Method and system for predicting carpooling traffic flow and carbon emission based on cumulative foreground theory

Abstract

The invention discloses a carpooling flow and carbon emission prediction method and system based on a cumulative foreground theory, and relates to the technical field of flow prediction, and the method comprises the following steps: receiving the related data of an urban traffic network, generating a connected network based on the related data of the urban traffic network, and generating a flow conservation constraint and a carpooling matching constraint based on the connected network; a variational inequality model is generated based on a generalized travel utility function, a flow conservation constraint and a carpooling matching constraint, and the generalized travel utility function is established based on a travel time foreground calculated based on a cumulative foreground theory; solving the variational inequality model based on a parallel adaptive projection gradient descent algorithm to obtain car-pooling user balanced flow based on a cumulative foreground theory; and calculating the influence effect of the car-pooling travel mode on urban traffic carbon emission through a pre-established environmental cost function and car-pooling user balanced flow based on a cumulative foreground theory.

Description

Technical Field

[0001] The present invention relates to technical fields such as carpool flow prediction and environmental impact assessment, and specifically to a carpool flow and carbon emission prediction method and system based on cumulative prospect theory. Background Art

[0002] Through real-time driver-passenger matching, route planning, mobile payment and other technologies, the carpooling service platform not only provides urban residents with an affordable, convenient and efficient travel experience, but also effectively alleviates traffic congestion, reduces energy consumption and reduces environmental pollution, which is of great significance to the sustainable development of modern urban transportation. In the transportation system, when faced with potential road risks caused by uncertain factors such as traffic congestion, road accidents, facility failures or bad weather, travelers usually set psychological reference points and evaluate the utility of travel plans based on personal experience and current situations, aversion to losses is greater than preference for gains, and nonlinear weighting is performed on the actual probability of environmental risks, overestimating the impact of extreme low-probability events. The cumulative prospect theory fully considers this complex psychology of travelers' decision-making process. By integrating the cumulative prospect theory with carpooling travel mode, the traffic model can more accurately capture the impact of environmental risks in modern urban transportation networks on travelers' decisions, improve the accuracy and practical significance of travel behavior modeling, and thus help traffic managers formulate robust and robust planning schemes and management policies, and promote the construction of a more efficient, environmentally friendly and sustainable modern urban transportation system.

[0003] A travel selection method based on cumulative prospect theory (CN201810672771.3) established a game strategy model based on cumulative prospect theory, and studied people's travel choices in different travel scenarios based on this model. A method for predicting residents' travel demand for urban shared transportation (CN202110918310.1) takes into account the randomness and diversity of travelers' subjective perception of cost, and establishes a random carpooling user equilibrium model to predict the service level of the urban transportation network and the residents' travel demand under the influence of this service level. The former invention takes into account the uncertainty risk of the transportation network, and the latter invention takes into account the carpooling travel mode of shared transportation, but the existing technology rarely takes both into account and fails to focus on the traffic prediction of urban shared transportation networks under uncertain risks. 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 method for predicting carpooling flow and carbon emissions based on cumulative prospect theory.

[0005] In the first aspect, the purpose of the present invention can be achieved by the following technical solution: a method for predicting carpooling flow and carbon emissions based on cumulative prospect theory, the method comprising the following steps:

[0006] Receiving urban traffic network related data, generating a connected network based on the urban traffic network related data, and generating flow conservation constraints and carpooling matching constraints based on the connected network;

[0007] Generate a variational inequality model based on a generalized travel utility function and flow conservation constraints and carpool matching constraints, wherein the generalized travel utility function is established based on the travel time prospect calculated by cumulative prospect theory;

[0008] The variational inequality model is solved based on the parallel adaptive projected gradient descent algorithm, and the equilibrium flow of carpooling users based on the cumulative prospect theory is obtained;

[0009] Through the pre-established environmental cost function, the impact of carpooling travel mode on urban transportation carbon emissions is calculated based on the equilibrium flow of carpooling users based on the cumulative prospect theory.

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

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

[0012] According to the topological structure, travel demand and transportation tools of the urban transportation network, a connected network G = (N, A) is constructed. In the connected network G, N represents the node set; A represents the road segment set, a∈A represents a certain road segment; W represents the OD pair set, w∈W represents a certain OD pair; P represents the path set, P w represents the set of OD paths to w, p∈P w represents a path; I represents a travel mode set, and i∈I represents a travel mode.

[0013] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of generating flow conservation constraints and carpooling matching constraints based on the connected network:

[0014] The urban traffic network includes self-driving drivers who drive their own cars and carpooling drivers and passengers, which are represented by set symbols SD, RD and R respectively. There is a set relation I = SD∪RD∪R. Let i = 1 represent self-driving drivers, and the set of self-driving drivers SD = {i = 1}; let i = 1 + N i Indicates that it is equipped with N i Carpool drivers with carpool passengers, carpool driver collection make Indicates the number of carpooling passengers being carried, the carpooling passenger collection in, Indicates the maximum passenger capacity of the vehicle. represents the number of carpooling passengers picked up by carpooling driver i.

[0015] Let x a,i represents the traffic flow of travel mode i on road segment a, Denotes a path p∈P w The traffic flow of travel mode i has a quantitative relationship:

[0016]

[0017] In the formula, represents the path-segment association coefficient;

[0018] Let Q w Representing OD's travel demand for w∈W, the following flow conservation constraint is established:

[0019]

[0020] The following carpooling matching constraints exist between the driver and the passenger:

[0021]

[0022] In the formula, Γ r (i) represents the mapping from carpooling driver i∈RD to carpooling passenger i′∈R, with Γ r :RD→R or

[0023] make represents the path flow vector, and its feasible domain is A path p∈P w With a travel mode i, a travel plan (p,i) is formed w .

[0024] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: the process of calculating the travel time prospect based on the cumulative prospect theory:

[0025] The travel time T of segment a a Defined as a continuous random variable with mean variance represents the average travel time of road section a, let Denotes a path p∈P w The travel time on is also a continuous random variable with a mean variance It represents the average travel time on the path p connecting the OD pair w. There is the following quantitative relationship between the segment travel time and the path travel time:

[0026]

[0027] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of establishing the generalized travel utility function is as follows:

[0028] make They represent the cumulative distribution function and probability density function of the path travel time respectively, and Prob(·) represents the probability of occurrence of a certain event (·), then there exists:

[0029]

[0030] According to the cumulative prospect theory, the value function of the path travel time is as follows:

[0031]

[0032] In the formula, parameters α and β are the risk coefficients of travelers, η is the loss aversion coefficient, Denotes a path p∈P w The reference point preset by the traveler; when When , travelers gain benefits and are risk averse; when When , travelers bear losses and prefer risks, and the value function g(·) decreases monotonically. Continuous but not differentiable;

[0033] Reference Points is the boundary between gains and losses, with path p∈P w Travel time budget As a reference point, the math is:

[0034]

[0035] Where the parameter μ represents the importance of on-time arrival and is related to the probability that the traveler will arrive at the destination within the allocated travel time budget:

[0036]

[0037] In the formula It is defined as travel time reliability within budget;

[0038] The probability weight function w(p′) reflects the nonlinear perception of travelers on probability, and its mathematical form is as follows:

[0039] w(p)=exp[-(-lnp′) γ ],0<γ<1

[0040] Where w(p′) and p′ represent the perceived probability and actual probability of an event respectively;

[0041] make Denotes a path p∈P w The travel time prospect on w + (·) and w - (·) represent the probability weight function w(·) of gaining profit and suffering loss, respectively. Then, any path p∈P in the urban transportation network w The travel time outlook is:

[0042]

[0043] make Represents travel plan (p,i) w The travel utility function is as follows:

[0044]

[0045] In the formula, Denotes a path p∈P w The length of i , γ i 、b i , r i They represent the time value, inconvenience coefficient, duration fee coefficient, mileage fee coefficient, and floating pricing coefficient of travel mode i respectively; c t Respectively represent vehicle operating costs;

[0046] Ride-sharing platforms introduce subsidies and premiums as additional costs in the form of:

[0047]

[0048] Where Γ rd (·) is Γ r The inverse mapping of (·) is Γ rd :R→RD; Denotes a path p∈P w The Lagrangian multiplier generated by the carpool matching constraints of the carpool driver i∈RD; after adding the additional cost, the travel plan (p,i) w The generalized travel utility function for:

[0049]

[0050] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of generating a variational inequality model based on the generalized travel utility function, the flow conservation constraint, and the carpooling matching constraint includes:

[0051] In an urban transportation network, when travelers pursue a travel plan with the greatest generalized travel utility, the traffic flow will form a carpooling user equilibrium state based on the cumulative prospect theory, which can be mathematically expressed as:

[0052]

[0053] In the formula, π w It represents the Lagrangian multiplier generated by the flow conservation constraint of OD on w, and it represents the maximum generalized travel utility of OD's travel plan for w;

[0054] Construct the following variational inequality model:

[0055]

[0056] In the formula, the vector function is a built-in function of the VI model, f * represents the unique flow solution for the VI model.

[0057] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of solving the variational inequality model based on the parallel adaptive projected gradient descent algorithm is as follows:

[0058] Define f w is the path flow vector of OD to w, that is Ω w is the vector f w The feasible domain of Ψ(f w ) is the vector function of OD to w, are the vector f obtained at the kth iteration w ,Ψ(f w );β k Indicates the step size at the kth iteration; represents the projection operator, that is, the vector [·] in the set Ω w The process of the parallel adaptive projection gradient descent algorithm PSAPGD is as follows:

[0059] Initialization, given accuracy ε>0, initial step size β0=1, μ=0.5, ν=0.9, k=1, and initial feasible path flow

[0060] Check the stop condition and set express In the set Ωw Projection on:

[0061]

[0062] if If yes, stop the calculation; otherwise, continue the execution;

[0063] Calculate intermediate variables r k :

[0064]

[0065] If r k >ν, then update the iteration step size; otherwise, update the iteration variable;

[0066] Update the iteration step and calculate the intermediate variables again r k :

[0067] β k :=β k *0.8 / r k

[0068]

[0069] If r k >ν, repeatedly update the iteration step; otherwise update the iteration variable;

[0070] Update iteration variables

[0071]

[0072] If r k ≤μ, let the step length be k+1 :=β k *1.5, k=k+1, recheck the stop condition.

[0073] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the process of calculating the impact of the carpooling travel mode on urban traffic carbon emissions based on the environmental cost function and the carpooling equilibrium flow is as follows:

[0074] Let x = (x a,i ,a∈A,i∈I) T represents the road flow vector, l a Represents the length of road section a and the average travel time of road section a is a function of x, that is Environmental cost e of road section a a is also a function of x, that is, e a =e a (x), expressed as:

[0075]

[0076] The environmental cost of the entire urban road network T EC (x) is:

[0077] T EC (x) = e(x) T x vh

[0078] In the formula, e(x)=(e a (x),a∈A) T represents the environmental cost vector in the urban road network; represents the vehicle flow vector, represents the motor vehicle flow rate on road section a, which is the sum of the self-driving driver flow rate and the carpooling driver flow rate, that is,

[0079] In the second aspect, in order to achieve the above-mentioned purpose, the present invention discloses a carpooling flow and carbon emission prediction system based on cumulative prospect theory, comprising:

[0080] A constraint generation module, used for receiving urban traffic network related data, generating a connected network based on the urban traffic network related data, and generating flow conservation constraints and carpooling matching constraints based on the connected network;

[0081] A model building module, for generating a variational inequality model based on a generalized travel utility function and a flow conservation constraint and a carpooling matching constraint, wherein the generalized travel utility function is established based on a travel time prospect calculated by a cumulative prospect theory;

[0082] The model solving module is used to solve the variational inequality model based on the parallel adaptive projected gradient descent algorithm to obtain the equilibrium flow of carpooling users based on the cumulative prospect theory;

[0083] The environmental evaluation module is used to calculate the impact of carpooling on urban transportation carbon emissions through a pre-established environmental cost function and the equilibrium flow of carpooling users based on the cumulative prospect theory.

[0084] Beneficial effects of the present invention:

[0085] The present invention predicts the flow distribution state in an urban transportation network that includes carpooling travel modes. First, the present invention constructs a connected network with flow conservation constraints and carpooling matching constraints based on data such as the road network topology, residents' travel needs, and commonly used transportation tools. Secondly, the present invention simulates the travel decision-making process of travelers under environmental risks with the cumulative prospect theory, calculates the travel time prospect, and establishes a generalized travel utility function. Finally, the present invention constructs a variational inequality model and designs a parallel adaptive projected gradient descent algorithm for solution, and studies the traffic flow under the equilibrium state of carpooling users based on the cumulative prospect theory. The present invention helps managers understand the impact of potential environmental risks such as facility failures, traffic accidents, and severe weather in urban transportation networks, so as to formulate scientific, reasonable, robust and robust transportation planning schemes and management policies, and promote the construction of a more efficient, environmentally friendly, and sustainable urban transportation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] 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.

[0087] Figure 1 It is a schematic flow chart of the method of the present invention;

[0088] Figure 2 The topological structure and travel demand diagram of the example network of this embodiment;

[0089] Figure 3 It is a road section flow diagram of the example network of this embodiment;

[0090] Figure 4 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0091] 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.

[0092] Embodiment 1:

[0093] like Figure 1 As shown, the method for predicting carpooling flow and carbon emissions based on cumulative prospect theory includes the following steps:

[0094] S101: receiving urban traffic network related data, generating a connected network based on the urban traffic network related data, and generating flow conservation constraints and carpooling matching constraints based on the connected network;

[0095] Urban transportation network-related data include the topological structure of the urban transportation network, travel demand and transportation tools;

[0096] like Figure 2 , based on the road network topology, residents' travel demand, common transportation tools and other data of the example network, a connected network G = (N, A) is constructed. The network has 13 nodes, 19 road segments and 4 OD pairs, namely N = {1,2,…,13}, A = {a = 1,2,…,19}, W = {w = 1,2,3,4}. The 4 OD pairs are (1,2), (1,3), (4,2) and (4,3), and their travel demands are Q 1 =100, Q 2 =120, Q 3 =100 and Q 4 =80, and each OD pair has at most 7, 5, 5, and 6 paths respectively, that is, |P 1 |=7、|P 2 |=5、|P 3 |=5 and |P 4 |=6. As shown below, Tables 1 and 2 show the road segment characteristics and path connection information of the example network, where FFT is the free flow time. Average travel time The classic Bureau of Public Road (BPR) function is used for calculation.

[0097] Table 1. Road segment characteristics of the example network

[0098]

[0099]

[0100] Table 2 Path connection information of the example network

[0101]

[0102] Let I represent the set of travel modes, and i∈I represent a certain travel mode. The urban traffic network mainly includes self-driving drivers, carpooling drivers and passengers, etc., which are represented by set symbols SD, RD and R respectively. There is a set relation I=SD∪RD∪R. Assume that the upper limit of vehicle passenger capacity is Then |I|=5; i=1 represents a self-driving driver, and the set of self-driving drivers SD={i=1}; i=2,3 represents a carpooling driver, and the set of carpooling drivers RD={i=2,3}; i=4,5 represents a carpooling passenger, and the set of carpooling passengers R={i=4,5}.

[0103] Let x a,i represents the traffic flow of travel mode i on road segment a, Denotes a path p∈P w The traffic flow of travel mode i has a quantitative relationship:

[0104]

[0105] In the formula, Represents the path-section association coefficient. When path p passes through section a, on the contrary

[0106] There is the following flow conservation constraint between any OD pair w∈W:

[0107]

[0108] The following carpooling matching constraints exist between the driver and the passenger:

[0109]

[0110] In the formula, represents the number of carpooling passengers carried by carpooling driver i; Γ r (i) represents the mapping from carpooling driver i∈RD to carpooling passenger i′∈R, with Γ r :RD→R or

[0111] make represents the path flow vector, and its feasible domain is In an urban transportation network, travelers choose both the path and the travel mode when making travel decisions. A path p∈P w With a travel mode i, a travel plan (p,i) is formed w .

[0112] S102: generating a variational inequality model based on a generalized travel utility function, a flow conservation constraint, and a carpooling matching constraint, wherein the generalized travel utility function is established based on a travel time prospect calculated by a cumulative prospect theory;

[0113] The travel time prospect is calculated based on the cumulative prospect theory, and a generalized travel utility function is established.

[0114] Assume that the travel time in the example network follows a normal distribution, and the mean and variance have the relationship Right now in is the coefficient of variation. The cumulative distribution function of the path travel time Probability density function They are:

[0115]

[0116] According to the cumulative prospect theory, when people make travel decisions, they judge their own gains or losses based on the reference point. The value function of the path travel time is as follows:

[0117]

[0118] In the formula, the parameters α, β (0 < α, β ≤ 1) reflect the risk attitude of the traveler and determine the degree to which the traveler's sensitivity to gains or losses away from the reference point decreases. The loss aversion coefficient η (η ≥ 1) indicates that the traveler is more sensitive to losses than to gains. Denotes a path p∈P w The reference point preset by the traveler; when When , travelers gain benefits and are risk averse; when When , the traveler bears the loss and prefers risk. The value function g(·) is monotonically decreasing. Continuous but not differentiable.

[0119] Reference Points is the boundary between gain and loss, and the traveler is risk-neutral at this time. w Travel time budget As a reference point, this is expressed mathematically as:

[0120]

[0121] Here, the parameter μ indicates the importance of arriving on time. For example, for trips that incur high costs for being late (such as catching a plane or a high-speed train), travelers usually allocate a higher travel time budget, making the value of μ larger. From a mathematical point of view, μ is related to the probability that the traveler will arrive at the destination within the allocated travel time budget:

[0122]

[0123] In the formula It is defined as the reliability of travel time within the budget, which represents the probability that the actual travel time does not exceed the travel time budget.

[0124] Since the travel time of the example network Following the normal distribution, when μ=0, at this time Conforms to the standard normal distribution, travel time budget can be expressed as:

[0125]

[0126] In the formula Represents the inverse of the standard normal distribution function.

[0127] The probability weight function w(p′) reflects the nonlinear perception of travelers to probability, especially the overreaction to low-probability events and the underreaction to high-probability events. Its mathematical form is as follows:

[0128] w(p)=exp[-(-lnp′) γ ],0<γ<1

[0129] Where w(p′) and p′ represent the perceived probability and actual probability of an event, respectively. Regardless of the value of γ, the probability weight function w(·) is always strictly monotonically increasing with respect to p′.

[0130] make Denotes a path p∈P w The travel time prospect on w + (·) and w - (·) represent the probability weight function w(·) when gaining profit and suffering loss respectively. At this time, any path p∈P w The travel time outlook is:

[0131]

[0132] make Represents travel plan (p,i) w The travel utility function is as follows:

[0133]

[0134] In the formula, Denotes a path p∈P w The length of i , γ i 、b i , r i They represent the time value, inconvenience coefficient, duration fee coefficient, mileage fee coefficient, and floating pricing coefficient of travel mode i respectively; c t They represent the vehicle operating costs, such as fuel, charging, parking fees, etc. The values ​​of each parameter are shown in Table 3.

[0135] Table 3 Parameter settings of numerical experiments

[0136]

[0137] Carpooling platforms introduce subsidies and premiums as additional costs to balance the potential supply and demand of the carpooling service market in the form of:

[0138]

[0139] Where Γ rd (·) is Γ r The inverse mapping of (·) is Γ rd :R→RD; Denotes a path p∈P w The Lagrangian multiplier generated by the carpool matching constraints of the carpool driver i∈RD. After adding the additional cost, the travel plan (p,i) w The generalized travel utility function for:

[0140]

[0141] Based on the generalized travel utility function, flow conservation constraints and carpooling matching constraints, the process of generating a variational inequality model includes:

[0142] In an urban transportation network, when travelers pursue a travel plan with the maximum generalized travel utility, the traffic flow will form a ridesharing user equilibrium based on cumulative prospect theory (CPT-RUE), which is mathematically expressed as:

[0143]

[0144] In the formula, π w It represents both the Lagrangian multiplier generated by the flow conservation constraint of OD on w and the maximum generalized travel utility of OD's travel plan for w.

[0145] In order to study the flow distribution under the CPT-RUE state, the following variational inequality (VI) model is constructed:

[0146]

[0147] In the formula, the vector function is a built-in function of the VI model, f * represents the unique flow solution for the VI model.

[0148] S103: Solving the variational inequality model based on a parallel adaptive projected gradient descent algorithm, and obtaining the equilibrium flow of carpooling users based on the cumulative prospect theory.

[0149] In order to solve the VI model, the network decomposition and parallel computing technology are used to improve the existing adaptive projected gradient algorithm, and the parallel self-adaptive projected gradient descent (PSAPGD) algorithm is proposed. The algorithm condition is that "the built-in function Ψ(f) is monotonic and Lipschitz continuous in the feasible domain Ω".

[0150] Define f w is the path flow vector of OD to w, that is Ω w is the vector f w The feasible domain of Ψ(f w ) is the vector function of OD to w, are the vector f obtained at the kth iteration w ,Ψ(f w );β k Indicates the step size at the kth iteration; represents the projection operator, that is, the vector [·] in the set Ω w The PSAPGD algorithm process is as follows:

[0151] Step 1: Initialization. Given accuracy ε>0, initial step size β0=1, μ=0.5, ν=0.9, k=1, and initial feasible path flow

[0152] Step 2 checks the stop condition. express In the set Ω w Projection on:

[0153]

[0154] if Then stop the calculation; otherwise, go to step 3.

[0155] Step 3: Calculate intermediate variables r k :

[0156]

[0157] If r k >ν, execute step 4; otherwise, execute step 5.

[0158] Step 4: Update the iteration step size and calculate the intermediate variables again. r k :

[0159] β k :=β k *0.8 / r k

[0160]

[0161] If r k >ν, repeat step 4; otherwise, go to step 5.

[0162] Step 5 updates the iteration variable.

[0163]

[0164] If r k ≤μ, let the step length be k+1 :=β k *1.5, k=k+1, go to step 2.

[0165] Specifically, the present invention is further described below by way of embodiments:

[0166] After solving the PSAPGD algorithm, the traffic distribution of the example network in the CPT-RUE state is shown in Table 4. Figure 3 The road flow of the example network in the CPT-RUE state is further presented. i∈I x a,i , Its value is divided into 4 intervals (0,50], (50,100], (100,150], (150,+∞), and the unmarked section flow ∑ i∈I x a,i =0.

[0167] Table 4. Traffic flow of road segments in the example network

[0168]

[0169] S104: Using the pre-established environmental cost function, the equilibrium flow of carpooling users based on the cumulative prospect theory is used to calculate the impact of carpooling on urban transportation carbon emissions.

[0170] Define x = (x a,i ,a∈A,i∈I) T is the road flow vector, l a is the length of section a, and the average travel time of section a Then the environmental cost of section a is e a =e a (x), expressed as:

[0171]

[0172] The environmental cost of the entire urban road network T EC (x) is:

[0173] T EC (x) = e(x) T x vh

[0174] In the formula, e(x)=(e a (x),a∈A) T represents the environmental cost vector in the urban road network; represents the vehicle flow vector, represents the motor vehicle flow rate on road section a, which is the sum of the self-driving driver flow rate and the carpooling driver flow rate, that is,

[0175] The motor vehicle flow and environmental cost of the road section in the example network at this time are shown in Table 5.

[0176] Table 5: Vehicle flow and environmental cost of road sections in the example network

[0177]

[0178] The environmental cost of the entire urban road network T EC =73394.89.

[0179] Embodiment 2: In the second aspect, as Figure 4 As shown, in order to achieve the above-mentioned purpose, the present invention discloses a carpooling flow and carbon emission prediction system based on cumulative prospect theory, comprising:

[0180] A constraint generation module 11 is used to receive urban traffic network related data, generate a connected network based on the urban traffic network related data, and generate flow conservation constraints and carpooling matching constraints based on the connected network;

[0181] A model building module 12, for generating a variational inequality model based on a generalized travel utility function and a flow conservation constraint and a carpooling matching constraint, wherein the generalized travel utility function is established based on a travel time prospect calculated by a cumulative prospect theory;

[0182] A model solving module 13 is used to solve the variational inequality model based on a parallel adaptive projected gradient descent algorithm to obtain a carpooling user equilibrium flow based on a cumulative prospect theory;

[0183] The environmental evaluation module 14 is used to calculate the impact of carpooling on urban transportation carbon emissions through a pre-established environmental cost function and the carpooling user equilibrium flow based on the cumulative prospect theory.

[0184] 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.

[0185] 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.

[0186] 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.

[0187] 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 method for predicting carpooling flow and carbon emissions based on cumulative prospect theory, characterized in that: The method comprises the following steps: Receiving urban traffic network related data, generating a connected network based on the urban traffic network related data, and generating flow conservation constraints and carpooling matching constraints based on the connected network; Generate a variational inequality model based on a generalized travel utility function and flow conservation constraints and carpool matching constraints, wherein the generalized travel utility function is established based on the travel time prospect calculated by cumulative prospect theory; The variational inequality model is solved based on the parallel adaptive projected gradient descent algorithm, and the equilibrium flow of carpooling users based on the cumulative prospect theory is obtained. Through the pre-established environmental cost function, the impact of carpooling travel mode on urban transportation carbon emissions is calculated based on the equilibrium flow of carpooling users based on the cumulative prospect theory.

2. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1 is characterized in that: The urban transportation network related data includes the topological structure, travel demand and transportation tools of the urban transportation network.

3. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 2 is characterized in that: The process of generating a connected network based on urban traffic network related data: According to the topological structure, travel demand and transportation tools of the urban transportation network, a connected network G = (N, A) is constructed. In the connected network G, N represents the node set; A represents the road segment set, a∈A represents a certain road segment; W represents the OD pair set, w∈W represents a certain OD pair; P represents the path set, P w represents the set of OD paths to w, p∈P w represents a path; I represents a travel mode set, and i∈I represents a travel mode.

4. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 3 is characterized in that: The process of generating flow conservation constraints and carpooling matching constraints based on the connected network: The urban traffic network includes self-driving drivers who drive their own cars and carpooling drivers and passengers, which are represented by set symbols SD, RD and R respectively. There is a set relation I = SD∪RD∪R. Let i = 1 represent self-driving drivers, and the set of self-driving drivers SD = {i = 1}; let i = 1 + N i Indicates that it is equipped with N i Carpool drivers with carpool passengers, carpool driver collection make Indicates the number of carpooling passengers being carried, the carpooling passenger collection in, Indicates the maximum passenger capacity of the vehicle. represents the number of carpooling passengers picked up by carpooling driver i; Let x a,i represents the traffic flow of travel mode i on road segment a, Denotes a path p∈P w The traffic flow of travel mode i has a quantitative relationship: In the formula, represents the path-segment association coefficient; Let Q w Representing OD's travel demand for w∈W, the following flow conservation constraint is established: The following carpooling matching constraints exist between the driver and the passenger: In the formula, Γ r (i) represents the mapping from carpooling driver i∈RD to carpooling passenger i′∈R, with Γ r :RD→R or make represents the path flow vector, and its feasible domain is A path p∈P w With a travel mode i, a travel plan (p,i) is formed w .

5. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1 is characterized in that: The process of calculating the travel time prospect based on the cumulative prospect theory is as follows: The travel time T of segment a a Defined as a continuous random variable with mean variance represents the average travel time of road section a, let Denotes a path p∈P w The travel time on is also a continuous random variable with a mean variance It represents the average travel time on the path p connecting the OD pair w. There is the following quantitative relationship between the segment travel time and the path travel time:

6. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1 is characterized in that: The establishment process of the generalized travel utility function is as follows: make They represent the cumulative distribution function and probability density function of the path travel time respectively, and Prob(·) represents the probability of occurrence of a certain event (·), then there exists: According to the cumulative prospect theory, the value function of the path travel time is as follows: In the formula, parameters α and β are the risk coefficients of travelers, η is the loss aversion coefficient, Denotes a path p∈P w The reference point preset by the traveler; when When , travelers gain benefits and are risk averse; when When , travelers bear losses and prefer risks, and the value function g(·) decreases monotonically. Continuous but not differentiable; Reference Points is the boundary between gains and losses, with path p∈P w Travel time budget As a reference point, the math is: Where the parameter μ represents the importance of on-time arrival and is related to the probability that the traveler will arrive at the destination within the allocated travel time budget: In the formula It is defined as travel time reliability within budget; The probability weight function w(p′) reflects the nonlinear perception of travelers on probability, and its mathematical form is as follows: w(p)=exp[-(-lnp′) γ ],0<γ<1 Where w(p′) and p′ represent the perceived probability and actual probability of an event respectively; make Denotes a path p∈P w The travel time prospect on w + (·) and w - (·) represent the probability weight function w(·) of gaining profit and suffering loss, respectively. Then, any path p∈P in the urban transportation network w The travel time outlook is: make Represents travel plan (p,i) w The travel utility function is as follows: In the formula, Denotes a path p∈P w The length of i , γ i , b i , r i They represent the time value, inconvenience coefficient, duration fee coefficient, mileage fee coefficient, and floating pricing coefficient of travel mode i respectively; c t Respectively represent vehicle operating costs; Ride-sharing platforms introduce subsidies and premiums as additional costs in the form of: Where Γ rd (·) is Γ r The inverse mapping of (·) is Γ rd :R→RD; Denotes a path p∈P w The Lagrangian multiplier generated by the carpool matching constraints of the carpool driver i∈RD; after adding the additional cost, the travel plan (p,i) w The generalized travel utility function for:

7. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1 is characterized in that: The process of generating a variational inequality model based on the generalized travel utility function, the flow conservation constraint and the carpooling matching constraint includes: In an urban transportation network, when travelers pursue a travel plan with the greatest generalized travel utility, the traffic flow will form a carpooling user equilibrium state based on the cumulative prospect theory, which can be mathematically expressed as: In the formula, π w It represents the Lagrangian multiplier generated by the flow conservation constraint of OD on w, and it represents the maximum generalized travel utility of OD's travel plan for w; Construct the following variational inequality model: In the formula, the vector function is a built-in function of the VI model, f * represents the unique flow solution for the VI model.

8. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1 is characterized in that: The process of solving the variational inequality model based on the parallel adaptive projected gradient descent algorithm is as follows: Define f w is the path flow vector of OD to w, that is Ω w is the vector f w The feasible domain of Ψ(f w ) is the vector function of OD with respect to w, we have are the vector f obtained at the kth iteration w ,Ψ(f w );β k Indicates the step size at the kth iteration; represents the projection operator, that is, the vector [·] in the set Ω w The process of the parallel adaptive projection gradient descent algorithm PSAPGD is as follows: Initialization, given accuracy ε>0, initial step size β0=1, μ=0.5, ν=0.9, k=1, and initial feasible path flow Check the stop condition and set express In the set Ω w Projection on: if If yes, stop the calculation; otherwise, continue the execution; Calculate intermediate variables r k : If r k >ν, then update the iteration step size; otherwise, update the iteration variable; Update the iteration step and calculate the intermediate variables again r k : β k ：＝β k *0.8 / r k If r k >ν, repeatedly update the iteration step; otherwise update the iteration variable; Update iteration variables If r k ≤μ, let the step length be k+1 :=β k *1.5, k=k+1, recheck the stop condition.

9. The method for predicting carpooling flow and carbon emissions based on cumulative prospect theory according to claim 1, characterized in that: The process of calculating the impact of carpooling on the urban environment is as follows: Let x = (x a,i ,a∈A,i∈I) T represents the road flow vector, l a Represents the length of road section a and the average travel time of road section a is a function of x, that is Environmental cost e of road section a a is a function of x, that is, e a =e a (x), expressed as: The environmental cost of the entire urban road network T EC (x) is: T EC (x)=e(x) T x vh In the formula, e(x)=(e a (x),a∈A) T represents the environmental cost vector in the urban road network; represents the vehicle flow vector, represents the motor vehicle flow rate on road section a, which is the sum of the self-driving driver flow rate and the carpooling driver flow rate, that is, 10. The carpooling flow and carbon emission prediction system based on cumulative prospect theory is characterized by: include: A constraint generation module, used for receiving urban traffic network related data, generating a connected network based on the urban traffic network related data, and generating flow conservation constraints and carpooling matching constraints based on the connected network; A model building module, for generating a variational inequality model based on a generalized travel utility function and a flow conservation constraint and a carpooling matching constraint, wherein the generalized travel utility function is established based on a travel time prospect calculated by a cumulative prospect theory; The model solving module is used to solve the variational inequality model based on the parallel adaptive projected gradient descent algorithm to obtain the equilibrium flow of carpooling users based on the cumulative prospect theory; The environmental evaluation module is used to calculate the impact of carpooling on urban transportation carbon emissions through a pre-established environmental cost function and the equilibrium flow of carpooling users based on the cumulative prospect theory.

Images