Traffic flow prediction method and system for multi-hop ride-sharing network of online car-hailing

Abstract

The present invention provides a traffic flow prediction method and system for a multi-hop ride-sharing travel network of online car-hailing, and relates to the field of traffic flow prediction. The traffic flow prediction method for a multi-hop ride-sharing travel network of online car-hailing includes: inputting the topological structure, travel mode and travel demand of the target city road network; establishing traffic flow constraints and ride-sharing matching constraints that meet the characteristics of multi-hop ride-sharing travel; calculating the travel costs of travelers in the target city road network through traffic flow constraints and ride-sharing matching constraints; completing the establishment of an equivalent implicit variational inequality model; solving the equivalent implicit variational inequality model to complete the prediction of the traffic flow of the multi-hop ride-sharing travel network of online car-hailing. This method can improve the computational efficiency of traffic flow prediction of the multi-hop ride-sharing travel network, and more clearly define the impact of online car-hailing ride-sharing travel on urban traffic, thereby taking efficient management measures, improving the construction of online car-hailing platforms, and promoting the development of modern urban transportation systems.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic flow prediction, and specifically to a method and system for predicting traffic flow in an online car-hailing multi-hop ride-sharing travel network. Background Art

[0002] In recent years, with the rise of the sharing economy and the development of communication technology, ride-sharing (i.e., carpooling) has attracted more and more attention as a simple but potentially very effective way to curb urban congestion. Ride-sharing refers to a mode of travel in which travelers with different starting and ending points are matched through online car-hailing platforms and become carpooling drivers and carpooling passengers respectively. Passengers share the travel expenses by paying a certain amount of compensation to the driver. This can reduce the travel expenses of travelers while meeting their travel needs. It has great potential in improving vehicle utilization and reducing traffic congestion. However, due to the differences in the starting and ending points of online car-hailing drivers and passengers, drivers may incur additional congestion and travel time when taking detours to pick up and drop off passengers. Therefore, it is difficult to determine the impact of online car-hailing ride-sharing on urban traffic flow.

[0003] A method and medium for predicting urban traffic flow based on graph convolutional neural network (application number CN201910637679.8), generates a distance matrix according to the latitude and longitude information corresponding to each node, calculates a reachable matrix according to the speed limit average and the distance matrix, constructs an initial traffic flow prediction model for predicting traffic flow speed, and inputs the traffic flow speed information and the reachable matrix into the initial traffic flow prediction model, so that the initial traffic flow prediction model outputs the traffic flow speed prediction value according to the traffic flow speed information and the reachable matrix, and trains the initial traffic flow prediction model to determine the final traffic flow prediction model. A method, system and device for predicting urban traffic flow based on generative adversarial network (application number CN202310681033.6), according to the travel demand data of each traffic flow area in the current stage of the study city, uses the Node2Vec graph embedding algorithm to determine the embedding matrix of the current stage of the study city; the embedding matrix of the current stage of the study city and the random noise conforming to the Gaussian distribution are used as input values, and are input into the trained generator to predict the traffic flow of the study city in the future stage. Both inventions take into account traffic flow forecasts on urban transportation networks, but neither takes into account the emerging model of ride-sharing. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the shortcomings of the prior art, the present invention provides a method and system for predicting traffic flow in a multi-hop ride-sharing network of online car-hailing, based on an equivalent implicit variational inequality model of road sections and a projection method incorporating parallel cost averaging (PM-PCA). The combination of the two can predict the traffic flow of online car-hailing ride-sharing in urban transportation networks. Through the present invention, the computational efficiency of traffic flow prediction in a multi-hop ride-sharing network can be improved, and the impact of online car-hailing ride-sharing on urban transportation can be more clearly understood, thereby adopting efficient management methods, improving the construction of online car-hailing platforms, and promoting the development of modern urban transportation systems.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] In a first aspect, a method for predicting traffic flow in a multi-hop ride-sharing network of online car-hailing vehicles comprises:

[0009] Input the topology, travel mode and travel demand of the target city’s road network;

[0010] According to the topological structure, travel mode and travel demand of the target city road network, traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel are established;

[0011] The travel costs of travelers in the target city road network are calculated through traffic flow constraints and ride-sharing matching constraints;

[0012] The equivalent implicit variational inequality model is established by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and the travel costs of travelers;

[0013] The projection algorithm combined with parallel cost averaging is used to solve the equivalent implicit variational inequality model to predict the traffic flow of the multi-hop ride-hailing travel network.

[0014] Preferably, the topological structure of the target urban road network is (N, A), where N is the set of OD points in the urban road traffic network, and A is the set of directed road segments connecting the OD points in the urban road traffic network;

[0015] The travel demand between different OD pairs is q w , where w represents an OD pair, W represents the set of OD pairs in the urban road traffic network, w∈W, and the starting and ending points of OD pair w are represented by o w ∈O and d w∈D means that O is the set of starting points, D is the set of end points, and both O and D are subsets of N;

[0016] According to the characteristics of multi-hop carpooling, travel modes are divided into three categories: SD represents the set of drivers who have no matched passengers and travel alone, RD represents the set of drivers who have matched passengers and participate in carpooling, and R represents the set of passengers who have matched drivers and participate in carpooling. Let I represent the set of three travel modes in the urban road traffic network, I = SD∪RD∪R.

[0017] Preferably, the traffic flow constraint is expressed as:

[0018]

[0019]

[0020] Where x a,i represents the traffic flow of travel mode i on road segment a; represents the traffic flow of travel mode i between OD pair w on road segment a; represents the traffic flow between OD pairs w on road segment a; in addition, let the vector as well as The vector form of traffic flow;

[0021] The starting and ending points in the urban road network are:

[0022]

[0023] Where OUT(n) represents the set of directed segments starting from node n, and IN(·) represents the set of directed segments ending at node n;

[0024] For road nodes that are not starting or ending points in the urban road network, there is the following flow relationship:

[0025]

[0026]

[0027] In the formula, N\{o w ,d w} represents a set of road nodes that are not starting or ending points.

[0028] Preferably, the carpool matching constraint is expressed as:

[0029]

[0030] Where Γ r (i) represents mapping the online car-hailing driver i to the corresponding online car-hailing passenger, that is, Γ r:RD→R,M i Indicates the number of passengers in the car of online car-hailing driver i;

[0031] The ride-sharing matching constraints are rewritten into a compact structure based on the vector form of traffic flow:

[0032] Bx=0 or ∑ w B w X w =0

[0033] Where B=[B1 B2...B w∈W ...B |W| ]; is a block matrix of |A|×|A|; It is a |RD|×(2|RD|+1) matrix.

[0034] Preferably, the calculation of the travel cost of travelers in the target city road network includes:

[0035] According to the characteristics of online car-hailing ride-sharing travel, the travel cost of travelers includes time cost, inconvenience cost and ride-sharing fee, which can be expressed as:

[0036]

[0037] The road flow vector represents the travel cost of a traveler of mode i between OD pairs w on road segment a, which is a function of x; represents the travel time of travelers on road section a, function t a (·) is about Strictly monotonically increasing empirical function; α i represents the time value of travel mode i, β i represents the inconvenience coefficient of travel mode i; B a,i represents the basic fee collected / paid by online car-hailing driver i / online car-hailing passenger i on road section a; m i represents the price fluctuation parameter of travel mode i;

[0038] According to the characteristics of the supply and demand balance of online car-hailing ride-sharing and the complementary relaxation of linear programming, the car-sharing subsidy and premium are combined with the travel cost function, and the generalized travel cost function of the traveler is expressed as:

[0039]

[0040] In the formula represents the generalized travel cost of a traveler of mode i between OD pairs w on road segment a; a,iis the Lagrange multiplier corresponding to the carpooling matching constraint, whose value is equal to the value of the premium and the subsidy; Γ rd (i) represents mapping the online car-hailing passenger i to the corresponding online car-hailing driver, that is, Γ rd :R→RD; let vector and denote the generalized travel cost and Lagrange multiplier respectively.

[0041] Preferably, the establishment of the equivalent implicit variational inequality model includes:

[0042] Establish a variational inequality model [RUE'], the Lagrange multiplier λ corresponds to the multiplication matching constraint. When the value of λ is given, establish a variational inequality model to express the multiplication equilibrium problem and find x * ∈Ω' RUE Satisfy the formula:

[0043]

[0044] Where λ * is a parameter of a given value, equal to the Lagrange multiplier in the shared equilibrium state; Ω' RUE represents the feasible set of x, that is, the set of road flow that satisfies the traffic flow constraint and the ride-sharing matching constraint;

[0045] Establish a fixed point model [RUE”], and establish a fixed point model [RUE”] equivalent to the variational inequality model [RUE']:

[0046]

[0047] Where ψ(·) represents the network load mapping from the generalized travel cost set to the traffic flow set;

[0048] Establish an implicit variational inequality model. The solution x(λ) of the carpooling equilibrium problem is an implicit function of λ, whose value is determined by solving [RUE'] or [RUE”] to find λ * ∈Ω λ Satisfy the formula:

[0049]

[0050] Where Ω λ =R |A|·|RD| , represents the feasible set of λ, which is the set of Lagrange multipliers that satisfy the traffic flow constraints and the ridesharing matching constraints.

[0051] Preferably, the steps of solving the equivalent implicit variational inequality model include:

[0052] Initialization step, set the accuracy ε>0, β (0)>0, k=l=0, given the initial Lagrange multiplier λ (0) And the initial flow solution x (0) ;

[0053] Find the direction step, update the vector Then, let each processor execute a Dijkstra algorithm to find the shortest path from a starting point to all destinations, transforming the travel demand q w Loaded on the shortest path between OD pair w, generating auxiliary flow solution y (k) and auxiliary travel time

[0054] Update the cost vector step, the vector To iterate:

[0055]

[0056] Check the cost accuracy step, if Enter the step of updating the inconvenience coefficient and the Lagrange multiplier; otherwise k=k+1, return to the step of finding the direction;

[0057] Update the inconvenience coefficient and Lagrange multiplier step, for the vector β (l) and λ (l+1) To update:

[0058]

[0059] λ (l+1) =λ (l) -β (l) (Bx (k) )

[0060] Check the accuracy of the Lagrange multipliers step, if Then terminate the calculation and output λ (l) and x (k) ; Otherwise, let l=l+1, k=0, and return to the search direction step.

[0061] In the second aspect, a traffic flow prediction system for a multi-hop ride-sharing network of online car-hailing includes the following modules:

[0062] An input module, configured to input the topology of the target city road network, travel modes and travel demands;

[0063] A first establishment module is configured to establish traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel according to the topological structure, travel mode and travel demand of the target city road network;

[0064] A first calculation module is configured to calculate the travel cost of travelers in the target city road network through traffic flow constraints and ride-sharing matching constraints;

[0065] The second building module is configured to complete the building of an equivalent implicit variational inequality model by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and travel costs of travelers;

[0066] The second computing module is configured to solve the equivalent implicit variational inequality model in combination with the parallel cost averaging projection algorithm to complete the traffic flow prediction of the online car-hailing multi-hop ride-sharing travel network.

[0067] In a third aspect, a computer-readable storage medium storing one or more programs is provided, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the methods described.

[0068] In a fourth aspect, a computing device is provided, including:

[0069] One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the methods described.

[0070] (III) Beneficial effects

[0071] The present invention provides a method for predicting traffic flow in a multi-hop ride-hailing network, which predicts the traffic flow distribution of an urban transportation network in the presence of ride-hailing services, thereby analyzing the impact of ride-hailing on urban transportation. By using the present invention, relevant departments can more clearly understand the impact of ride-hailing on urban transportation, thereby taking efficient management measures, improving the construction of online ride-hailing platforms, and promoting the development of modern urban transportation systems. The present invention has the following advantages: ① Innovatively studying the ride-hailing matching between drivers and passengers with different starting and ending points, which is more in line with actual ride-hailing scenarios, and the research difficulty is increased and more challenging; ② Innovatively proposing an equivalent implicit variational inequality model based on road sections, and solving the traffic flow equilibrium solution by combining the projection method incorporating parallel cost averaging (PM-PCA) with parallel cost averaging, thereby improving the computational efficiency of traffic flow prediction in a multi-hop ride-hailing network. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a flow chart of a method for predicting traffic flow in a multi-hop ride-sharing network for online car-hailing services according to the present invention;

[0073] Figure 2 This is the topological structure of the target city road network of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the accompanying drawings of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] Example

[0076] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting traffic flow in a multi-hop ride-sharing network of online car-hailing, comprising:

[0077] Input the topology, travel mode and travel demand of the target city’s road network;

[0078] According to the topological structure, travel mode and travel demand of the target city road network, traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel are established;

[0079] The travel costs of travelers in the target city road network are calculated through traffic flow constraints and ride-sharing matching constraints;

[0080] The equivalent implicit variational inequality model is established by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and the travel costs of travelers;

[0081] The projection algorithm combined with parallel cost averaging is used to solve the equivalent implicit variational inequality model to predict the traffic flow of the multi-hop ride-hailing travel network.

[0082] Specific:

[0083] According to the target city road network, a multi-hop ride-sharing travel network model is constructed. The specific process is as follows:

[0084] Step 1: Obtain and input the topological structure, travel mode and travel demand of the target city road network.

[0085] Based on the geographical location and traffic elements, the target city area is divided into multiple traffic zones. Each traffic zone can be used as the starting point or the end point of a traffic trip. The set of starting and ending points are represented by O and D respectively. The starting and ending points with travel demand form OD pairs w, and the set of OD pairs in the network is represented by W. There is a travel demand q between OD pairs w. w .

[0086] By consulting the map, the topological structure (N, A) of the target city road network is obtained. Each traffic zone is a node in the topological structure, represented by n. The set of nodes is represented by N. The urban roads connecting each traffic zone are directed sections in the topological structure, represented by a. The set of directed sections is represented by A. The topological structure of the urban road network can be stored by adjacency matrix, adjacency table, etc.

[0087] Travel demand w It can be obtained through traditional methods such as travel demand surveys and OD demand inversion, as well as emerging methods such as intelligent traffic monitoring systems, mobile phone signaling and GPS.

[0088] The travel mode in the target city road network is denoted by i. The set of travel modes is denoted by I. The travel modes in the network include an empty car driver driving alone, a ride-hailing driver with passengers in the car, and a ride-hailing passenger riding in a ride-hailing car, denoted by i∈SD, i∈RD, and i∈R, respectively.

[0089] After sorting, the topological structure of the target city road network is as follows: Figure 2 As shown in Table 1, there are 24 nodes, 76 road segments, and 528 OD pairs in this road network. The travel modes in this road network are shown in Table 1.

[0090] Table 1 Travel modes of target city road network

[0091]

[0092] Step 2: Establish traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel.

[0093] ① Traffic flow constraints

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] Where x a,i represents the traffic flow of travel mode i on road segment a; represents the traffic flow of travel mode i between OD pair w on road segment a; represents the traffic flow between OD pairs w on road segment a; in addition, let the vector as well as Represents the vector form of traffic flow; OUT(n) represents the set of directed road segments starting from node n, and IN(·) represents the set of directed road segments ending at node n; where N\{o w ,d w} represents a set of road nodes that are not starting or ending points.

[0100] ② Ride-sharing matching constraints

[0101]

[0102] Step 3: Calculate the travel costs of travelers in the target city’s road network.

[0103] Through preliminary investigation, the values ​​of various parameters in the target city road network are shown in Table 2.

[0104] Table 2 Parameters of target city road network

[0105]

[0106] The generalized travel cost in the target road network is calculated as follows:

[0107]

[0108] The specific calculation function is:

[0109]

[0110] Where t a,0 is the free flow travel time of section a, which can be obtained by the floating car method; a is the capacity of section a, which can be obtained by visiting the design unit or consulting relevant information.

[0111] 2) Based on the equilibrium principle of carpooling users, an equivalent implicit variational inequality model is established.

[0112] Find λ * ∈Ω λ Satisfy the formula:

[0113] (λ-λ * ) T [Bx(λ * )]≥0 or

[0114] Where Ω λ =R |A|·|RD| , represents the feasible set of λ, that is, the set of Lagrange multipliers that satisfy the traffic flow constraints and the ridesharing matching constraints.

[0115] 3) The equivalent implicit variational inequality model is solved by the PM-PCA algorithm to predict the traffic flow of the multi-hop ride-hailing travel network.

[0116] The PM-PCA algorithm is used to solve the variational inequality. Table 3 gives m i =0.05, the traffic distribution of the target city road network under the equilibrium state of carpooling users.

[0117] Table 3 Traffic distribution of target city road network

[0118]

[0119] Table 4 gives m i =0.05. The distribution of travel choices of the target city road network under the equilibrium state of carpooling users.

[0120] Table 4 Distribution of travel choices

[0121]

[0122] From Table 4, we can see that about 3.2% of travelers become passengers. In other words, due to the emergence of multi-hop ride-sharing services, the number of vehicles has decreased by 3.2%.

[0123] Table 5 gives the speedup rate and time saving rate under different numbers of processor cores, respectively, to test the parallel computing performance of the PM-PCA algorithm. The high-performance computing (HPC) system used for this test consists of Intel Xeon Gold 6240 central processing units (CPUs), each with 36 processors and 192GB of 2933MHz DDR4 RAM with a clock speed of 2.60GHz.

[0124] Table 5 Algorithm efficiency test on target road network

[0125]

[0126] It can be seen from Table 5 that with the increase of the number of processor cores, the PM-PCA algorithm can effectively shorten the calculation time and greatly improve the acceleration rate and time saving rate.

[0127] The present invention also provides a traffic flow prediction system for a multi-hop ride-sharing network of online car-hailing, including the following modules:

[0128] An input module, configured to input the topology of the target city road network, travel modes and travel demands;

[0129] A first establishment module is configured to establish traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel according to the topological structure, travel mode and travel demand of the target city road network;

[0130] A first calculation module is configured to calculate the travel cost of travelers in the target city road network through traffic flow constraints and ride-sharing matching constraints;

[0131] The second building module is configured to complete the building of an equivalent implicit variational inequality model by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and travel costs of travelers;

[0132] The second computing module is configured to solve the equivalent implicit variational inequality model in combination with the parallel cost averaging projection algorithm to complete the traffic flow prediction of the online car-hailing multi-hop ride-sharing travel network.

[0133] The present invention also provides a traffic flow prediction device for an online car-hailing multi-hop ride-sharing travel network, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned traffic flow prediction method for the online car-hailing multi-hop ride-sharing travel network are implemented.

[0134] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned method for environmental impact assessment of online ride-hailing travel.

[0135] The embodiments of the present application can be provided as methods or computer program products. Therefore, the present application can adopt the form of complete hardware embodiment, complete software embodiment, or the embodiment in combination with software and hardware. Moreover, the present application can adopt the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0136] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0137] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0138] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0139] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

Claims

1. A traffic flow prediction method for a multi-hop ride-sharing network of online car-hailing, characterized in that: include: Input the topology, travel mode and travel demand of the target city’s road network; According to the topological structure, travel mode and travel demand of the target city road network, traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel are established; The travel costs of travelers in the target city road network are calculated through traffic flow constraints and ride-sharing matching constraints; The equivalent implicit variational inequality model is established by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and the travel costs of travelers; Combined with the parallel cost average projection algorithm, the equivalent implicit variational inequality model is solved to complete the traffic flow prediction of the multi-hop ride-sharing network of online car-hailing; The topological structure of the target urban road network is (N, A), where N is the set of OD points in the urban road traffic network, and A is the set of directed road segments connecting the OD points in the urban road traffic network; The travel demand between different OD pairs is q w , where w represents an OD pair, W represents the set of OD pairs in the urban road traffic network, w∈W, and the starting and ending points of OD pair w are represented by o w ∈O and d w ∈D means that O is the set of starting points, D is the set of end points, and both O and D are subsets of N; According to the characteristics of multi-hop ridesharing, the travel modes are divided into three categories: SD represents the set of drivers who drive alone without matching passengers, RD represents the set of drivers who match passengers and participate in ridesharing, and R represents the set of passengers who match drivers and participate in ridesharing. Let I represent the set of three types of travel modes in the urban road traffic network, I = SD∪RD∪R; The traffic flow constraint is expressed as: Where x a,i represents the traffic flow of travel mode i on road segment a; represents the traffic flow of travel mode i between OD pairs w on road segment a; represents the traffic flow between OD pairs w on road segment a; in addition, let the vector as well as The vector form of traffic flow; The starting and ending points in the urban road network are: Where OUT(n) represents the set of directed segments starting from node n, and IN(·) represents the set of directed segments ending at node n; For road nodes that are not starting or ending points in the urban road network, there is the following flow relationship: In the formula, N\{o w ,d w } represents the road node set that is not the starting point or the ending point; The ride-sharing matching constraint is expressed as: Where Γ r (i) represents mapping the online car-hailing driver i to the corresponding online car-hailing passenger, that is, Γ r :RD→R,M i Indicates the number of passengers in the car of online car-hailing driver i; The ride-sharing matching constraints are rewritten into a compact structure based on the vector form of traffic flow: Bx=0 or ∑ w B w X w =0 Where B=[B1B2…B w∈W …B |W| ]; is a block matrix of |A|×|A|; is a |RD|×(2|RD|+1) matrix; The calculation of the travel cost of travelers in the target city road network includes: According to the characteristics of online car-hailing ride-sharing travel, the travel cost of travelers includes time cost, inconvenience cost and ride-sharing fee, which can be expressed as: The road flow vector represents the travel cost of a traveler of mode i between OD pairs w on road segment a, which is a function of x; represents the travel time of travelers on road section a, function t a (·) is about Strictly monotonically increasing empirical function; α i represents the time value of travel mode i, β i represents the inconvenience coefficient of travel mode i; B a,i represents the basic fee collected / paid by online car-hailing driver i / online car-hailing passenger i on road section a; m i represents the price fluctuation parameter of travel mode i; According to the characteristics of the supply and demand balance of online car-hailing ride-sharing and the complementary relaxation of linear programming, the car-sharing subsidy and premium are combined with the travel cost function, and the generalized travel cost function of the traveler is expressed as: In the formula represents the generalized travel cost of a traveler of mode i between OD pairs w on road segment a; a,i is the Lagrange multiplier corresponding to the carpooling matching constraint, whose value is equal to the value of the premium and the subsidy; Γ rd (i) represents mapping the online car-hailing passenger i to the corresponding online car-hailing driver, that is, Γ rd :R→RD; let vector and denote the generalized travel cost and Lagrange multiplier, respectively; The establishment of the equivalent implicit variational inequality model includes: Establish a variational inequality model [RUE'], the Lagrange multiplier λ corresponds to the multiplication matching constraint. When the value of λ is given, establish a variational inequality model to express the multiplication equilibrium problem and find x * ∈Ω′ RUE Satisfy the formula: Where λ * is a parameter of a given value, equal to the Lagrange multiplier in the multiplication equilibrium state; Ω′ RUE represents the feasible set of x, that is, the set of road flow that satisfies the traffic flow constraints and the ride-sharing matching constraints; Establish a fixed point model [RUE”], and establish a fixed point model [RUE”] equivalent to the variational inequality model [RUE']: Where ψ(·) represents the network load mapping from the generalized travel cost set to the traffic flow set; Establish an implicit variational inequality model. The solution x(λ) of the carpooling equilibrium problem is an implicit function of λ, whose value is determined by solving [RUE'] or [RUE”] to find λ * ∈Ω λ Satisfy the formula: Where Ω λ =R |A|·|RD| , represents the feasible set of λ, which is the set of Lagrange multipliers that satisfy the traffic flow constraints and the ride-sharing matching constraints; The steps of solving the equivalent implicit variational inequality model include: Initialization step, set the accuracy ε>0, β (0) > 0, k = l = 0, given the initial Lagrange multiplier λ (0) And the initial flow solution x (0) ; Find the direction step, update the vector Then, let each processor execute a Dijkstra algorithm to find the shortest path from a starting point to all destinations, transforming the travel demand q w Loaded on the shortest path between OD pair w, generating auxiliary flow solution y (k) and assisted travel time Update the cost vector step, the vector To iterate: Check the cost accuracy step, if Enter the step of updating the inconvenience coefficient and the Lagrange multiplier; otherwise k=k+1, return to the step of finding the direction; Update the inconvenience coefficient and Lagrange multiplier step, for the vector β (l) and λ (l+1) To update: l (l+1) =λ (l) -b (l) (Bx (k) ) Check the accuracy of the Lagrange multipliers step, if Then terminate the calculation and output λ (l) and x (k) ; Otherwise, let l=l+1, k=0, and return to the search direction step.

2. A prediction system based on the traffic flow prediction method of the online car-hailing multi-hop ride-sharing travel network according to claim 1, characterized in that: Includes the following modules: An input module, configured to input the topology of the target city road network, travel modes and travel demands; A first establishment module is configured to establish traffic flow constraints and ridesharing matching constraints that meet the characteristics of multi-hop ridesharing travel according to the topological structure, travel mode and travel demand of the target city road network; A first calculation module is configured to calculate the travel cost of travelers in the target city road network through traffic flow constraints and ride-sharing matching constraints; The second building module is configured to complete the building of an equivalent implicit variational inequality model by combining the topological structure of the target city's road network, travel mode, travel demand, traffic flow constraints, ride-sharing matching constraints, and travel costs of travelers; The second computing module is configured to solve the equivalent implicit variational inequality model in combination with the parallel cost averaging projection algorithm to complete the traffic flow prediction of the online car-hailing multi-hop ride-sharing travel network.

3. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods of claim 1 .

4. A computing device, characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods according to claim 1.

Images