Large event activity traffic distribution prediction method based on interestingnessTechnical Field
The invention relates to a large-scale event activity traffic distribution prediction method based on interestingness.
Background
In recent years, with the rapid development of social economy and the vigorous development of national sports culture industry, the frequency of hosting large-scale events in each large city is also higher and higher, and the national and social communities are also paying more attention to various sports events. The number of people going out during the large-scale event is huge, and the traffic service requirement is high, so that the local and peripheral road networks bear huge traffic pressure, and traffic guarantee works face huge challenges. In order to reasonably cope with challenges brought by large-scale event activities and ensure smooth progress of the activities, predicting traffic volume of the large-scale event activities in advance is one of the basis and keys for formulating effective traffic organization and management and control schemes.
In the prediction of traffic volume, the prediction of traffic distribution is a key ring, and the traditional gravity model only considers the attraction strength of traffic cells and the resistance between the traffic cells, so that the travel distribution of two traffic cells is considered to be in direct proportion to the traffic occurrence quantity and attraction quantity of the two cells and in inverse proportion to the resistance between the traffic cells. In the prior art, the traffic distribution prediction method combining the gravity model and the Fratar model not only fully utilizes the current situation distribution information, but also considers the influence of road network change and land utilization on the travel of people, and improves the accuracy and the applicability of the prediction result. The traffic distribution and traffic flow distribution combination model of the destination preference of the traveler is considered, and the important influence of the destination preference of the traveler on the system balance is compared and verified. The two methods make up the defects of the traditional gravity model.
Because the traditional gravity model has weaker pertinence to traffic prediction of large-scale events, the model prediction error is possibly larger, even misleading is caused to the establishment of related traffic organizations and management and control schemes, and the adaptability of the two patents for improving the gravity model and the large-scale event events is lower in matching, therefore, the invention provides a traffic distribution prediction method of the large-scale event events based on interestingness.
The method is characterized in that the travel distribution among cells is changed into the travel distribution among cells and the place where the event is held on the basis of a traditional gravity model, the attraction parameters of the event are set according to the scale, the type and the like of the large-scale event, the travel parameters are set according to the interestingness of the residents of the cells to the event, and when the interestingness difference of the group participation events among the cells is obvious, the interestingness can be used for adjusting the distance value and setting an impedance function to correct the model.
Disclosure of Invention
The invention provides a large-scale event activity traffic distribution prediction method based on interestingness. On the basis of the traditional gravity model, the gravity model is improved aiming at traffic characteristics of large-scale event activities based on interestingness, so that accuracy and pertinence of large-scale event activity traffic distribution prediction are achieved.
The technical scheme adopted by the invention is a large-scale event activity traffic distribution prediction method based on interestingness, which comprises the following steps:
Step 1, collecting current occurrence traffic quantity, current OD distribution and other basic data of each cell for large-scale event activities, wherein the current occurrence quantity of each cell is Oi, the attraction quantity of a large-scale event activity sub-field is Dj, i=1, 2..n, j=1, 2..m, n represents the number of cells for large-scale event activity traffic, and m represents the number of large-scale event activity sub-fields.
Step 2, determining the form of a large event activity traffic distribution model based on interest degree as follows:
The method comprises the steps of setting a model, wherein Xij is a distribution predicted value of a small area i to a large-scale event activity sub-field j, k is a model parameter, Oi is a traffic occurrence quantity of the small area i, alphai is a occurrence quantity parameter of the small area i, setting according to the interestingness of residents in the small area i to attend and watch similar event activities, Dj is an attraction quantity of the event activity sub-field j, the model only aims at the large-scale event activities with a plurality of sub-fields located in different areas, betaj is an attraction quantity parameter of the event activity sub-field j, f (cij) is an impedance function, cij is an impedance function parameter, and taking account of large differences between the event activities and daily traveling behaviors, and besides considering the distance, time and cost between the small area i and the event activity sub-field j, introducing the interestingness of the residents in the small area to the event to represent traveling will of the residents.
The model satisfies constraintsThat is, the occurrence amount of the traffic zone is equal to the attraction amount of the large-scale event.
Step 3, determining the expression of the parameters alphai、βj and f (cij) of the large event activity traffic distribution model based on the interest level
The occurrence quantity parameter alphai of the traffic cell i, namely the attention index of residents in the cell i to the large-scale event activity, is related to the interest degree of the residents in the cell to the event activity, so that the calculation formula is as follows:
Wherein, alphai is the occurrence parameter, gi is the number of sports facilities related to the event in the cell i, and bi is the number of times of searching the entry related to the event in the cell i through the webpage.
The attraction parameter βj of the large-scale event division field j relates to the scale of the event, the number of items of the event and the related range, and when the scale of the event is larger, the number of items is larger and the related range is wider, the value of the attraction parameter is larger to adjust the attraction of the event, and the calculation formula is as follows:
wherein, betaj is the attraction parameter of the event activity sub-field j, nj is the maximum accommodation number of the event sub-field j; cj is the number of countries or regions participating in the event field j, wherein the number of regions participating in the event is taken if the event activity is at a state level, and the number of countries participating in the event is taken if the event activity is at an international event; pj is the number of items of the event division field j; average number of items for the same type of event activity.
Impedance function f (cij) of large event activity traffic distribution model based on interestingness and adopting exponential functionBesides, the impedance function is inversely proportional to the economic level of the community, the population number, the interest degree of the community residents in the event activity and the like, and the calculation formula is as follows, wherein the difference between the travel of the large event activity and the usual travel of residents is considered:
Wherein f (ci) is an impedance function, Rij is the distance from the cell i to the large-scale event activity sub-field j, and Ni is the total number of people in the cell i; gi is the total yield value of the cell i; Si is the commodity sales sum of the cell i; The method comprises the steps of planning an average value of commodity sales total in each cell in a region, wherein gi is the number of sports facilities related to the event in a cell i, and bi is the number of times of searching entries related to the event in the cell i through a webpage.
Step 4, calibrating model parameters k and lambda
And determining a sample data set through the occurrence quantity of the current traffic cell, the attraction predicted value of the large-scale event activity and the OD distribution of the current large-scale event activity, calibrating a model by adopting a least square method, and determining model parameters k and lambda.
The invention has the beneficial effects that:
The traffic distribution prediction method for the large-scale event activities based on the interestingness provided by the invention breaks through the defects that the conventional gravity model has weak pertinence to the large-scale activities, so that the model prediction error is larger, and even misleading is caused to the establishment of related traffic organization and management and control schemes. Factors such as the interest degree of residents on the event activities and the attraction degree of the large event activities are considered, a more accurate prediction model is provided for predicting the traffic volume of the large event activities in the future, the prediction accuracy is improved, the related units can adopt a more effective traffic organization mode, and the guarantee is provided for smooth holding of the large event activities.
Detailed Description
The following description is made of specific steps of the present invention.
Step 1, collecting current occurrence traffic quantity, current OD distribution and other basic data of each cell for large-scale event activities, wherein the current occurrence quantity of each cell is Oi, the attraction quantity of a large-scale event activity sub-field is Dj, i=1, 2..n, j=1, 2..m, n represents the number of cells for large-scale event activity traffic, and m represents the number of large-scale event activity sub-fields.
Step 2, determining the form of a large event activity traffic distribution model based on interest degree as follows:
The method comprises the steps of setting a model, wherein Xij is a distribution predicted value of a small area i to a large-scale event activity sub-field j, k is a model parameter, Oi is a traffic occurrence quantity of the small area i, alphai is a occurrence quantity parameter of the small area i, setting according to the interestingness of residents in the small area i to attend and watch similar event activities, Dj is an attraction quantity of the event activity sub-field j, the model only aims at the large-scale event activities with a plurality of sub-fields located in different areas, betaj is an attraction quantity parameter of the event activity sub-field j, f (cij) is an impedance function, cij is an impedance function parameter, and taking account of large differences between the event activities and daily traveling behaviors, and besides considering the distance, time and cost between the small area i and the event activity sub-field j, introducing the interestingness of the residents in the small area to the event to represent traveling will of the residents.
The model satisfies constraintsThat is, the occurrence amount of the traffic zone is equal to the attraction amount of the large-scale event.
Step 3, determining the expression of the parameters alphai、βj and f (cij) of the large event activity traffic distribution model based on the interest level
The occurrence quantity parameter alphai of the traffic cell i, namely the attention index of residents in the cell i to the large-scale event activity, is related to the interest degree of the residents in the cell in the event activity, and when the interest degree of the residents in the cell in the event activity is higher, the venue facilities related to the event activity are more, so that the calculation formula is as follows:
Wherein, alphai is the occurrence parameter, gi is the number of sports facilities related to the event in the cell i, and bi is the number of times of searching the entry related to the event in the cell i through the webpage.
The attraction parameter βj of the large-scale event division field j relates to the scale of the event, the number of items of the event and the related range, and when the scale of the event is larger, the number of items is larger and the related range is wider, the value of the attraction parameter is larger to adjust the attraction of the event, and the calculation formula is as follows:
wherein, betaj is the attraction parameter of the event activity sub-field j, nj is the maximum accommodation number of the event sub-field j; cj is the number of countries or regions participating in the event field j, wherein the number of regions participating in the event is taken if the event activity is at a state level, and the number of countries participating in the event is taken if the event activity is at an international event; pj is the number of items of the event division field j; average number of items for the same type of event activity.
Impedance function f (cij) of large event activity traffic distribution model based on interestingness and adopting exponential functionBesides, the impedance function is inversely proportional to the economic level of the community, the population number, the interest degree of the community residents in the event activity and the like, and the calculation formula is as follows, wherein the difference between the travel of the large event activity and the usual travel of residents is considered:
Wherein f (ci) is an impedance function, Rij is the distance from the cell i to the large-scale event activity sub-field j, and Ni is the total number of people in the cell i; gi is the total yield value of the cell i; Si is the commodity sales sum of the cell i; The method comprises the steps of planning an average value of commodity sales total in each cell in a region, wherein gi is the number of sports facilities related to the event in a cell i, and bi is the number of times of searching entries related to the event in the cell i through a webpage.
Step 4, calibrating model parameters k and lambda
And determining a sample data set through the occurrence quantity of the current traffic cell, the attraction predicted value of the large-scale event activity and the OD distribution of the current large-scale event activity, calibrating a model by adopting a least square method, and determining model parameters k and lambda.