of Shanxi Reservoir,the index evaluation system was extracted.Using analytic hierarchy method,entropy weight method and other methods to calculate the subjective and objective weights corresponding to each index,and then based on the combined weight assignment method of Bo abandonment theory,the respective deviations between the possible weights and each basic weight are minimized,and the combined weights of each index are calculated by combining the empirical weight and mathematical weight,such as the subjective weight of each index of the once-in-50-year design flood is 0.043,0.331,0.169,0.060,0.236,0.161,and the objective weight is 0.233,0.172,0.142,0.170,0.166,0.118,the comprehensive weight is 0.086,0.295,0.163,0.085,0.220,0.152,on this basis,Bayesian,fuzzy evaluation and other methods are used to optimize the scheme of historical floods and design floods,the scheme is graded and stored,multiple floods are selected for testing,similar floods are matched according to flood peaks and floods,and historical scheduling is used as a training set.Decision trees are used to calculate the hierarchy of flood control schemes,and fuzzy optimization and Bayesian models are used to optimize schemes and rank each scheme.The results show that in the event of a flood in the following 50 years,the decision-making results of the three decision-making methods are not much different.For a flood in 50 years,the first-level scheme selected by the fuzzy optimization model is scheme 12,and the first-level scheme selected by the Bayesian model is scheme 13,and the two schemes are not optimal through the decision tree calculation,indicating that there is a better scheme in the historical scheduling,then the parameters are readjusted to generate four new schemes through scheduling calculation,the three schemes are reused for decision-making,and the optimal scheme is selected for scheduling by synthesizing the three model optimization results.