A Traffic Distribution Prediction Method Based on a Dual-Constraint Radial Model

Abstract

This invention discloses a traffic distribution prediction method based on a dual-constraint radiation model. First, basic data related to urban traffic travel are collected. Second, the generalized travel cost between traffic zones is calculated, and intervention opportunity areas are determined. Then, a traffic distribution prediction model based on the generalized cost radiation model is established, and an initial demand distribution matrix is ​​calculated. Finally, using the initial demand distribution matrix as a priori, iterative proportional fitting is performed, and the fitted travel demand distribution matrix is ​​output as the final prediction result. The advantages are that it is based on the generalized cost and intervention opportunity mechanism, and satisfies the dual constraints of travel occurrence and attraction, thus overcoming the inherent defects of gravity models and radiation models. It possesses a solid theoretical foundation and macroscopic statistical consistency, improving the accuracy of traffic distribution prediction.

Description

Technical Field

[0001] This invention relates to urban traffic demand distribution prediction technology, and in particular to a traffic distribution prediction method based on a dual-constraint radial model. Background Technology

[0002] Urban traffic demand distribution forecasting is one of the core issues in traffic planning and management. Its purpose is to predict the travel volume between traffic zones (i.e., the origin-destination traffic distribution matrix), which is of great significance for urban traffic network optimization, resource allocation, and the formulation and implementation of future traffic policies.

[0003] Traditional traffic distribution prediction methods are mainly based on gravity models, which are widely used due to their simplicity and ease of computation. For example, Chinese patent number 2022101727395 discloses an improved gravity model traffic distribution prediction method, and Chinese patent number 2022102764160 discloses a two-stage traffic distribution prediction method. However, gravity models are essentially empirical formulas, lacking a behavioral theoretical foundation, and their parameter calibration relies on prior data, leading to systematic biases in the prediction results and significant room for improvement in prediction accuracy. In recent years, research has proposed traffic distribution prediction methods based on radiation models. For example, Chinese patent application CN117172377A discloses an intercity traffic distribution prediction method that considers differences in urban attraction intensity. The radiation model derives the travel probability from the perspective of traveler behavior decision-making based on the intervention opportunity mechanism, which to some extent overcomes the lack of behavioral theoretical basis of the gravity model. However, it assumes that travelers choose their destinations according to straight-line distance and does not consider the real road network structure and generalized travel costs (such as time and expenses). More importantly, its prediction results are not guaranteed to be consistent with known traffic generation and attraction, and the prediction accuracy is still low, making it difficult to directly apply to actual planning scenarios that need to meet macro-statistical constraints. Summary of the Invention

[0004] This invention aims to overcome the shortcomings of existing technologies and provide a traffic distribution prediction method based on a dual-constraint radiation model. This method is based on generalized cost and intervention opportunity mechanisms and satisfies the dual constraints of trip occurrence and attraction, thus overcoming the inherent defects of gravity and radiation models. While maintaining the theoretical foundation of behavior, it improves the macroscopic statistical consistency of traffic distribution prediction and has high prediction accuracy.

[0005] The technical solution adopted by this invention to solve the above-mentioned technical problems is as follows: a traffic distribution prediction method based on a dual-constraint radiation model, comprising the following steps: Step 1: Collect basic data related to urban transportation and travel; Step 2: Calculate the generalized travel cost between traffic zones and determine the opportunity area for intervention; Step 3: Establish a traffic distribution prediction model based on the generalized cost radiation model and calculate the initial demand distribution matrix; Step 4: Using the initial demand distribution matrix as a priori, perform iterative proportional fitting; Step 5: Output the fitted travel demand distribution matrix as the final prediction result.

[0006] Compared with existing technologies, the advantages of this invention are as follows: This invention, through an intervention opportunity mechanism, establishes travel distribution prediction on the theoretical basis of individual behavioral decision-making, overcoming the shortcomings of traditional gravity model empirical formulas; simultaneously, it correlates intervention opportunities with urban road network structure through generalized travel costs, compensating for the deficiencies of radiation models; finally, through iterative proportional fitting of occurrence and attraction quantities as dual constraints, it ensures the consistency between the prediction results and macro-statistical data. Therefore, this invention combines a behavioral theoretical foundation with macro-statistical consistency, and the prediction results more closely match the actual distribution patterns of urban traffic travel, exhibiting higher prediction accuracy and reliability, and providing a scientific reference for urban traffic planning, road network evaluation, and traffic policy formulation.

[0007] In a further technical solution, the basic urban transportation data collected in step 1 includes: urban road network data, traffic zone characteristic data, and transportation demand data, as detailed below: Step 1.1: Collect urban road network data, including road network node (i.e., intersection) numbers, road segment numbers and start and end points, as well as the length, design speed and monetary cost (including fuel cost and toll cost) of each road segment, and the delay time of each node, and record it. The set of all nodes in the road network. This is the set of all road segments in the road network. For any road segment... Let its length be . (Unit: km), Design speed is (Unit: km / h), Currency: (Unit: Yuan); for any node Record the delay time as (Unit: hours); Step 1.2: Collect traffic zone characteristic data, including traffic zone number, traffic zone centroid number, and population within the traffic zone, and record it. This is the set of all traffic zones. For any traffic zone... Let its centroid be . The population is ; Step 1.3: Collect traffic demand data, including future trip occurrences and attraction volumes for each traffic zone. For any traffic zone... Record the number of trips taken. Travel attraction volume is ; In a further technical solution, step 2 involves calculating the generalized travel cost between traffic zones and determining the intervention opportunity area, as detailed below: Step 2.1: For any road segment Calculate its basic generalized travel cost using the following formula. : ; in, For any road segment Travel time (unit: hours); The time value coefficient for a city (unit: yuan / hour) is determined based on the city's average wage level. For any road segment The basic generalized cost of travel; Step 2.2: For any road segment The endpoint is determined by the following formula. The delay time, after being converted into monetary costs, is included in the generalized travel cost of that route segment: ; in, For any road segment that includes the destination delay after adjustment The generalized cost of travel; Step 2.3: For any traffic zone pair Based on the adjusted generalized travel costs of each road segment, the minimum generalized travel cost is calculated using the shortest path algorithm: ; in, As a traffic zone from any starting point As any destination traffic zone and They are respectively the starting point traffic community centroid ( (a certain node in the middle) and the terminal traffic area centroid ( (a certain node in the middle). For the node To the node The set of all feasible paths, For any one of these paths, it is composed of several consecutive road segments; Indicates from traffic community To the traffic community The minimum generalized travel cost between; Step 2.4: For any traffic zone pair It will be from the traffic community Departure, minimum generalized travel cost less than The set of all other traffic zones, defined as the set of all traffic zones that the traffic zone is paired with. Intervention opportunity area : ; in, To get from the traffic community To the traffic community The minimum generalized travel cost between; Indicates satisfaction Any other traffic zone.

[0008] In a further technical solution, step 3 involves establishing a traffic distribution prediction model based on a generalized cost radiation model and calculating an initial demand distribution matrix, as detailed below: Step 3.1: For any traffic zone Define its destination opportunity attractiveness If no restrictions are placed on the type of traffic distribution, then... For traffic community population ,like and The dimensions are inconsistent, so a conversion factor is used to unify them to population equivalents; if the transportation distribution is limited to the distribution of commuting travel demand... For traffic community In this case, when collecting traffic zone feature data, it is also necessary to collect the number of jobs in the traffic zone; if the traffic distribution is limited to the distribution of tourism travel demand, For traffic community The number of school places in the traffic area is also required when collecting traffic area feature data. Step 3.2: For any traffic zone pair From the traffic community To the traffic community travel probability Calculated by the following formula: ; in, For traffic community Population size; To get from the traffic community To the traffic community The total population of the intervention opportunity area, i.e. ; Indicates traffic community Intervention opportunity area The population size of any traffic zone K in the middle; Step 3.3: Calculate from traffic zone To the traffic community initial travel volume and the initial demand distribution matrix : ; ; in, Indicates traffic community The number of trips, Indicates from traffic community To any traffic community The minimum generalized travel cost between them Represents a set The model.

[0009] In a further technical solution, in step 4, using the initial demand distribution matrix as a priori, an iterative proportional fitting is performed, as follows: Step 4.1: Set iteration parameters, including the maximum number of iterations. and convergence threshold Maximum number of iterations and convergence threshold All settings are based on the required prediction accuracy. The smaller the value, the higher the precision requirement; initialize the iteration counter. Current demand distribution matrix ; Step 4.2: Execute the... The next iteration is as follows: Step 4.2.1: Perform iterative scaling and row adjustment: ; in, Indicates from traffic community To the traffic community Initial travel volume; Step 4.2.2: Perform iterative scaling and column adjustment: ; in, Indicates traffic community The number of trips attracted.

[0010] Step 4.3: Convergence Judgment: Calculate the relative error between the current demand distribution matrix and the target constraint: ; ; like or If the iteration converges, then the iteration stops and proceeds to step 5; otherwise, the iteration has not converged, and we set... Then return to step 4.2 for the next iteration.

[0011] In a further technical solution, in step 5, after iterative convergence, the current demand distribution matrix is ​​output. This serves as the final prediction result for the distribution of urban traffic demand. Attached Figure Description

[0012] Figure 1 This is a flowchart of the traffic distribution prediction method of the present invention; Figure 2 This is a diagram of the urban road network structure according to an embodiment of the present invention. Detailed Implementation

[0013] To facilitate understanding and implementation of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0014] It should be noted that the term "traffic zone" as used in this invention is a general professional term and is well known to those skilled in the art. In urban traffic planning, the urban traffic network is divided into several functional sub-zones according to certain principles, and these functional sub-zones are called traffic zones.

[0015] Example: Figure 1 As shown, a traffic distribution prediction method based on a dual-constraint radial model includes the following steps: Step 1: Collect basic data related to urban transportation and travel, specifically: Step 1.1: Collect urban road network data, including road network node (i.e., intersection) numbers, road segment numbers and start and end points, as well as the length, design speed and monetary cost (including fuel cost and toll cost) of each road segment, and the delay time of each node, and record it. The set of all nodes in the road network. This is the set of all road segments in the road network. For any road segment... Let its length be . (Unit: km), Design speed is (Unit: km / h), Currency: (Unit: Yuan); for any node Record the delay time as (Unit: hours); Step 1.2: Collect traffic zone characteristic data, including traffic zone number, traffic zone centroid number, population within the traffic zone, and number of jobs within the traffic zone, and record them. This is the set of all traffic zones. For any traffic zone... Let its centroid be . The population is ; Step 1.3: Collect traffic demand data, including future trip occurrences and attraction volumes for each traffic zone. For any traffic zone... Record the number of trips taken. Travel attraction volume is ; Step 2: Calculate the generalized travel cost between traffic zones and determine the opportunity area for intervention, specifically: Step 2.1: For any road segment Calculate its basic generalized travel cost using the following formula. : ; in, For any road segment Travel time (unit: hours); The time value coefficient for a city (unit: yuan / hour) is determined based on the city's average wage level. For any road segment The basic generalized cost of travel; Step 2.2: For any road segment The endpoint is determined by the following formula. The delay time, after being converted into monetary costs, is included in the generalized travel cost of that route segment: ; in, For any road segment that includes the destination delay after adjustment The generalized cost of travel; Step 2.3: For any traffic zone pair Based on the adjusted generalized travel costs of each road segment, the minimum generalized travel cost is calculated using the shortest path algorithm: ; in, As a traffic zone from any starting point As any destination traffic zone and They are respectively the starting point traffic community centroid ( (a certain node in the middle) and the terminal traffic area centroid ( (a certain node in the middle). For the node To the node The set of all feasible paths, For any one of these paths, it is composed of several consecutive road segments; Indicates from traffic community To the traffic community The minimum generalized travel cost between; Step 2.4: For any traffic zone pair It will be from the traffic community Departure, minimum generalized travel cost less than The set of all other traffic zones, defined as the set of all traffic zones that the traffic zone is paired with. Intervention opportunity area : ; in, To get from the traffic community To the traffic community The minimum generalized travel cost between; Indicates satisfaction Any other traffic zone.

[0016] In this embodiment, the urban road network includes 4 nodes and 5 road segments. The attributes of the nodes and road segments are shown in Tables 1 and 2.

[0017] Table 1 Node Attributes

[0018] Table 2 Road Segment Attributes

[0019] The city's road network is divided into three traffic zones. The centroid, population size, and traffic generation and attraction of each zone are shown in Table 3.

[0020] Table 3 Traffic Community Attributes

[0021] set up Yuan / hour, for road segment (1,2): Hour, Yuan, The minimum generalized travel cost between traffic zones is calculated sequentially for all road segments, as shown in Table 4, following step 2.

[0022] Table 4 Minimum Generalized Travel Cost

[0023] Taking A→C as an example, Other communities with a cost less than 11.98 are Therefore, the intervention area , The same calculations were performed for the other traffic zones, and the intervention areas were determined as shown in Table 5.

[0024] Table 5. Intervention Area and Population

[0025] Step 3: Establish a traffic distribution prediction model based on the generalized cost radiation model and calculate the initial demand distribution matrix, specifically: Step 3.1: For any traffic zone Define its destination opportunity attractiveness The current traffic distribution is limited to the distribution of commuting travel demand. For traffic community The number of job openings is as follows: ; Step 3.2: For any traffic zone pair From the traffic community To the traffic community travel probability Calculated by the following formula: ; in, For traffic community Population size; To get from the traffic community To the traffic community The total population of the intervention opportunity area, i.e. ; Indicates traffic community Intervention opportunity area The population size of any traffic zone K in the middle; Step 3.3: Calculate from traffic zone To the traffic community initial travel volume and the initial demand distribution matrix : ; ; in, Indicates traffic community The number of trips, Indicates from traffic community To any traffic community The minimum generalized travel cost between them Represents a set The model.

[0026] Taking traffic community A as an example, the probability of travel is calculated as follows: ; ; ;

[0027] The remaining origin and destination points are calculated similarly, and the normalized selection probabilities and initial demand distribution matrices are obtained as follows: ; Step 4: Using the initial demand distribution matrix as a priori, perform iterative proportional fitting, specifically as follows: Step 4.1: Set iteration parameters, including the maximum number of iterations. and convergence threshold Initialize the iteration counter Current demand distribution matrix ; Step 4.2: Execute the... The next iteration is as follows: Step 4.2.1: Perform iterative scaling and row adjustment: ; in, Indicates from traffic community To the traffic community Initial travel volume; Step 4.2.2: Perform iterative scaling and column adjustment: ; in, Indicates traffic community The number of trips attracted.

[0028] Step 4.3: Convergence Judgment: Calculate the relative error between the current demand distribution matrix and the target constraint: ; ; like or If the iteration converges, then the iteration stops and proceeds to step 5; otherwise, the iteration has not converged, and we set... Then return to step 4.2 for the next iteration.

[0029] In this embodiment, during the first iteration, the distribution matrix is ​​adjusted by fitting the iterative ratio. The current column sum is as follows: For traffic zone A: For traffic area B: For traffic area C: ;Target column and (i.e., attraction): , , Adjustment coefficients: For traffic zone A: 3200 / 2979 = 1.075, for traffic zone B: 4200 / 6013 = 0.698, for traffic zone C: 4600 / 3011 = 1.527. This yields the matrix after the first iteration. ; If the current row does not meet the occurrence constraint, proceed with step 4.2 iteratively until, after 8 iterations, the matrix converges to the following result: ; Step 5: Output the current demand distribution matrix This serves as the final prediction result for the distribution of urban traffic demand.

[0030] It should be understood that the above embodiments are only for illustrating the technical ideas of the present invention. For those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A traffic distribution prediction method based on a dual-constraint radial model, characterized in that, Includes the following steps: Step 1: Collect basic data related to urban transportation and travel; Step 2: Calculate the generalized travel cost between traffic zones and determine the opportunity area for intervention; Step 3: Establish a traffic distribution prediction model based on the generalized cost radiation model and calculate the initial demand distribution matrix; Step 4: Using the initial demand distribution matrix as a priori, perform iterative proportional fitting; Step 5: Output the fitted travel demand distribution matrix as the final prediction result.

2. The traffic distribution prediction method based on a dual-constraint radial model according to claim 1, characterized in that, The basic urban transportation data collected in step 1 includes: urban road network data, traffic zone characteristic data, and traffic demand data. The urban road network data includes road network node numbers, road segment numbers, road segment start and end points, as well as the length, design speed, and monetary cost of each road segment. The traffic zone characteristic data includes traffic zone numbers, traffic zone centroid numbers, and the population within each traffic zone. The traffic demand data includes the future travel occurrence and travel attraction of each traffic zone.

3. The traffic distribution prediction method based on a dual-constraint radial model according to claim 2, characterized in that, In step 2, the generalized travel cost between traffic zones is calculated and the intervention opportunity area is determined, as follows: Step 2.1: For any road segment Calculate its basic generalized travel cost using the following formula. : ； in, It is the set of all road segments in the urban road network. The set of all nodes in the urban road network. Represent the start and end points of any road segment, respectively; any road segment Its length is Design speed is The monetary fee is ; For any road segment Travel time; The time value coefficient of the city; For any road segment The basic generalized cost of travel; Step 2.2: For any road segment The endpoint is determined by the following formula. The delay time, after being converted into monetary costs, is included in the generalized travel cost of that route segment: ； in, Indicates the end point The delay time; For any road segment that includes the destination delay after adjustment The generalized cost of travel; Step 2.3: For any traffic zone pair Based on the adjusted generalized travel costs of each road segment, the minimum generalized travel cost is calculated using the shortest path algorithm: ； in, Indicates a traffic zone with any starting point. Indicates any destination traffic zone. This is a collection of all traffic zones. and They are respectively the starting point traffic community Centroid and endpoint traffic area The centroid point; For the node To the node The set of all feasible paths, For any one of these paths, it is composed of several consecutive road segments; Indicates from traffic community To the traffic community The minimum generalized travel cost between; Step 2.4: For the given traffic zone pair It will be from the traffic community Departure, minimum generalized travel cost less than The set of all other traffic zones, defined as the set of all traffic zones that the traffic zone is paired with. Intervention opportunity area : ； in, To get from the traffic community To the traffic community The minimum generalized travel cost between; Indicates satisfaction Any other traffic zone.

4. The traffic distribution prediction method based on a dual-constraint radial model according to claim 3, characterized in that, In step 3, a traffic distribution prediction model based on the generalized cost radiation model is established and the initial demand distribution matrix is ​​calculated, as follows: Step 3.1: For any traffic zone Define its destination opportunity attractiveness If no restrictions are placed on the type of traffic distribution, then... For traffic community population ,like and The dimensions are inconsistent, so a conversion factor is used to unify them to population equivalents; if the transportation distribution is limited to the distribution of commuting travel demand... For traffic community The number of job openings; if the transportation distribution is limited to the distribution of tourism travel demand, For traffic community The number of school places; Step 3.2: For any traffic zone pair From the traffic community To the traffic community travel probability Calculated by the following formula: ； in, For traffic community Population size; To get from the traffic community To the traffic community The total population of the intervention opportunity area, i.e. ; Indicates traffic community Intervention opportunity area The population size of any traffic zone K in the middle; Step 3.3: Calculate from traffic zone To the traffic community initial travel volume and the initial demand distribution matrix : ； ； in, Indicates traffic community The number of trips, Indicates from traffic community To any traffic community The minimum generalized travel cost between them Represents a set The model.

5. The traffic distribution prediction method based on a dual-constraint radial model according to claim 4, characterized in that, In step 4, using the initial demand distribution matrix as a priori, iterative proportional fitting is performed, as follows: Step 4.1: Set iteration parameters, including the maximum number of iterations. Convergence threshold greater than 0 Initialize the iteration counter Current demand distribution matrix ; Step 4.2: Execute the... The next iteration is as follows: Step 4.2.1: Perform iterative scaling and row adjustment: ； in, Indicates from traffic community To the traffic community Initial travel volume; Step 4.2.2: Perform iterative scaling and column adjustment: ； in, Indicates traffic community The number of trips attracted; Step 4.3: Convergence Judgment: Calculate the relative error between the current demand distribution matrix and the target constraint: ； ； like or If the iteration converges, then the iteration stops and proceeds to step 5; otherwise, the iteration has not converged, and we set... Then return to step 4.2 for the next iteration.

6. The traffic distribution prediction method based on a dual-constraint radial model according to claim 5, characterized in that, In step 5, after the iteration converges, the current demand distribution matrix is ​​output. This serves as the final prediction result for the distribution of urban traffic demand.

Images