Limited rationality ride-sharing travel flow prediction method and system

Abstract

The invention discloses a bounded rationality ride-sharing travel flow prediction method and system, and relates to the technical field of flow prediction, and the method comprises the steps: firstly constructing a network topology structure of a ride-sharing network, traffic flow constraint and a generalized travel cost function, and supposing that travelers in the ride-sharing network have bounded rationality; in order to predict a path flow set and an acceptable travel scheme set of an urban traffic network in a limited rational ride-sharing user equilibrium state, a mathematical programming model with equilibrium constraint is established and converted into a mixed integer programming model; and commercial software is applied, and parallel computing is combined for solving. Compared with a past traffic flow prediction method, the method of the invention considers an emerging trip mode of ride-sharing trip and a bounded rationality trip decision process at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic prediction, and in particular to a limited rational carpooling traffic prediction method and system. Background Art

[0002] Carpooling is based on the driver's own travel needs. It shares the travel costs by picking up passengers with the same or similar travel routes, thereby reducing energy consumption and improving road congestion. With the development of satellite navigation, mobile payment and other technologies, a mature business model of carpooling has been formed both at home and abroad. In real life, when travelers choose travel plans based on cost, due to the influence of factors such as travel experience and travel habits, they do not aim to optimize travel utility, but to ensure that the difference between their travel utility and the optimal travel utility is within an acceptable range. Based on a multi-modal urban transportation network that includes carpooling, considering the limited rationality of travelers when making travel plan decisions, predicting the flow distribution of the transportation network in equilibrium can not only help travelers make reasonable decisions and improve personal travel satisfaction; it can also help third-party platforms understand the behavioral selection rules of travelers, optimize travel services and business strategies; it can also help relevant departments formulate reasonable and effective planning plans and management policies, and build a green and efficient urban transportation system.

[0003] A traffic prediction method for urban shared travel (application number CN202110906993.9) comprehensively considers the generalized travel cost from multiple angles, including the travel time perceived by the travelers, the loss of in-car comfort, and the online operation mechanism of the ride-sharing service, and distributes the traffic flow based on the logit travel selection behavior and the equilibrium principle of random P2P ride-sharing users, to obtain the path traffic flow of the urban multi-modal complex network. A method for analyzing the charging demand of electric vehicles considering the limited rationality of users (application number CN201911193203.6) establishes a travel choice probability model on the activity-travel chain of travelers, analyzes the transfer law of travelers on the activity-travel chain and the charging law of electric vehicles on the activity-travel chain, and combines the number of daily trips of travelers to obtain the daily charging demand of electric vehicles. Although both existing inventions take into account the errors in travelers' evaluation of travel utility, they believe that travelers choose paths and travel modes with the goal of optimizing utility, ignoring the behavioral inertia of travelers as limited rational subjects, that is, travelers may implement satisfactory rather than optimal path and travel mode selection patterns. Summary of the invention

[0004] In order to solve the deficiencies mentioned in the above background technology, the purpose of the present invention is to provide a limited rational method and system for predicting carpooling travel flow.

[0005] In a first aspect, the purpose of the present invention can be achieved by the following technical solution: a bounded rational carpooling flow prediction method, the method comprising the following steps:

[0006] Obtain ridesharing travel related data, generate a ridesharing network based on the ridesharing travel related data, and obtain flow conservation constraints and ridesharing matching constraints, as well as a generalized travel cost function based on the ridesharing network;

[0007] Generate a set of path flows under the equilibrium state of bounded rational carpooling users based on the carpooling network, as well as a set of acceptable travel plans that can carry the flow;

[0008] Based on the mathematical programming model with equilibrium constraints, the critical point sequence is solved, and the set of path flows and the set of acceptable travel plans that can carry the flow are determined by using the critical point sequence under the equilibrium state of finite rational carpooling users with a given indifference interval vector.

[0009] The mathematical programming model with equilibrium constraints is optimized to obtain a mixed integer programming model, and the mixed integer programming model is solved by applying an algorithm to obtain the prediction result of carpooling travel flow.

[0010] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: the ride-sharing travel-related data includes the topological structure and travel demand of the urban transportation network of the ride-sharing travel mode.

[0011] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of generating a ridesharing network based on ridesharing travel related data:

[0012] Construct a strongly connected urban transportation network G = (N, A) that includes ridesharing as the ridesharing network G, where N and A represent the node set and the road segment set in the ridesharing network G, respectively. Let W represent the set of OD pairs in the ridesharing network, w represents a certain OD pair, w∈W, and let P represent the set of all paths in the ridesharing network, P w Denotes the path set of OD to w, and P = ∪ w∈W P w ; p represents a path, and p∈P w , let I represent the set of travel modes in the ridesharing network, i represents a certain travel mode, i∈I; the main travel modes in the ridesharing network include empty car drivers, carpooling drivers and carpooling passengers, represented by the sets SD, RD and R respectively, with I=SD∪RD∪R; carpooling drivers and carpooling passengers are the drivers and passengers of the carpooling service, and empty car drivers do not participate in the carpooling service and drive alone; let Indicates the maximum passenger capacity of a vehicle, N i represents the number of carpooling passengers carried by carpooling driver i, Then i=1 represents an empty vehicle driver, and the empty vehicle driver set SD={i=1}; i=1+N i Indicates that it is equipped with N i Rideshare drivers of rideshare passengers, rideshare driver collection Indicates the passengers being carried, the set of passengers being carried

[0013] In combination with the first aspect, in some implementations of the first aspect, the method further includes: a process of obtaining a flow conservation constraint and a ridesharing matching constraint based on the ridesharing network:

[0014] Let a∈A represent a certain section of the ridesharing network, represents the flow of travel mode i on path p connecting OD pair w, x a,i represents the flow of travel mode i on road segment a, represents the path-section association coefficient, then the path flow and road traffic x a,i The relationship between them is:

[0015]

[0016] Let x a represents the traffic volume of road section a, t a (x a ) represents the travel time function of section a, Denote the travel time of path p between OD pairs w, we have:

[0017]

[0018] Let QW represent the travel demand of OD between w∈W, and construct the flow conservation constraint in the ridesharing network G as:

[0019]

[0020] Since there are limited seats available for sharing in the car, there is a certain quantitative relationship between the carpooling drivers and the carpooling passengers. Based on the quantitative relationship, the carpooling matching constraint is established as follows:

[0021]

[0022] In the formula, Γ r (i) represents the mapping relationship from carpooling driver i∈RD to carpooling passenger i′∈R, i.e., Γ r :RD→R or N i is a positive integer, indicating the number of seats shared by the ridesharing driver i∈RD.

[0023] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of generating a generalized travel cost function based on the ridesharing network:

[0024] make represents the path flow vector in the ridesharing network G, The generalized travel cost function of travel mode i on path p connecting OD pair w is:

[0025]

[0026] In the formula, They represent the path length of path p between OD pair w respectively; represents the Lagrangian multiplier corresponding to the travel mode i∈RD on the path p between OD pairs w; Γ rd (i) represents the mapping relationship between carpooling passengers i∈R and carpooling drivers i′∈RD, Γ rd :R→RD or is Γ r The inverse mapping of (i), parameter ρ i represents the time value of travel mode i∈I; γ i represents the inconvenience coefficient of travel mode i∈RD∪R; b i , With r i They represent the coefficients for calculating the time fee, mileage fee and floating pricing of travel mode i∈RD∪R respectively; c t Represents the cost of running a vehicle, including fuel, charging and parking costs.

[0027] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of generating a set of path flows in a bounded rational ridesharing user equilibrium state and a set of acceptable travel solutions capable of carrying the flows by using the generalized travel cost function of the ridesharing network:

[0028] In the ride-sharing network G, travelers choose both routes and travel modes. A route p∈P w With a travel mode i∈I, it forms a travel plan (p, i) between OD pair w. w , when a travel plan (p, i) w The generalized travel cost Generalized travel costs compared to the optimal travel solution The difference does not exceed a certain threshold called the indifference interval; let ε w represents the indifference interval of OD to w, and takes the value ε w ≥0; ε=(ε w , w∈W) Trepresents the indifference interval vector of the ridesharing network G; represents the feasible domain of f. When bounded rational travelers travel in the ride-sharing network G, the traffic flow in the network will form the following bounded rational ride-sharing user equilibrium state:

[0029]

[0030] The above formula shows that under the BR-RUE state, travelers between OD pairs w can choose the generalized travel cost Any travel plan (p, i) w , the path flow f following the BR-RUE state is called the BR-RUE path flow pattern, symbolically represented by f ε , at this time, the travel plans in the ridesharing network G can be divided into three states: acceptable, 0-acceptable and unacceptable, which meet the following conditions respectively:

[0031] Acceptable:

[0032] 0 - Acceptable:

[0033] Unacceptable:

[0034] When ε is given, there are usually more than one ε-BR-RUE path flow pattern fε in the ridesharing network G, forming a BR-RUE flow set, which is symbolically represented as The acceptable travel plans corresponding to all fε in also form a set, which is the ε-acceptable travel plan set, and the symbol is expressed as

[0035] For a ridesharing network G with multiple OD pairs, the path flow vector is decomposed into f = (f 1 , f 2 , ..., f w , ..., f |W| ), weight represents the path flow between OD pair w; define ε w -BR-RUE flow set is arbitrary The corresponding component f w ,Right now So that the ε-BR-RUE flow set can be expressed as The acceptable travel plans between OD and w form a set, ε w -A set of acceptable travel options, represented by Right now

[0036] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of solving the critical point sequence based on the mathematical programming model with equilibrium constraints:

[0037] Given any two indifference interval vectors and If for some OD pair w∈W there exists Then the set relation All are established, and the set In the indifference interval ε w When it grows to a certain value, it expands simultaneously;

[0038] With the indifference interval ε w The value of increases from 0 to +∞, collecting all critical points in the process and combining them into an ascending sequence, mathematically expressed as Where k = 1, 2, ..., K, there are This monotone positive sequence divides the non-negative real number interval [0, +∞) into K+1 subintervals, namely The sequence ends are and When the indifference interval ε w Take a critical point Right now When w - A collection of acceptable travel options is represented as With ε w from Grow to New, not part of the previous set The travel plan will be added to the collection In represents a newly added travel plan between OD and w;

[0039] When ε is given, the ε of any OD to w in the ridesharing network G can be determined by using the critical point sequence. w - A collection of acceptable travel options So as to obtain a set of acceptable travel solutions Then the travel demand Q w Assigned to these acceptable travel options, we get ε w -BR-RUE traffic collection The relationship between each OD and w in the comprehensive ride-sharing network G Get BR-RUE traffic collection

[0040] Construct a mathematical programming model with equilibrium constraints, namely the MPEC model, to solve the critical point sequence

[0041]

[0042] In the above formula, the objective function solves the kth critical point of OD for w is a strict inequality constraint, is a complementary constraint, is a parameter with a value of {0, 1}. When the travel plan (p, i) is acceptable w Not included in the collection Inside, on the contrary make represents the maximum generalized travel cost of all acceptable travel options between OD pair w, is the acceptable travel option (p, i) w The variable represents the difference between its generalized travel cost and the maximum acceptable generalized travel cost, that is,

[0043] In combination with the first aspect, in some implementations of the first aspect, the method further includes: a process of determining a set of path flows and a set of acceptable travel solutions capable of carrying the flows in a finite rational carpooling user equilibrium state for a given indifference interval vector using a critical point sequence:

[0044] Solving the MPEC model can obtain the critical point sequence of any OD pair w∈W , where k = 1, 2, ..., K, and the corresponding ε of each critical point is obtained w - A collection of acceptable travel options Then the travel demand Q w Assign to these acceptable travel options to construct ε w -BR-RUE traffic collection For OD pair w, when the indifference interval ε w Gradually increase to One or more new acceptable travel options will be added to ε w - A collection of acceptable travel options New path traffic patterns are also added to ε w -BR-RUE traffic collection and in these new path traffic patterns, the newly added set The acceptable travel options all carry a certain amount of traffic, Indicates when ε w-BR-RUE traffic solution set, Respectively represent ε w = 0 when ε w -Acceptable travel plan set, BR-RUE flow set, decomposition set based on critical points There is a collection relationship The complement The mathematical expression is:

[0045]

[0046] In the formula, represents the path flow vector f of OD pair w w The feasible domain of w = 0, then k = 0, and the set is defined It is expressed as:

[0047]

[0048] when When w - A collection of acceptable solutions BR-RUE traffic collection for:

[0049]

[0050] The set of acceptable travel solutions for the ride-sharing network G BR-RUE traffic collection

[0051] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of applying an algorithm to the mixed integer programming model to obtain a prediction result of the ride-sharing travel flow:

[0052] Mixed integer programming, or MIP model, is as follows:

[0053]

[0054] In the above formula, the auxiliary variable is a 0-1 integer variable; parameter M is a positive number; parameters m1, m2∈(0, 1] and have very small values;

[0055] If the segment travel time function t a (x a ) is a linear function, and the MIP model is a mixed integer linear programming model, which is solved by CPLEX or Gurobi software; if the section travel time function t a (x a) is a nonlinear function, and the MIP model is a mixed integer nonlinear programming model, which is solved by Gurobi software;

[0056] Combining multi-process parallel computing technology with network decomposition methods, solving complex ridesharing networks with multiple OD pairs The critical point sequence

[0057] Apply the network decomposition method to decompose the ridesharing network G into |W| subnetworks with single OD pairs; create an asynchronous process pool and assign processes to process the subnetworks, let represents the kth critical point of OD on w The corresponding 0-1 parameter vector is solved in turn within the sub-network corresponding to OD for w∈W. The calculation process is as follows:

[0058] Initialize k = 1; for make v w (k) = 0;

[0059] Solving the MIP model yields Acceptable travel options If the travel plan Set the corresponding parameters

[0060] Solve for the kth critical point of OD on w and produce Acceptable travel options If the travel plan Set the corresponding parameters

[0061] If v w(k) =1, the calculation of OD pair w is completed, update w=w+1, k=1, and then perform the calculation of the next OD pair, otherwise, update k=k+1, v w(k) =v w(k-1) , solve the next critical point for OD with respect to w.

[0062] In a second aspect, in order to achieve the above-mentioned object, the present invention discloses a bounded rational carpooling flow prediction system, comprising:

[0063] A data processing module is used to obtain ride-sharing travel related data, generate a ride-sharing network based on the ride-sharing travel related data, and obtain flow conservation constraints and ride-sharing matching constraints, as well as a generalized travel cost function based on the ride-sharing network;

[0064] A set generation module is used to generate a set of path flows under the equilibrium state of bounded rational carpooling users, as well as a set of acceptable travel plans that can carry the flow;

[0065] The set optimization module is used to solve the critical point sequence based on the mathematical programming model with equilibrium constraints, and use the critical point sequence to determine the path flow set and the acceptable travel plan set that can carry the flow in the equilibrium state of the finite rational carpooling users with a given indifference interval vector;

[0066] The model solving module is used to optimize the mathematical programming model with equilibrium constraints to obtain a mixed integer programming model, and apply the algorithm to solve the mixed integer programming model to obtain the prediction result of the carpooling travel flow.

[0067] Beneficial effects of the present invention:

[0068] The present invention first constructs the network topology, traffic flow constraints and generalized travel cost function of the ride-sharing network, and assumes that travelers in the ride-sharing network are all of limited rationality, and pursue acceptable rather than optimal travel plans when making travel decisions. In order to predict the path flow set and acceptable travel plan set of the urban transportation network under the equilibrium state of limited rational ride-sharing users, the present invention establishes a mathematical programming model with equilibrium constraints and converts it into a mixed integer programming model, and solves it using commercial software and parallel computing. Compared with previous traffic flow prediction methods, the present invention simultaneously considers the emerging travel mode of ride-sharing and the limited rational travel decision-making process. The present invention can help travelers make reasonable travel plans and improve personal travel satisfaction; help travel service providers understand the behavioral selection rules of travelers and possible individual differences, and carry out targeted travel service optimization and business strategy adjustments in combination with user needs; help government departments formulate effective traffic planning plans and traffic management policies, make full use of road resources, alleviate traffic congestion, and build an efficient and green urban transportation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0070] Figure 1 It is a schematic flow chart of the method of the present invention;

[0071] Figure 2 is a schematic diagram of the topological structure and travel demand of an example network of the present invention;

[0072] Figure 3is a schematic diagram of a set of acceptable travel solutions for an example network of the present invention;

[0073] Figure 4 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments 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] Embodiment 1:

[0076] like Figure 1 As shown, a bounded rational carpooling flow prediction method comprises the following steps:

[0077] S101: Acquire ridesharing travel related data, generate a ridesharing network based on the ridesharing travel related data, and obtain flow conservation constraints and ridesharing matching constraints, as well as a generalized travel cost function based on the ridesharing network;

[0078] Among them, a certain urban transportation network including ride-sharing is simplified to form Figure 2 The network structure of the carpooling network G has 3 nodes and 4 road segments, namely N = {1, 2, 3}, A = {(1), (2), (3), (4)}; there are 2 OD pairs (1, 2) and (1, 3), corresponding to indexes w = 1 and w = 2, respectively, and their travel demands are Q 1 =2, Q 2 =4, that is, W = {(1, 2), (1, 3)}. The four road segments form six paths: (1) → (3), (1) → (4), (2) → (3), (2) → (4), (1), and (2). The first four paths connect OD pair w = 1, and the last two paths connect OD pair w = 2, that is, P 1 ={(1)→(3), (1)→(4), (2)→(3), (2)→(4)}, P 2 ={(1),(2)}. The length of each road segment is 100. The road segment travel time function t a (x a ) are respectively And t4(x4)=50+x4.

[0079] Assume that each ride-sharing driver can only carry two ride-sharing passengers at the same time, that is There are 5 travel modes in the ridesharing network, namely |I|=5, i=1∈SD, i=2,3∈RD, i=4,5∈R.

[0080] Let a∈A represent a certain section of the ridesharing network, represents the flow of travel mode i on path p connecting OD pair w, x a，i represents the flow of travel mode i on road segment a, represents the path-section association coefficient, then the path flow and road traffic x a，i The relationship between them is:

[0081]

[0082] Let x a represents the traffic volume of road section a, t a (x a ) represents the travel time function of section a, Denote the travel time of path p between OD pairs w, we have:

[0083]

[0084] Let Q w Represents the travel demand of OD between w∈W, and constructs the flow conservation constraint in the ridesharing network G as:

[0085]

[0086] Since there are limited seats available for sharing in the car, there is a certain quantitative relationship between the carpooling drivers and the carpooling passengers. The carpooling matching constraints are established based on this quantitative relationship, which are as follows:

[0087]

[0088] In the formula, Γ r (i) represents the mapping relationship from carpooling driver i∈RD to carpooling passenger i′∈R, i.e., Γ r :RD→R or N i is a positive integer, indicating the number of seats shared by the ridesharing driver i∈RD.

[0089] make represents the path flow vector in the ridesharing network G, represents the generalized travel cost function of travel mode i on path p connecting OD pair w, specifically:

[0090]

[0091] In the above formula, They represent the path length of path p between OD pair w respectively; represents the Lagrangian multiplier corresponding to the travel mode i∈RD on the path p between OD pairs w; Γrd (i) represents the mapping relationship between carpooling passengers i∈R and carpooling drivers i′∈RD, that is, Γ rd :R→RD or is Γ r The inverse mapping of (i). Parameter ρ i represents the time value of travel mode i∈I; γ i represents the inconvenience coefficient of travel mode i∈RD∪R; b i ,φ i With r i They represent the coefficients for calculating the time fee, mileage fee and floating pricing of travel mode i∈RD∪R respectively; c t Represents the vehicle operating costs, such as fuel, charging, parking fees, etc. Generalized travel cost function The parameter values ​​are shown in Table 1.

[0092] Table 1 Parameter settings of numerical experiments

[0093]

[0094] S102: Generate a set of path flows under a bounded rational carpooling user equilibrium state and a set of acceptable travel plans that can carry the flows using a generalized travel cost function of the carpooling network;

[0095] In the ride-sharing network G, travelers choose both routes and travel modes. A route p∈P w With a travel mode i∈I, it forms a travel plan (p, i) between OD pair w. w When a travel plan (p, i) w The generalized travel cost Generalized travel costs compared to the optimal travel solution The difference (i.e. As long as the travel plan does not exceed a certain threshold called the “indifference interval”, the travel plan may be chosen by travelers with limited rationality. w represents the indifference interval of OD to w, and takes the value ε w ≥0; E=(ε w , w∈W) T represents the indifference interval vector of the ridesharing network G; represents the feasible domain of f. When bounded rational travelers travel in the ridesharing network G, the traffic flow in the network will form the following bounded rational ridesharing user equilibrium (BR-RUE) state:

[0096]

[0097] The above formula shows that under the BR-RUE state, travelers between OD pairs w can choose the generalized travel cost Any travel plan (p, i) w The path flow f following the BR-RUE state is called the “BR-RUE path flow pattern”, further symbolized as f ε , at this time, the travel plans in the ridesharing network G can be divided into three states: acceptable, 0-acceptable and unacceptable, which meet the following conditions respectively:

[0098] Acceptable:

[0099] 0 - Acceptable:

[0100] Unacceptable:

[0101] When ε is given, there are usually more than one ε-BR-RUE path flow pattern fε in the ridesharing network G, forming a “BR-RUE flow set” Symbols are represented by All f ε The corresponding acceptable travel plans also form a set, defined as the "ε-acceptable travel plan set", which is represented by

[0102] Furthermore, for the ridesharing network G with multiple OD pairs, the path flow vector is decomposed into f = (f 1 , f 2 , …, f w , …f |W| ), weight represents the path flow between OD pair w. Define “ε w -BR-RUE traffic set" is any The corresponding component f w ,Right now This makes E-BR-RUE traffic aggregation can be expressed as The acceptable travel plans between OD pairs w form a set, defined as “ε w -Acceptable travel plan set", symbolized by Right now

[0103] S103: Solving the critical point sequence based on the mathematical programming model with equilibrium constraints, and using the critical point sequence to determine the path flow set and the acceptable travel plan set that can carry the flow in the equilibrium state of the finite rational carpooling user with a given indifference interval vector;

[0104] It can be shown that given any two indifference interval vectors and If for some OD pair w∈W there exists Then the set relation All are established, and the set In the indifference interval ε w They grow and expand simultaneously when they reach a certain value, which are defined as "critical points".

[0105] With the indifference interval ε w The value of increases from 0 to +∞, collecting all critical points in the process and combining them into an ascending sequence, mathematically expressed as have This monotone positive sequence divides the non-negative real number interval [0, +∞) into K+1 subintervals, namely The sequence ends are When the indifference interval ε w Take a critical point Right now When w - A collection of acceptable travel options is represented as With ε w from Grow to Some new ones that did not belong to the previous set The travel plan will be added to the collection In Represents a newly added travel plan for OD between w.

[0106] When ε is given, the ε of any OD to w in the ridesharing network G can be determined by using the critical point sequence. w - A collection of acceptable travel options So as to obtain a set of acceptable travel solutions Then the travel demand Q w Assigning to these acceptable travel options, we can get ε w -BR-RUE traffic collection The relationship between each OD and w in the comprehensive ride-sharing network G Get BR-RUE traffic collection

[0107] Construct a mathematical programming with equilibrium constraints (MPEC) model to solve the critical point sequence

[0108]

[0109]

[0110] In the above formula, the objective function solves the kth critical point of OD for w The third and sixth constraints are strict inequality constraints, and the fifth constraint is a complementarity constraint. is a parameter with a value of {0, 1}. If the travel plan (p, i) is acceptable w Not included in the collection Inside, on the contrary make represents all acceptable travel plans between OD pairs w (i.e. ), also known as the “maximum acceptable generalized travel cost”. is the acceptable travel option (p, i) w (Right now ) variable, representing its generalized travel cost and the maximum acceptable generalized travel cost The gap, that is

[0111] Solving the MPEC model can not only obtain the critical point sequence of any OD pair w∈W (k=1,2,…,K), and the ε corresponding to each critical point can also be obtained w - A collection of acceptable travel options Then the travel demand Q w Assign to these acceptable travel options to construct ε w -BR-RUE traffic collection For OD pair w, when the indifference interval ε w Gradually increase to One or more new acceptable travel options will be added to ε w - A collection of acceptable travel options At the same time, new path traffic patterns are also added to ε w -BR-RUE traffic collection and in these new path traffic patterns, the newly added set The acceptable travel options all carry a certain amount of traffic. Indicates when ε w -BR-RUE traffic solution set, Respectively represent ε w = 0 when ε w -Acceptable travel plan set, BR-RUE flow set. Decompose the set according to the critical point There is a collection relationship The complement The mathematical expression is:

[0112]

[0113] In the formula, represents the path flow vector f of OD pair w w The feasible domain of w = 0, then k = 0, and the set is defined It is expressed as:

[0114]

[0115] Therefore, when When w - A collection of acceptable solutions Traffic Collection for:

[0116]

[0117] The set of acceptable travel solutions for the ride-sharing network G BR-RUE traffic collection

[0118] S104: Optimize the pre-established mathematical programming model with equilibrium constraints based on the path flow set of finite rational carpooling users in equilibrium state with a given indifference interval vector and the set of acceptable travel plans that can carry the flow, obtain a mixed integer programming model, and apply the algorithm to solve the mixed integer programming model to obtain the carpooling travel flow prediction result.

[0119] Since the complementary constraints and inequality constraints greatly increase the difficulty of solving the MPEC model, the complementary constraints are further transformed into linear constraints, and the strict inequality constraints are relaxed into inequality constraints, so that the original MPEC model is transformed into the following mixed integer programming (MIP) model:

[0120]

[0121]

[0122] In the above formula, the auxiliary variable It is a 0-1 integer variable; parameter M = 999999, parameter m1 = m2 = 10 -5 Since the segment travel time function t in the example network a (x a ) is a linear function, and the MIP model is a mixed integer linear programming model, which can be solved by CPLEX or Gurobi software.

[0123] To improve computational efficiency, we can further combine multi-process parallel computing technology with network decomposition methods to solve the carpooling network of multiple OD pairs. The critical point sequence

[0124] First, the network decomposition method is applied to decompose the ridesharing network G into |W| sub-networks with single OD pairs; secondly, an asynchronous process pool is created and an appropriate number of processes are assigned to process these sub-networks. represents the kth critical point of OD on w The corresponding 0-1 parameter vector. In the sub-network corresponding to OD for w∈W, solve its critical points one by one The specific calculation process is as follows:

[0125] Step 0: Initialize k = 1; for Vw∈W, let v w(k) =0.

[0126] Step 1: Solve the MIP model to produce Acceptable travel options If the travel plan Set the corresponding parameters

[0127] Step 2: Find the kth critical point of OD for w and produce Acceptable travel options If the travel plan Set the corresponding parameters

[0128] Step 3 If v w(k) =1, the calculation of OD pair w is completed, update w=w+1, k=1, and return to step 1 to perform the calculation of the next OD pair. Otherwise, update k=k+1, v w(k) =v w(k-1) , repeat step 2 to solve the next critical point for OD versus w.

[0129] Solve the MIP model using the CPLEX tool to obtain the critical point sequence of the two OD pairs in the ridesharing network G They are:

[0130]

[0131] OD for w=1 Acceptable travel options as follows:

[0132]

[0133] OD for w = 2 Acceptable travel options as follows:

[0134]

[0135]

[0136] application Figure 3 The set of acceptable travel solutions for the example network The expression is more intuitive. The numbers in the grid indicate that the travel plan is added to the set. The size of k in the middle.

[0137] After obtaining the set of acceptable travel solutions, the travel demand is allocated according to the flow conservation constraint and the ridesharing matching constraint to obtain the BR-RUE path flow set.

[0138] Embodiment 2: In the second aspect, as Figure 4 As shown, in order to achieve the above-mentioned purpose, the present invention discloses a bounded rational carpooling flow prediction system, comprising:

[0139] The data processing module 11 is used to obtain ride-sharing travel related data, generate a ride-sharing network based on the ride-sharing travel related data, and obtain flow conservation constraints and ride-sharing matching constraints, as well as a generalized travel cost function based on the ride-sharing network;

[0140] A set generation module 12 is used to generate a set of path flows under a bounded rational carpooling user equilibrium state, and a set of acceptable travel plans that can carry the flow;

[0141] The set optimization module 13 is used to solve the critical point sequence based on the mathematical programming model with equilibrium constraints, and use the critical point sequence to determine the path flow set of the finite rational carpooling user in the equilibrium state of a given indifference interval vector and the set of acceptable travel plans that can carry the flow;

[0142] The model solving module 14 is used to optimize the mathematical programming model with equilibrium constraints to obtain a mixed integer programming model, and apply an algorithm to solve the mixed integer programming model to obtain a carpooling travel flow prediction result.

[0143] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0144] It needs to be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, on which a computer program is stored, and the computer program is executed by a processor to execute the above method. The storage medium can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0145] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0146] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure may have various changes and improvements, and these changes and improvements fall within the scope of the present disclosure to be protected.

Claims

1. A bounded rational method for predicting ridesharing flow, characterized in that: The method comprises the following steps: Obtain ridesharing travel related data, generate a ridesharing network based on the ridesharing travel related data, and obtain flow conservation constraints and ridesharing matching constraints, as well as a generalized travel cost function based on the ridesharing network; Generate a set of path flows under the equilibrium state of bounded rational carpooling users based on the carpooling network, as well as a set of acceptable travel plans that can carry the flow; Based on the mathematical programming model with equilibrium constraints, the critical point sequence is solved, and the set of path flows and the set of acceptable travel plans that can carry the flow are determined by using the critical point sequence under the equilibrium state of finite rational carpooling users with a given indifference interval vector. The mathematical programming model with equilibrium constraints is optimized to obtain a mixed integer programming model, and the mixed integer programming model is solved by applying an algorithm to obtain the prediction result of carpooling travel flow.

2. A bounded rational carpooling flow prediction method according to claim 1, characterized in that: The ride-sharing travel related data includes the topological structure of the urban transportation network and travel demand of the ride-sharing travel mode.

3. A bounded rational carpooling flow prediction method according to claim 2, characterized in that: The process of generating a ridesharing network based on ridesharing travel related data: Construct a strongly connected urban transportation network G = (N, A) that includes ridesharing as the ridesharing network G, where N and A represent the node set and the road segment set in the ridesharing network G, respectively. Let W represent the set of OD pairs in the ridesharing network, w represents a certain OD pair, w∈W, and let P represent the set of all paths in the ridesharing network, P w Denotes the path set of OD to w, and P = ∪ w∈W P w ; p represents a path, and p∈P w , let I represent the set of travel modes in the ridesharing network, i represents a certain travel mode, i∈I; the main travel modes in the ridesharing network include empty car drivers, carpooling drivers and carpooling passengers, represented by the sets SD, RD and R respectively, with I=SD∪RD∪R; carpooling drivers and carpooling passengers are the drivers and passengers of the carpooling service, and empty car drivers do not participate in the carpooling service and drive alone; let Indicates the maximum passenger capacity of a vehicle, N i represents the number of carpooling passengers carried by carpooling driver i, Then i=1 represents an empty vehicle driver, and the empty vehicle driver set SD={i=1}; i=1+N i Indicates that it is equipped with N i Rideshare drivers of rideshare passengers, rideshare driver collection Indicates the passengers being carried, the set of passengers being carried 4. The bounded rational method for predicting ridesharing flow according to claim 3 is characterized in that: The process of obtaining the flow conservation constraint and the ride-sharing matching constraint based on the ride-sharing network: Let a∈A represent a certain section of the ridesharing network, represents the flow of travel mode i on path p connecting OD pair w, x a,i represents the flow of travel mode i on road segment a, represents the path-section association coefficient, then the path flow and road traffic x a,i The relationship between them is: Let x a represents the traffic volume of road section a, t a (x a ) represents the travel time function of section a, Denote the travel time of path p between OD pairs w, we have: Let Q w Represents the travel demand of OD between w∈W, and constructs the flow conservation constraint in the ridesharing network G as: Since there are limited seats available for sharing in the car, there is a certain quantitative relationship between the carpooling drivers and the carpooling passengers. Based on the quantitative relationship, the carpooling matching constraint is established as follows: In the formula, Γ r (i) represents the mapping relationship from carpooling driver i∈RD to carpooling passenger i′∈R, i.e., Γ r :RD→R or N i is a positive integer, indicating the number of seats shared by the ridesharing driver i∈RD.

5. The bounded rational method for predicting ridesharing flow according to claim 4, characterized in that: The process of obtaining the generalized travel cost function based on the ride-sharing network is as follows: make represents the path flow vector in the ridesharing network G, The generalized travel cost function of travel mode i on path p connecting OD pair w is: In the formula, They represent the path length of path p between OD pair w respectively; represents the Lagrangian multiplier corresponding to the travel mode i∈RD on the path p between OD pairs w; Γ rd (i) represents the carpooling passenger i∈R to the carpooling driver i ′ ∈RD mapping relationship, Γ rd :R→RD or is Γ r The inverse mapping of (i), parameter ρ i represents the time value of travel mode i∈I; γ i represents the inconvenience coefficient of travel mode i∈RD∪R; b i , With r i They represent the coefficients for calculating the time fee, mileage fee and floating pricing of travel mode i∈RD∪R respectively; c t Represents the cost of running a vehicle, including fuel, charging and parking costs.

6. The bounded rational method for predicting ridesharing flow according to claim 1, characterized in that: The process of generating a set of path flows under a bounded rational carpooling user equilibrium state based on the carpooling network and a set of acceptable travel plans that can carry the flows: In the ride-sharing network G, travelers choose both routes and travel modes. A route p∈P w With a travel mode i∈I, it forms a travel plan (p,i) between OD pair w w , when a travel plan (p,i) w The generalized travel cost Generalized travel costs compared to the optimal travel solution The difference does not exceed a certain threshold called the indifference interval; let ε w represents the indifference interval of OD to w, and takes the value ε w ≥0; ε=(ε w ,w∈W) T represents the indifference interval vector of the ridesharing network G; represents the feasible domain of f. When bounded rational travelers travel in the ride-sharing network G, the traffic flow in the network will form the following bounded rational ride-sharing user equilibrium state: The above formula shows that under the BR-RUE state, travelers between OD pairs w can choose the generalized travel cost Any travel plan (p,i) w , the path flow f following the BR-RUE state is called the BR-RUE path flow pattern, symbolically represented by f ε , at this time, the travel plans in the ridesharing network G can be divided into three states: acceptable, 0-acceptable and unacceptable, which meet the following conditions respectively: Acceptable: 0 - Acceptable: Unacceptable: When ε is given, there is usually more than one ε-BR-RUE path flow pattern f in the ridesharing network G. ε , forming the BR-RUE flow set, which is represented by All f ε The corresponding acceptable travel plans also form a set, which is the ε-acceptable travel plan set, and the symbol is expressed as For a ridesharing network G with multiple OD pairs, the path flow vector is decomposed into f = (f 1 ,f 2 ,…,f w ,…f |W| ), weight represents the path flow between OD pair w; define ε w -BR-RUE flow set is arbitrary The corresponding component f w ,Right now Make the ε-BR-RUE flow set can be expressed as The acceptable travel plans between OD and w form a set, ε w -A set of acceptable travel options, represented by Right now 7. The bounded rational method for predicting ridesharing flow according to claim 1, characterized in that: The process of solving the critical point sequence based on the mathematical programming model with equilibrium constraints: Given any two indifference interval vectors and If for some OD pair w∈W there exists Then the set relation All are established, and the set In the indifference interval ε w When it grows to a certain value, it expands simultaneously; With the indifference interval ε w The value of increases from 0 to +∞, collecting all critical points in the process and combining them into an ascending sequence, mathematically expressed as Among them, k=1,2,…,K, we have This monotone positive sequence divides the non-negative real number interval [0, +∞) into K+1 subintervals, namely The sequence ends are and When the indifference interval ε w Take a critical point Right now When w - A collection of acceptable travel options is represented as With ε w from Grow to New, not part of the previous set The travel plan will be added to the collection In represents a newly added travel plan between OD and w; When ε is given, the ε of any OD to w in the ridesharing network G can be determined by using the critical point sequence. w - A collection of acceptable travel options So as to obtain a set of acceptable travel solutions Then the travel demand Q w Assigned to these acceptable travel options, we get ε w -BR-RUE traffic collection The relationship between each OD and w in the comprehensive ride-sharing network G Get BR-RUE traffic collection Construct a mathematical programming model with equilibrium constraints, namely the MPEC model, to solve the critical point sequence In the above formula, the objective function solves the kth critical point of OD for w is a strict inequality constraint, is a complementary constraint, k=1,2,…,K is a parameter with a value of {0,1}. When the travel plan (p,i) is acceptable w Not included in the collection Inside, on the contrary make represents the maximum generalized travel cost of all acceptable travel options between OD pair w, is the acceptable travel option (p,i) w , indicating the generalized travel cost and the maximum acceptable generalized travel cost The gap, that is 8. The bounded rational method for predicting ridesharing flow according to claim 1, characterized in that: The process of using the critical point sequence to determine the path flow set and the acceptable travel plan set that can carry the flow in the equilibrium state of the finite rational carpooling user with a given indifference interval vector: Solving the MPEC model can obtain the critical point sequence of any OD pair w∈W , where k = 1, 2, ..., K, and the corresponding ε of each critical point is obtained w - A collection of acceptable travel options Then the travel demand Q w Assign to these acceptable travel options to construct ε w -BR-RUE traffic collection For OD pair w, when the indifference interval ε w Gradually increase to One or more new acceptable travel options will be added to ε w - A collection of acceptable travel options New path traffic patterns are also added to ε w -BR-RUE traffic collection and in these new path traffic patterns, the newly added set The acceptable travel options all carry a certain amount of traffic, Indicates when ε w -BR-RUE traffic solution set, Respectively represent ε w = 0 when ε w -Acceptable travel plan set, BR-RUE flow set, decomposition set based on critical points There is a collection relationship The complement The mathematical expression is: In the formula, represents the path flow vector f of OD pair w w The feasible domain of w = 0, then k = 0, and the set is defined It is expressed as: when When w - A collection of acceptable solutions BR-RUE traffic collection for: The set of acceptable travel solutions for the ride-sharing network G BR-RUE traffic collection 9. The bounded rational method for predicting ridesharing flow according to claim 1, characterized in that: The process of applying the algorithm to the mixed integer programming model to obtain the prediction result of the ride-sharing travel flow: The mixed integer programming model, namely the MIP model, is as follows: In the above formula, the auxiliary variable is a 0-1 integer variable; parameter M is a positive number; parameters m1, m2∈(0,1] and have very small values; If the segment travel time function t a (x a ) is a linear function, and the MIP model is a mixed integer linear programming model, which is solved by CPLEX or Gurobi software; if the section travel time function t a (x a ) is a nonlinear function, and the MIP model is a mixed integer nonlinear programming model, which is solved by Gurobi software; Combining multi-process parallel computing technology with network decomposition methods, solving complex ridesharing networks with multiple OD pairs The critical point sequence The network decomposition method is applied to decompose the ridesharing network G into |W| sub-networks with single OD pairs; Create an asynchronous process pool and assign processes to handle subnetworks. represents the kth critical point of OD on w The corresponding 0-1 parameter vector is solved in turn within the sub-network corresponding to OD for w∈W. The calculation process is as follows: Initialize k = 1; for make ν w(k) =0; Solving the MIP model yields - A collection of acceptable travel options If the travel plan Set its corresponding parameters Solve for the kth critical point of OD for w and produce - A collection of acceptable travel options If the travel plan Set the corresponding parameters If ν w(k) =1, the calculation of OD pair w is completed, update w=w+1, k=1, and then perform the calculation of the next OD pair, otherwise, update k=k+1, ν w(k) =v w(k-1) , solve the next critical point for OD with respect to w.

10. A bounded rational carpooling flow prediction system, characterized in that: include: A data processing module is used to obtain ride-sharing travel related data, generate a ride-sharing network based on the ride-sharing travel related data, and obtain flow conservation constraints and ride-sharing matching constraints, as well as a generalized travel cost function based on the ride-sharing network; A set generation module is used to generate a set of path flows under the equilibrium state of bounded rational carpooling users, as well as a set of acceptable travel plans that can carry the flow; The set optimization module is used to solve the critical point sequence based on the mathematical programming model with equilibrium constraints, and use the critical point sequence to determine the path flow set and the acceptable travel plan set that can carry the flow in the equilibrium state of the finite rational carpooling users with a given indifference interval vector; The model solving module is used to optimize the mathematical programming model with equilibrium constraints to obtain a mixed integer programming model, and apply the algorithm to solve the mixed integer programming model to obtain the prediction result of the carpooling travel flow.

Images