A Road Network Optimization Method Of The "Green Heart Region" In The Changzhutan City Group Under Spatial Interaction Modeling Of Urban Agglomeration Tourist Attractions

In the current society,privatecar travel mode has a considerable share in both intraand inter-city,and has an important impact on intra-and inter-city traffic.There are specific patterns and regularities in the travel interactions of private car groups within regions,which can characterize the travel demand and behavior patterns of residents to a considerable extent.Inter-regional private car trajectory analysis can be applied to the analysis of residents’travel behavior and serve the construction of urban infrastructure.Chang-Zhu-Tan city cluster is an important city cluster located in the central region of China,and the ecological "green heart area" of Chang-Zhu-Tan is located at the location where the three cities meet,mainly including the suburban areas of the three cities.The purpose of planning the ecological "green heart area" is to strengthen the construction of ecological civilization and play its ecological barrier and ecological service function,and the reasonable construction of road network in the area is directly related to whether the "green heart area" can play its function and meet the development goals of perfect facilities and equal services.This study is conducted from the perspective of spatial interaction.This study focuses on the travel and interaction behaviors of private cars between cities in urban clusters and tourist attractions in the region from the perspective of spatial interaction,and study the spatial interaction differences between private cars and tourist attractions in the region,and build a road network optimization method that takes into account the interaction differences of tourist attractions.This paper mainly includes the following research contents.1)A preference model of private cars’ choice of tourist scenic spots based on private cars’ interaction trajectories was proposed.By constructing the movement-stay chain of private car travel,we extract the private car groups that have spatial interaction with scenic spots,and propose four features such as travel time,tour time and expected difference,tour scenic spot level,and expected difference,and tour spending and expected difference from the perspective of private car’s choice demand for tourist scenic spots.The machine learning algorithm is used to construct a private car’s choice preference model for tourist scenic spots.Using the random forest algorithm,the classification accuracy of the choice preference classification model reaches up to 93.23%when the number of classifiers is 100,and the number of minimum split samples is 10.2)A interaction variability index of tourist attractions based on the Gini coefficient and interaction intensity was defined,and we find that the overall interaction variability of tourist attractions in the "green heart area" of Chang-Zhu-Tan city cluster is large and unevenly distributed.Based on the definition of accessibility,the number of people who demand tourist attractions in the departure area,the density of POI facilities in tourist attractions,and the average time of private cars arriving at a tourist attraction are used to define the travel demand,the attractiveness of tourist attractions and the construction level of transportation facilities in the interaction intensity,and the Gini coefficient is introduced to define the interaction variability index of tourist attractions.The empirical analysis finds that 90.35% of the scenic spots in the "green heart area" have interaction variability values greater than 0.4,with a mean value of 0.8065.The overall variability is large,and fairness is low.3)A optimization method of road network structure considering the interaction variability of tourist attractions was proposed.The interactive variability factor of tourist scenic spots is introduced,which constitutes the objective function of optimization together with the road construction cost factor,and propose constraints from two aspects: the road network form structure and the road network scale.And the solution process is designed by applying the idea of genetic algorithm,and the constraints are reflected in the design of penalty function and the initial solution of optimization solution.The experimental results converge in the 56 th generation,and the objective value decreases from 0.3707 in the initial solution to 0.3278 in the converged solution,which is a decrease of 11.6%.Meanwhile,the mean value of interaction variability of tourist attractions before the optimization is 0.8065,and after optimization is 0.7293,which is a decrease of 9.6%,indicating that the optimized road network solution solved can effectively improve the traffic equity of tourist attractions in the region.