The Research Of Expressway Tow And Rescue Resources Demand Forecasting Based On Case-based Reasoning

With the rapid growth of expressway mileage and the amount of motor vehicles in China,the task of maintaining the safe and efficient operation of expressway has become more and more difficult.As an important guarantee and basis for the implementation of expressway unobstructed traffic flows maintenance work,the tow and rescue resources play an important role in reducing the loss of life and property caused by traffic accidents and improving the efficiency of road network.And for the management work of tow and rescue resources allocating and utilizing,a reasonable prediction of expressway tow and rescue resources demand is a prerequisite for follow-up resource allocation and scheduling work.It is of great theoretical and practical significance to study the demand forecast of expressway tow and rescue resources.By introducing the theoretical framework of the case-based reasoning method and analyzing its advantages in the research field of emergency rescue,the case-based reasoning is introduced into the demand forecast of the expressway tow and rescue resources.By analyzing the common knowledge expression method,a two-tuple representation method of resource prediction case is proposed.Through analyzing the influencing factors of expressway traffic accidents,such as expressway alignment and roadbed pavement,the case characteristic space including the length of the road is constructed.According to the actual use situation of tow and rescue resources and summarize of the equipment statistical data,the representation method of resources demand is also constructed by rescue equipment's different types and abilities.Aiming at deleting the redundant information existing in the construction process of the case feature attribute space,the principle,construction process,key parameters and application advantages of random forest method is elaborated and the reduction rate of out-of-bag classification error is used to measure the importance of the features.The application of the feature selection algorithm is given based on the actual data and the key factor including the road length and traffic volume are selected the feature set of demand forecasting.Through the analysis of the advantages and disadvantages of the traditional case retrieval method,the case retrieval method combined with radial basis function neural network is proposed to improve the efficiency of retrieval process.By introducing the general principle of radial basis function neural network and common learning method,aiming at solving the problem that the hidden layer of the radial basis network is difficult to be determined,the dynamic decay adjustment method is used to improving the radial basis network learning process.Through the example data verification,the improved radial basis function neural network case retrieval can greatly improve the learning speed in the case of guaranteeing the retrieval precision.