Research On The Prediction Of Ordinary Residential Travel Rate Based On Location And Supporting Facilities Completenes

In recent years,economy has maintained a sustained and rapid development trend in China,and the urbanization process has been accelerating day by day,which has provided a strong impetus for the massive development of urban land.However,the high-intensity land development activities have led to an increasingly prominent problem of the imbalance between supply and demand of urban transportation,the contradiction between the land development and the transportation system has gradually become prominent,Traffic impact analysis of construction project is an effective measure to coordinate this contradiction.Trip rate is a basic parameter for quantitatively analyzing transportation demand in transportation impact analysis work.The selection of trip rate will directly affect the results of transportation impact analysis.As a typical construction project,ordinary residence houses have a direct impact on the operating quality of the transportation system due to the size of their trips.Therefore,research on trip rates for ordinary residence houses is crucial.This requires us to start from the traffic characteristics of residential areas,deeply study the influence mechanism of residents’ trip behavior and willingness to trip,clarify the main factors affecting trip,and construct more accurate trip prediction models.In this way,it can provide support for the conclusion of traffic impact analysis and achieve quantitative traffic evaluation to guide the goals of land development and urban development.Based on a comprehensive review of domestic and foreign research,this paper takes ordinary residence houses in Guangzhou as the research object.Based on data mining,the paper abstracts ordinary residential buildings as point elements from a macroscopic scale,creates Thiessen polygons to initially divide traffic zones,and uses POI data to quantify the location of traffic zones.Using the location potential as a factor,an improved K-means clustering method is used to cluster the initial traffic zones into core areas,sub-core areas,peripheral areas,and edge areas.The results are tested using spatial autocorrelation methods.Based on the location division results,typical residential houses were selected for a survey from four aspects: building characteristics,supporting facilities,residents’ trip characteristics,and trip volume.The survey results were compiled and analyzed from two dimensions: overall and by zone,and differences in residents’ trip purposes and modes between different zones were identified.Preliminary conclusions were drawn on the effects of building characteristics,location,and supporting facilities on trip behavior.Based on the theory of activity space and the tolerance distance of residents’ daily trip to various living service facilities,different types of daily living service facilities are analyzed and classified using the Analytic Hierarchy Process to quantify the impact of supporting facilities on trip and establish a more strongly correlated supporting facility improvement index with trip volume.Based on this,a completeness index of supporting facilities that is more strongly correlated with travel volume is established.On this basis,a multiple regression and neural network travel prediction model considering location and completeness of supporting facilities is respectively constructed.The grey wolf algorithm is used for optimization to build a combination model with higher prediction accuracy,with a fitting goodness of 90.46%,which respectively improved by 2.68% and 0.48%.Finally,the combination model is used to propose recommended values for the morning peak travel rate of residential areas in Guangzhou.