Method for traffic assignment in large-scale urban traffic network based on overrelaxation iteration method

Abstract

A new method for large-scale urban traffic network flow distribution based on super relaxation iteration method, according to road traffic network and urban travel distribution, the shortest path on the road traffic network is calculated and the initial path set is established, and the shortest path set is updated by updating the traffic and travel time of the road section, then the gradient projection algorithm is applied to update the path traffic, further, based on the basic principle of super relaxation iteration method, the path traffic of the current OD pair is corrected, and the Armijo criterion is applied to determine the relaxation factor in the super relaxation iteration method, finally, the calculation accuracy index is calculated, and the convergence state is judged. On the basis of the path flow distribution technology, the principle of super relaxation iteration method is applied, a new method for large-scale traffic network flow distribution is proposed, and the test results in the large-scale traffic network show that the large-scale urban traffic network flow distribution method based on super relaxation iteration method has higher calculation efficiency and faster convergence speed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of traffic network modeling, in particular to a large-scale urban traffic network flow distribution method based on over-relaxation iteration method. BACKGROUND

[0002] In the strategic planning of traffic system, traffic distribution is the process of finding the traffic distribution pattern in a known road traffic network. It helps traffic planners and operators better understand the role of traffic infrastructure, so as to develop a sustainable traffic system. Among them, user equilibrium is the basic principle of traffic distribution problem, and the core is that any traveler cannot unilaterally change his travel route to reduce travel time.

[0003] In the past few decades, designing effective solution algorithms for this problem has attracted widespread attention, which mainly includes three categories, including (1) link-based algorithm; (2) path-based algorithm and (3) origin-based algorithm, so far, for the traffic distribution problem of multiple OD pairs, there are two traffic update strategies that can be applied to the above algorithms, respectively, Jacobian iteration method and Gauss-Seidel iteration method, Jacobian iteration method and Gauss-Seidel iteration method are both classical iterative methods for solving linear equations, and since the convergence speed of Gauss-Seidel iteration method is faster than that of Jacobian iteration method, Gauss-Seidel iteration method can better improve the calculation efficiency of solving the traffic distribution problem of urban road traffic network.

[0004] However, the over-relaxation iteration method is a new iterative method for solving linear equations, which has a faster convergence speed, so it is imminent to introduce the over-relaxation iteration method into the field of traffic distribution to further improve the calculation efficiency of solving the traffic distribution problem of urban road traffic network, which has good development prospects. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a large-scale urban traffic network flow distribution method based on over-relaxation iteration method.

[0006] In order to solve the above technical problems, the technical scheme adopted by the present application is: a large-scale urban traffic network flow distribution method based on over-relaxation iteration method, specifically comprising the following steps: S1, collecting basic parameters of urban road traffic network, motor vehicle traffic trip demand distribution, parameters required for applying Armijo criterion to determine relaxation factor; the traffic network basic parameters include free flow travel time of each link, traffic flow carrying capacity on the link, link impedance function, traffic distribution accuracy threshold, the traffic trip demand distribution includes traffic volume between the origin and the destination, the free flow travel time, the traffic flow carrying capacity are input into the link impedance function, and the link impedance function is input into the optimization objective function equation, specifically Its constraints are: ,

[0007] ,

[0008] ,

[0009] ,

[0010] Where A represents the set of road segments, a represents road segment a, a∈A; W represents the set of OD pairs, w represents OD pair w, w∈W; K represents the set of paths k between OD pairs w, k∈K; q w This indicates the travel demand between OD and w; f k w This represents the flow of OD to path k between w; Indicates the path between OD and w Traffic; v a 'and v a "represents the non-shortest path k and the shortest path respectively." Traffic flow on section a of the road; t a The impedance function of road segment a; The coefficient representing the relationship between road segment a and path k is 1 when road segment a is on the k-th path of OD pair w, and 0 otherwise; the formula for the road segment impedance function is as follows. , where t a (v) is the road segment impedance function, t0 is the free-flow travel time, and C a Let α and β represent the traffic flow carrying capacity on road segment a, where α and β are preset parameters set by the user. aS2, generating the shortest path tree for each origin O according to free flow travel time, determining the shortest path between each OD pair according to the shortest path tree, adding the shortest path between each OD pair into the path set corresponding to the OD pair, loading all traffic demand between the OD pair on the initial path corresponding to the OD pair, and then accumulating the traffic volume on the path to the road segments in the road network to update the road segment flow, road segment travel cost and road segment impedance function derivative, to complete the path set initialization; S3, updating the shortest path tree for each origin O according to the updated road segment travel cost, comparing the shortest path for each OD pair with all paths in the path set of the OD pair, if the shortest path is not in the existing path set, adding the shortest path into the path set and temporarily setting the path flow on the path to 0, to complete the augmentation of the path set; S4, determining the relaxation factor according to the non-precise linear search Armijo criterion in combination with the parameters required for applying the Armijo criterion to determine the relaxation factor, calculating the predicted reduction and the actual reduction along the descent direction respectively, gradually shortening the step length from the initial step length to search, and outputting the first step length satisfying the sufficient descent condition as the relaxation factor value required for the current iteration; S5, adjusting the path flow by applying the gradient projection-over-relaxation algorithm, determining the descent direction according to the travel time of the non-shortest path and the shortest path, determining the step length according to the second derivative information of the impedance function, and updating the path flow information through the projection operation, based on the basic principle of the over-relaxation iteration method, updating the path flow information of the current OD pair by weighting the path flow information generated by the previous OD pair and the path flow information obtained by the last iteration of the current OD pair, determining the weight according to the relaxation factor value, then, accumulating the path flow information of the current OD pair to the road segments in the road network to update the road segment flow, road segment travel cost and derivative of the road segment impedance function, and then updating the path flow information of the next OD pair until the path flow adjustment of all OD pairs is completed; S6, calculating the traffic assignment relative error, and outputting the road segment flow and travel time on the road network when the accuracy threshold is satisfied, otherwise returning to step S3.

[0011] Preferably, step S4 specifically comprises:

[0012] Step S41, let m=0, where m is a non-negative integer;

[0013] Step S42, calculating the actual reduction along the steepest descent direction ared(n) and the predicted reduction along the steepest descent direction pred(n) of the nth iteration;

[0014] ,

[0015] ,

[0016] ,

[0017] ,

[0018] wherein, is the objective function value at the n-th iteration, is the objective function value at the (n+1)-th iteration;c k w (n) and respectively represent the travel time of the non-shortest path k and the travel time of the shortest path at the n-th iteration; μ is a fixed parameter satisfying 0<μ<1, and s is a fixed parameter representing the search starting point when the non-precise Armijo rule is used to determine the relaxation factor, represents the relationship coefficient of the link a-path , which takes the value of 1 when the link a is on the first path of the OD pair w, otherwise, it takes the value of 0;

[0019] Step S43, it is judged whether the sufficient descent condition is satisfied, if yes, i.e.ared(n)>η·pred(n), wherein η is a fixed parameter satisfying 0<η<1, the relaxation factor is set as , otherwise, m n =m+1, and steps S42-S43 are repeated.

[0020] Preferably, step S5 specifically comprises:

[0021] Step S51, the descent direction p w , of the OD pair w is determined by the gradient projection algorithm; k w and respectively represent the travel time of the non-shortest path k of the OD pair w and the travel time of the shortest path of the OD pair w;

[0022] Step S52, the path flow of the OD pair w is updated according to the first and second derivative information of the link impedance function, and the specific formula is:

[0023] ,

[0024] ,

[0025] ,

[0026] wherein, f k w (n+1) represents the flow on the path k between the OD pair w at the (n+1)-th iteration, represents the flow on the path f k w (n) represents the flow on path k between OD pair w at the n th iteration, and a(n) is the step size of the gradient projection algorithm, s k w is the target equation is the second derivative of t a (v a ) is the derivative of the link cost with respect to the flow;

[0027] In step S53, based on the basic principle of the successive over-relaxation iteration method, the path flow information generated by the previous OD pair and the path flow information obtained by the last iteration of the current OD pair are weighted, the path flow information generated by the previous OD pair is weighted by 1-w n , and the un-updated path flow is weighted by w n The path flow of the current OD pair is updated by using the gradient projection algorithm, and the specific path flow updating formula is The above formula represents the iteration format of the successive over-relaxation iteration (SOR), represents the path set of OD pair w, and f w (n) represents the path flow information generated by the previous OD pair, and f w (n+1) represents the un-updated path flow;

[0028] In step S54, the link flow, the link travel cost, and the derivative of the link impedance function are updated; and the step S51 is returned until the path flow adjustment of all OD pairs is completed.

[0029] Preferably, the step S6 is specifically: judging whether the relative error is less than or equal to the traffic distribution accuracy threshold value, if yes, that is, the accuracy threshold value is met, the link flow and the travel time on the road network are output, and if no, the step S3 is returned, and the relative error RG is calculated according to the formula: , wherein, φ w is the shortest travel time of OD pair w, and v a represents the flow on link a, and t a (v a ) represents the link cost of link a.

[0030] The scope of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combinations of the above technical features or their equivalent features. For example, the technical solutions formed by mutually replacing the above features and the technical features disclosed in the present application (but not limited to) having similar functions.

[0031] Compared with the prior art, the present application has the following advantages: the gradient projection algorithm with the over-relaxation iteration method is used, the Armijo criterion is used to determine the relaxation factor, a high-precision solution of the traffic distribution problem of the urban road traffic network is generated, the convergence speed is fast, it is simple and easy to understand, and it has a good development prospect. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 is a method flowchart of the present application;

[0033] Figure 2 is a flowchart of the traffic adjustment module of the gradient projection algorithm with the over-relaxation iteration;

[0034] Figure 3 is a convergence performance comparison chart of the new algorithm proposed in the present application, wherein Figure 3 (a) is a convergence error and CPU time consumption chart of the Annahiem network; Figure 3 (b) is a convergence error and CPU time consumption chart of the Chicago backbone network;

[0035] Figure 4 is a CPU time consumption and iteration round comparison chart of the new algorithm proposed in the present application, wherein Figure 4 (a) is a CPU time consumption and iteration round comparison chart of the Annahiem network; Figure 4 (b) is a CPU time consumption and iteration round comparison chart of the Chicago backbone. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical scheme and advantages of the present application clearer and more apparent, the present application will be further described in detail below with specific embodiments and with reference to the drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the application range of the present application.

[0037] In the present application, the phrase "embodiment" means that the specific features, structures or characteristics described in combination with the embodiment can be included in at least one embodiment of the present application. The appearance of this phrase at various places in the specification does not necessarily mean the same embodiment, nor is it an independent or alternative embodiment that is not mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0038] As Figure 1 shown, the large-scale urban traffic network traffic distribution method based on the over-relaxation iteration method specifically includes the following steps.

[0039] Step S1, collecting basic parameters of urban road traffic network, distribution of motor vehicle traffic demand, and parameters required for determining the relaxation factor using the Armijo criterion; the basic parameters of the traffic network include free-flow travel time of each road segment, traffic flow carrying capacity on the road segment, road segment impedance function, and traffic assignment accuracy threshold; the distribution of traffic demand includes traffic volume between the origin and the destination; the free-flow travel time and the traffic flow carrying capacity are input into the road segment impedance function, and the formula of the road segment impedance function is where t a (v) is the road segment impedance function, t0 is the free-flow travel time, C a is the traffic flow carrying capacity on the road segment a, and a and b are preset parameters, and v a represents the flow of the road segment a, and the road segment impedance function is input into the objective function equation, specifically , and the constraint condition is ,

[0040]

[0041] ,

[0042] ,

[0043] where A represents a set of road segments, a represents the road segment a, and a∈A; W represents a set of OD pairs, w represents the OD pair w, and w∈W; K represents a set of paths k between the OD pair w, and k∈K; q w represents the travel demand between the OD pair w; f k w represents the flow of the path k between the OD pair w; v a ’ and v a ” represent the road segment flow on the non-shortest path k and the shortest path , respectively; t a (v) represents the impedance function of the road segment a; represents the relationship coefficient of the road segment a-path k, which is 1 when the road segment a is on the kth path between the OD pair w, and 0 otherwise.

[0044] Step S2, generating a shortest path tree for each origin O according to the free-flow travel time, determining the shortest path between each OD pair according to the shortest path tree, adding the shortest path between each OD pair as an initial path into the path set corresponding to the OD pair, loading all traffic travel demands between the OD pair on the initial path corresponding to the OD pair, and then accumulating the vehicle flow on the path to the road segments in the road network to update the road segment flow, road segment travel cost, and derivative of the road segment impedance function, and initializing the path set.

[0045] Step S3: Based on the updated travel cost of the road segment, update the shortest path tree for each starting point O. Compare the shortest path of each OD pair with all paths in the path set of the OD pair. If the shortest path is not in the existing path set, add the shortest path to the path set and temporarily set the path flow on that path to 0, thus expanding the path set.

[0046] Step S4: Combining the parameters required for determining the relaxation factor using the Armijo criterion, determine the relaxation factor according to the non-exact linear search Armijo criterion. Calculate the expected and actual reduction amounts along the descent direction, and gradually shorten the step size from the initial step size to perform the search. Output the first step size that satisfies the sufficient descent condition, which is the relaxation factor value required for the current iteration. Specifically:

[0047] Step S41, let m = 0, where m is a non-negative integer;

[0048] Step S42: Calculate the actual descent amount ared(n) along the steepest descent direction in the nth iteration, and the expected descent amount pred(n) along the steepest descent direction;

[0049] ,

[0050] ,

[0051] ,

[0052] ,

[0053] in, It is the value of the objective function in the nth iteration. It is the value of the objective function in the (n+1)th iteration; c k w (n) and Let these represent the travel time of the non-shortest path k and the shortest path, respectively, in the nth iteration. The travel time; μ is a fixed parameter satisfying 0 < μ < 1, s is a fixed parameter representing the search starting point when determining the relaxation factor using the inexact Armijo criterion, Indicates road segment a - path The relationship coefficient, when road segment a is in the OD pair with w. If the path is on the same path, the value is 1; otherwise, the value is 0.

[0054] Step S43: Determine if the sufficient descent condition is met. If it is, i.e., ared(n) > η·pred(n), where η is a fixed parameter satisfying 0 < η < 1, let the relaxation factor... Otherwise let m n=m+1, repeat steps S42-S43.

[0055] Step S5 (e.g.) Figure 2 The gradient projection-super-relaxation algorithm is applied to adjust path flow. The descent direction is determined based on the travel time of the non-shortest and shortest paths, and the step size is determined based on the second derivative of the impedance function. Path flow information is updated through projection operations. Based on the fundamental principle of the super-relaxation iterative method, the path flow information of the current OD pair is updated by weighting the path flow information generated by previous OD pairs with the path flow information obtained from the previous iteration of the current OD pair. The weights are determined by the relaxation factor in step 4. Next, the path flow information of the current OD pair is accumulated onto road segments in the road network to update the segment flow, segment travel cost, and the derivative of the segment impedance function. Then, the path flow information of the next OD pair is updated until the path flow adjustment for all OD pairs is completed.

[0056] Step S5 specifically involves: Step S51, determining the descent direction p of OD for w using the gradient projection algorithm. w , c k w and Let k represent the travel time of the non-shortest path k from OD to w, and k represent the shortest path from OD to w. Travel time;

[0057] Step S52: Update the path flow of OD to w based on the first and second derivatives of the road segment impedance function. The specific formula is as follows:

[0058] ,

[0059] ,

[0060] ,

[0061] Among them, f k w (n+1) represents the flow of OD on path k between w in the (n+1)th iteration. This represents the path between OD and w at the (n+1)th iteration. Traffic on f k w α(n) represents the flow rate of OD on path k between w and w in the nth iteration, α(n) is the step size of the gradient projection algorithm, and s k w For the objective equation The second derivative, t a '(v a () represents the derivative of road segment cost with respect to traffic flow;

[0062] Step S53, based on the basic principle of the over-relaxation iteration method, the GP-SOR algorithm is used to update the path flow of the current OD pair by weighting the path flow information generated by the previous OD pair and the path flow information obtained by the last iteration of the current OD pair, and the weight of the path flow information generated by the previous OD pair is 1-w n , the weight of the un-updated path flow is w n , the specific path flow updating formula is

[0063] The above formula represents the iteration format of the over-relaxation iteration, i.e. SOR, denotes the path set of the OD pair w, and ϕ is a function abstraction symbol, f w (n) represents the path flow information generated by the previous OD pair, f w (n+1) is the un-updated path flow.

[0064] Step S54, update the link flow, link travel cost and link impedance function derivative; return to step S51 until the path flow adjustment work of all OD pairs is completed.

[0065] Step S6, calculate the traffic assignment relative error, when the accuracy threshold is met, output the link flow and travel time on the road network, otherwise return to step 3, which is: judge whether the relative error is less than or equal to the traffic assignment accuracy threshold, if yes, i.e. meet the accuracy threshold, output the link flow and travel time on the road network, if no, return to step S3, the relative error RG calculation formula is: , wherein, φ w is the shortest travel time of the OD pair w, v a denotes the flow on link a, t a (v a ) represents the link cost of link a.

[0066] Figures 3 and 4 show the calculation results of the gradient projection algorithm with different flow update strategies. In these figures, the horizontal axis represents the CPU time (or iteration round), and the vertical axis represents the convergence error RG(or CPU time). It can be seen that when the Armijo criterion is used to determine the relaxation factor, the gradient projection algorithm with the super-relaxation iteration (GP-SOR) can outperform the gradient projection algorithm with the Gauss-Seidel iteration (GP-GS) in all test networks with a large advantage. Specifically, for the Annapolis network (see Fig. 3(a)), the GP-GS algorithm needs about 9 seconds to reach the precision of level IV, while the GP-SOR algorithm only needs 2 seconds to converge to the same precision, and the convergence efficiency is increased by about 77%. In addition, as shown in Fig. 4, in the case of applying the Armijo criterion to determine the relaxation factor, the GP-SOR algorithm requires fewer iteration times than the GP-GS algorithm. Specifically, for the Chicago backbone (Fig. 4(b)), the GP-SOR algorithm can reduce the iteration times by about 2 times compared with the GP-GS algorithm.

[0067] The above embodiments are only for illustrating the technical concepts and characteristics of the present application, and the purpose is to enable those skilled in the art to understand the content of the present application and to implement it, and cannot limit the protection scope of the present application. Any equivalent changes or modifications made in accordance with the spirit and essence of the present application shall be covered within the protection scope of the present application.

Claims

1. A method for large-scale urban traffic network flow allocation based on the super-relaxation iterative method, characterized in that: Specifically, the following steps are included: S1. Collect basic parameters of the urban road traffic network, the distribution of motor vehicle travel demand, and the parameters required for determining relaxation factors using the Armijo criterion. The basic traffic network parameters include free-flow travel time for each road segment, traffic flow carrying capacity on each road segment, road segment impedance function, and traffic assignment accuracy threshold. The traffic demand distribution includes traffic volume between the origin and destination. Input the free-flow travel time and traffic flow carrying capacity into the road segment impedance function, and input the road segment impedance function into the optimization objective function equation. Specifically... Its constraints are: ， ， ， ， Where A represents the set of road segments, a represents road segment a, a∈A; W represents the set of OD pairs, w represents OD pair w, w∈W; K represents the set of paths k between OD pairs w, k∈K; q w This indicates the travel demand between OD and w; f k w This represents the flow of OD to path k between w. Indicates the path between OD and w Traffic; v a 'and v a "represents the non-shortest path k and the shortest path respectively." Traffic flow on section a of the road; t a The impedance function of road segment a; The coefficient representing the relationship between road segment a and path k is 1 when road segment a is on the k-th path of OD pair w, and 0 otherwise; the formula for the road segment impedance function is as follows. , where t a (v) is the road segment impedance function, t0 is the free-flow travel time, and C a Let α and β represent the traffic flow carrying capacity on road segment a, where α and β are preset parameters set by the user. a This represents the traffic flow in road segment a. S2. Based on the free-flow travel time, generate a shortest path tree for each starting point O, and determine the shortest path between each OD pair based on the shortest path tree. Add the shortest path between each OD pair as the initial path to the path set corresponding to that OD pair. Load all traffic travel demands between OD pairs onto the initial path corresponding to that OD pair, and then accumulate the traffic flow on the path to the road segment in the road network to update the road segment flow, road segment travel cost and road segment impedance function derivative, thus completing the initialization of the path set. S3. Based on the updated travel cost of the road segment, update the shortest path tree of each starting point O. Compare the shortest path of each OD pair with all paths in the path set of the OD pair. If the shortest path is not in the existing path set, add the shortest path to the path set and temporarily set the path flow on the path to 0 to complete the augmentation of the path set. S4. Combining the parameters required when applying the Armijo criterion to determine the relaxation factor, determine the relaxation factor according to the Armijo criterion of non-exact linear search. Calculate the expected reduction and actual reduction along the descent direction respectively. Search by gradually shortening the step size from the initial step size. Output the first step size that satisfies the sufficient descent condition, which is the relaxation factor value required for the current iteration. S5. Apply the gradient projection-super-relaxation algorithm to adjust path flow. Determine the descent direction based on the travel time of the non-shortest path and the shortest path, determine the step size based on the second derivative information of the impedance function, and update the path flow information through projection operation. Based on the basic principle of the super-relaxation iteration method, the path flow information generated by previous OD pairs and the path flow information obtained by the current OD pair in the previous iteration are weighted. The weight is determined according to the relaxation factor value to update the path flow information of the current OD pair. Then, the path flow information of the current OD pair is accumulated onto the road segments in the road network to update the road segment flow, road segment travel cost, and the derivative of the road segment impedance function. Then, the path flow information of the next OD pair is updated until the path flow adjustment of all OD pairs is completed. S6, calculate the relative error of traffic assignment. If the accuracy threshold is met, output the traffic flow and travel time of each road segment on the road network; otherwise, return to step S3.

2. The method for large-scale urban traffic network flow allocation based on the super-relaxation iterative method according to claim 1, characterized in that: Step S4 specifically includes: Step S41, let m = 0, where m is a non-negative integer; Step S42: Calculate the actual descent amount ared(n) along the steepest descent direction in the nth iteration, and the expected descent amount pred(n) along the steepest descent direction; ， ， ， ， in, It is the value of the objective function in the nth iteration. It is the value of the objective function in the (n+1)th iteration; and Let these represent the travel time of the non-shortest path k and the shortest path, respectively, in the nth iteration. The travel time; μ is a fixed parameter satisfying 0 < μ < 1, S is a fixed parameter representing the search starting point when determining the relaxation factor using the inexact Armijo criterion, Indicates road segment a - path The relationship coefficient, when road segment a is in the OD pair with w. If the path is on the same path, the value is 1; otherwise, the value is 0. Step S43: Determine if the sufficient descent condition is met. If it is, i.e., ared(n) > η·pred(n), where η is a fixed parameter satisfying 0 < η < 1, let the relaxation factor... Otherwise let m n =m+1, repeat steps S42-S43.

3. The method for large-scale urban traffic network flow allocation based on the super-relaxation iterative method according to claim 2, characterized in that: Step S5 specifically includes: Step S51: Determine the descent direction p of OD with respect to w using the gradient projection algorithm. w , ;c k w and Let k represent the travel time of the non-shortest path k from OD to w, and k represent the shortest path from OD to w. Travel time; Step S52: Update the path flow of OD to w based on the first and second derivatives of the road segment impedance function. The specific formula is as follows: ， ， ， Among them, f k w (n+1) represents the flow of OD on path k between w in the (n+1)th iteration. This represents the path between OD and w in the (n+1)th iteration. Traffic on f k w α(n) represents the flow rate of OD on path k between w and w in the nth iteration, α(n) is the step size of the gradient projection algorithm, and s k w For the objective equation The second derivative, t' a (v a () represents the derivative of road segment cost with respect to traffic flow; Step S53: Based on the fundamental principle of the over-relaxation iterative method, the path flow information generated by previous OD pairs and the path flow information obtained by the current OD pair in the previous iteration are weighted together. The weight of the path flow information generated by previous OD pairs is 1-w. n The unupdated path traffic weight is w. n The gradient projection algorithm is used to update the path flow of the current OD pair. The specific path flow update formula is as follows: The above formula represents the iterative format of the over-relaxation iteration, i.e., SOR. Let ϕ represent the set of paths from OD to w, where ϕ is an abstract symbol for a function, and f w (n) represents the path traffic information previously generated by the OD pair, f w (n+1) represents the path traffic that has not been updated; Step S54: Update segment flow, segment travel cost, and segment impedance function derivative; return to step S51 until path flow adjustment for all OD pairs is completed.

4. The method for large-scale urban traffic network flow allocation based on the super-relaxation iterative method according to claim 3, characterized in that: Step S6 specifically involves determining whether the relative error is less than or equal to the traffic assignment accuracy threshold. If it is less than or equal to the threshold, the accuracy threshold is met, and the traffic flow and travel time for each road segment on the road network are output. If it is greater than the threshold, the process returns to step S3. The formula for calculating the relative error RG is as follows: , where φ w For the shortest travel time of OD to w, v a t represents the traffic flow on road segment a. a (v a ) represents the segment cost of segment a.

Images