Design And Implementation Of Bridge Operation Accident Decision System Based On Case Reasoning

Bridges are one of the most important infrastructures in the country and are closely related to the lives of people.It is of great importance for the government and the public to monitor bridge accidents.Thus,enhancing the understanding of bridge accidents and studying how to deal with them timely and effectively during the accidents are vitally important.Case-based reasoning is a decision-making technique that uses existing experience to solve new problems and can be used to make rapid judgments and deal with bridge operation accidents.However,traditional case reasoning methods are inefficient and complex due to the single and unintuitive case representation.To address the above problems,this thesis adopts knowledge mapping technology to build a knowledge map of bridge operation accident cases and constructs a multi-loop case retrieval method based on different nodes,based on which a bridge operation assisted decision-making system is built to realize bridge operation accident case assisted decision-making.Our contributions include:(1)We construct a knowledge graph for the cases of bridge operation accidents by taking into account the unique features of bridge operation accidents.Firstly,the schema layer of the knowledge graph is designed for bridge operation accident problems.By collecting information from more than 300 accident cases and conducting statistics and analysis according to 18 key information elements,the classes and attributes of accident cases during bridge operation,as well as the types of relationships between entities,are defined.Secondly,we populate the knowledge graph by bridge operation accident cases in a top-down manner.(2)We construct a bridge operation accident case retrieval model based on multiple circles.The model consists of three parts: circle and weight division,similarity calculation,and algorithm design and implementation.First,according to the degree of influence of the nodes on the event,the nodes are divided into core layer and general layer,and the analytic hierarchy process is used to determine the weights of each layer node.Second,we apply Euclidean distance and Hamming distance formulas,the similarity of the cases is calculated from the core layer and general layer nodes respectively,and the total similarity of the cases is obtained according to the node weights.Finally,we transform the similarity calculation method into an executable algorithm and implement it through a program.To verify our multi-layered retrieval method,we conduct experiments by randomly selecting three groups cases from the case library.The results show that the retrieval model based on multiple circles performs excellently in reducing the total number of nodes required for matching and calculation,simplifying the calculation process,and improving the retrieval efficiency.(3)We design and develop a bridge operation accident decision system based on the proposed method.First,we analyze the system requirements and determine the overall architecture of the system.Then,we develop various functional modules of the system based on the Flask+Vue front-end and back-end separation technology.The major functions of the system include user login,case visualization display,similar case retrieval decision,case management and case reuse.Specifically,the similar case retrieval decision function can provide multiple results to assist decision-makers in making decisions.Finally,we conduct functional and performance and tests to validate the developed system.The test results show that our system can be applied to assist effective and efficient bridge operation accident decision-making.