Traffic capacity coordination and allocation method and command transmission system based on passenger flow prediction

Abstract

The present invention belongs to the field of transportation technology, and specifically relates to a transportation capacity joint dispatching allocation method and a command transmission system based on passenger flow prediction, which includes: S1, obtaining urban traffic passenger flow data, and using the entropy weight method to analyze passenger flow influencing factors; S2, constructing a long-short-term memory network passenger flow prediction model to predict urban traffic passenger flow data; S3, realizing transportation capacity allocation based on a large traffic model according to passenger flow prediction data; S4, realizing dynamic adjustment and coordinated dispatch of transportation capacity according to the transportation capacity allocation result. The present invention realizes passenger flow prediction based on the prediction of the long-short-term memory network model, scientifically arranges transportation capacity, optimizes road network layout, and improves resource utilization efficiency; the present invention comprehensively uses modern information technology and management means to realize the coordination and linkage of passenger flow monitoring, capacity dispatching, emergency response, and passenger services, and the capacity joint dispatching analysis based on the large traffic model can optimize the allocation of transportation resources and improve the efficiency of transportation operations.

Description

Technical Field

[0001] The present invention belongs to the technical field of transportation, and in particular relates to a transportation capacity coordination and allocation method based on passenger flow prediction and an instruction transmission system. Background Art

[0002] With the increase of popular business districts, popular scenic spots and popular routes, the existing urban traffic organization mode has had a huge impact, bringing more and more problems to the public's transportation safety and travel efficiency. First of all, various modes of transportation, including taxis and online car-hailing, have fixed stations. With the increase in passenger flow, the density of people is also increasing, and there may be problems with passenger transfers and crossovers, causing abnormal congestion in local areas. Secondly, the untimely dispatch of transportation vehicles may cause congestion at the entrance of the station, and the resulting safety problems will gradually increase. In addition, it is inconvenient to connect multiple modes of transportation, and passenger lines are constantly crossing in the transfer path. The interweaving of people will inevitably turn the orderly flow of people into disorder in a short period of time, the chaos of connection and transfer will increase rapidly, and safety hazards will appear. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention provides a transportation capacity coordination and allocation method based on passenger flow prediction and a command transmission system, which realizes passenger flow prediction based on long-short-term memory network model prediction, scientifically arranges transportation capacity, optimizes road network layout, and improves resource utilization efficiency; comprehensively uses modern information technology and management methods to achieve passenger flow monitoring, capacity scheduling, emergency response, and passenger service coordination and linkage, and capacity coordination analysis based on large traffic models can optimize transportation resource allocation and improve transportation operation efficiency.

[0004] To achieve the above object, the present invention discloses the following technical solution:

[0005] A method for joint dispatching and allocating transportation capacity based on passenger flow prediction, comprising:

[0006] S1: Obtain the urban traffic passenger flow data to obtain the actual value p of the urban traffic passenger flow data in , and use the entropy weight method to analyze the factors affecting passenger flow, specifically:

[0007] Summarize the passenger transport hub data, railway passenger transport data, airport passenger transport data and traffic checkpoint statistics, and use the above data to superimpose to obtain the real value p of urban traffic passenger flow data in ;

[0008] Determine the factors affecting passenger flow and use the entropy weight method to analyze the impact of passenger flow factors on urban traffic passenger flow data;

[0009] S2: Construct a long short-term memory network passenger flow prediction model to predict urban traffic passenger flow data, which includes the following sub-steps:

[0010] S21: constructing a long short-term memory network passenger flow prediction model including a forget gate, an input gate and an output gate, obtaining the urban traffic passenger flow data in step S1, and inputting the long short-term memory network passenger flow prediction model according to the time series;

[0011] S22: Calculate the output data of the forget gate and the input gate respectively, generate candidate passenger flow data memory units, and generate the current passenger flow data memory unit state as follows:

[0012]

[0013] Among them, C t is the memory unit state at the current moment; f t is the output of the forget gate; C t-1 is the memory unit state at the previous moment; i t is the output of the input gate; is a new candidate memory unit;

[0014] S23: Calculate the output data of the output gate and determine the hidden state of the current passenger flow data memory unit:

[0015] h t =O t ×tanh(C t );

[0016] Among them, h t is the hidden state at the current moment; O t is the output of the output gate; tanh is the second layer activation function;

[0017] S24: Update the weight matrix and bias inside the long short-term memory network passenger flow prediction model according to the hidden state of the passenger flow data memory unit; calculate the predicted value p of the urban traffic passenger flow data using the loss function of the long short-term memory network passenger flow prediction model out and the actual value p of the urban traffic passenger flow data in step S1 in The gap is:

[0018] Q=||p out -p in || 2 ≤Sq;

[0019] Where Q is the loss function of the long short-term memory network passenger flow prediction model; p out is the predicted value of urban traffic passenger flow data; p in is the real value of urban traffic passenger flow data, obtained by using the long short-term memory network passenger flow prediction model; Sq is the loss function indicator of the long short-term memory network passenger flow prediction model;

[0020] S25: Divide the passenger flow level into four levels according to the predicted value of urban traffic passenger flow data: low peak period Le1, flat peak period Le2, peak period Le3 and extreme peak period Le4;

[0021] S3: according to the urban traffic passenger flow forecast data in step S2, realize the traffic capacity allocation based on the traffic big model;

[0022] Based on the traffic big model, traffic capacity joint coordination analysis is carried out. According to the predicted value of urban traffic passenger flow data in step S2, according to the passenger flow level and the passenger flow influencing factors in step S1, taxis, online car-hailing vehicles and buses are dispatched and routed. The formula of the traffic big model is as follows:

[0023] (t, w, b, p, l, d) = TG (pout, Ld, Ca, Le, S);

[0024] Among them, t is the number of taxis; w is the number of online-hailing cars; b is the number of buses; p is the configured station; l is the traffic route; d is the departure frequency; TG is the traffic model built based on Traffic GPT; Ld is the real-time traffic data of the road network; Ca is the capacity analysis result of the transportation tool; Le is the passenger flow level; S is the weight of the factors affecting passenger flow;

[0025] S4: According to the results of traffic capacity allocation, dynamic adjustment and coordinated scheduling of capacity are realized: according to the output results of the traffic model in step S3, specifically: the number of taxis t, the number of online-hailing cars w and the number of buses b, the departure frequency d is controlled, the traffic routes l are planned and the stations p are configured; taxis, online-hailing cars and buses are dispatched to timely reduce the peak period Le3 and extreme peak period Le4 of passenger flow level to flat period Le2 or low peak period Le1, and complete the joint coordination and allocation of traffic capacity.

[0026] Preferably, the factors affecting passenger flow determined in step S1 specifically include: external factors and travel mode factors;

[0027] The external factors include: weather parameters, temperature parameters and date parameters;

[0028] The factors affecting travel modes include: route direction, station coverage, departure interval, operating speed, vehicle facilities and service operation parameters.

[0029] Preferably, the entropy weight method is used in step S1 to analyze the impact of passenger flow influencing factors on urban traffic passenger flow data, specifically:

[0030] S11: Standardize the factors affecting passenger flow; the method for determining the factors affecting positive passenger flow is:

[0031]

[0032] Among them, x ij is the original value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; x′ ij is the standardized value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; min is the minimum value function; max is the maximum value function; n is the n-city traffic passenger flow data in the evaluation system; i is the number of the city traffic passenger flow data;

[0033] The method for determining the factors affecting negative passenger flow is as follows:

[0034]

[0035] S12: Determine the proportion of each passenger flow influencing factor in the urban traffic passenger flow data; calculate the proportion of the j-th passenger flow influencing factor in the i-th city traffic passenger flow data, and regard it as the probability used in the relative entropy calculation:

[0036]

[0037] Among them, y ij is the proportion of the jth passenger flow influencing factor in the traffic passenger flow data of the i-th city; m is the m passenger flow influencing factors of the traffic passenger flow data of each city in the evaluation system, and j is the passenger flow influencing factor number;

[0038] S13: Calculate the information entropy of each passenger flow influencing factor; for the jth passenger flow influencing factor, the calculation formula of its information entropy is:

[0039]

[0040] Among them, e j is the information entropy value of the jth passenger flow influencing factor;

[0041] S14: Determine the urban traffic passenger flow data and the influence weights of each passenger flow influencing factor:

[0042]

[0043] Among them, w ij is the influence weight of the jth passenger flow influencing factor in the traffic passenger flow data of the i-th city;

[0044] S15: Comprehensively determine the weights of factors affecting passenger flow:

[0045]

[0046] Among them, S j is the weight of the jth passenger flow influencing factor.

[0047] Preferably, the long short-term memory network passenger flow prediction model in step S2 is a recurrent neural network; the forget gate controls the information of the unit state at the previous moment that needs to be forgotten; the input gate controls the information of the candidate state at the current moment that needs to be saved; the output gate controls the information of the internal state at the current moment that needs to be output to the outside.

[0048] Preferably, in step S22, the output data of the forget gate and the input gate are calculated to generate a candidate passenger flow data memory unit, which specifically includes the following sub-steps:

[0049] S221: Use the forget gate to control the information of factors affecting passenger flow in the current unit state; the output of the forget gate is a value between 0 and 1, 1 means using this information, 0 means not using this information, calculate the forget gate output:

[0050] f t =σ(W f ×[h t-1 , x t ]+b f );

[0051] Among them, f t is the output of the forget gate; σ is the sigmoid activation function; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous moment; x t is the current input connection; b f is the bias term of the forget gate;

[0052] S222: Using the input gate to control the influence of the factors affecting the passenger flow data, including two parts: the first layer of activation function sigmoid layer determines the passenger flow data information that needs to be updated, and the second layer of activation function tanh layer generates new candidate passenger flow data memory units; the input gate output is calculated as:

[0053] i t =σ(W i ×[h t-1 , x t ]+b i );

[0054] Among them, i t is the output of the input gate; W i is the weight matrix of the input gate; b i is the bias term of the input gate;

[0055] S223: Generate candidate passenger flow data memory unit as:

[0056]

[0057] in, is a new candidate memory unit; W c is the weight matrix of the candidate memory unit; b c is the bias term of the candidate memory unit;

[0058] S224: Generate the current passenger flow data memory unit state using the forget gate and the input gate.

[0059] Preferably, the output data of the output gate is calculated in step S23, specifically:

[0060] The output gate is used to control the output of the memory cell state at the current moment:

[0061] O t =σ(W o ×[h t-1 , x t ]+b o );

[0062] Among them, O t is the output of the output gate; W o is the weight matrix of the output gate; b o is the bias of the output gate.

[0063] Preferably, the transport capacity analysis result Ca of the transport tool in step S3 can ensure sufficient transport capacity based on the transport capacity allocation result of the traffic large model, including the required number of taxis A1, the equilibrium number of online car-hailing orders A2, and the number of vehicles N that should be configured for the bus line;

[0064] The method for obtaining the transportation capacity analysis result Ca of the transportation tool is:

[0065] Ca=A1+A2+N;

[0066] Among them, A1 is the number of taxis required; A2 is the equilibrium number of online ride-hailing orders; N is the number of vehicles that should be deployed on the bus line;

[0067] The required number of taxis A1 is:

[0068] A1=Qa÷(q1×T d ×N d ÷24);

[0069] Among them, Qa is the total passenger flow; q1 is the average daily carrying capacity of each taxi; T d N is the total daily operating time of the taxi; d is the average number of passengers carried per trip;

[0070] The equilibrium number of online ride-hailing orders A2 is determined by the revenue per vehicle and the average revenue per order, specifically:

[0071] A2=Sa÷Sd ;

[0072] Among them, Sa is the income of online car-hailing bicycles; S d The average revenue per online ride-hailing vehicle;

[0073] The number of vehicles N that should be deployed on the bus line is determined by the length of the bus line, the average operating speed of the buses, the proportion of rest time at the first and last stops, the departure frequency during peak hours, and the surplus coefficient of vehicles that cannot be operated due to maintenance, specifically:

[0074] N=(l N ×h N ×γ N )÷[v×(1-t N )];

[0075] Among them, l N is the length of the bus line; h N is the frequency of departure during peak hours; N is the surplus coefficient; v is the average operating speed of public transport vehicles; t N It is the proportion of rest time waiting for departure at the first and last stations.

[0076] Preferably, in step S4, the strategy adopted to reduce the peak period Le3 and the extreme peak period Le4 of the passenger flow level to the flat peak period Le2 or the low peak period Le1 is as follows:

[0077] When the passenger flow level is at the peak period Le3, guide empty taxis to the congested area, increase the number of taxis t, and increase the supply of transportation capacity; improve the transportation capacity of the online car-hailing platform, attract drivers to go online, increase the number of online car-hailing vehicles w, and shorten the waiting time for passengers; increase the density of bus trips, shorten the departure interval, and increase the number of buses b to cope with sudden passenger flow;

[0078] When the passenger flow level is at the extreme peak period Le4, start the taxi emergency dispatch mechanism, increase the number of taxis t, and give priority to the travel of passengers with emergency and special needs; start the online car-hailing carpooling mode, increase the vehicle passenger rate, encourage drivers to go online, increase the number of online car-hailing cars w, and increase the platform capacity; open temporary bus routes or extend the operating hours of existing routes, use big data analysis to predict passenger flow, increase the number of buses b, and deploy capacity in advance;

[0079] The control response strategy of increasing the number of taxis t, the number of online-hailing cars w, and the number of buses b reduces the peak and extreme peak periods of passenger flow levels to flat peak periods Le2 or low peak periods Le1.

[0080] The second aspect of the present invention proposes an instruction transmission system for a traffic capacity joint dispatching and allocation method based on passenger flow prediction, which includes: an instruction input module, an instruction parsing module, an instruction allocation module, an instruction transmission module, an execution and feedback module, and a monitoring and evaluation module;

[0081] The command input module inputs specific commands for passenger evacuation based on passenger flow forecast and transport capacity joint coordination analysis results, including: traffic signal control and vehicle dispatch;

[0082] After receiving the instruction, the instruction parsing module performs parsing processing, identifies the type, target, and parameter information of the instruction, and verifies the legitimacy and integrity of the instruction to ensure that the instruction is executed correctly;

[0083] The instruction distribution module distributes the instruction to the corresponding subsystem or execution unit according to the parsed instruction information;

[0084] The instruction transmission module transmits the instruction to the execution device via the communication network, and encrypts and compresses the instruction data to ensure the security and efficiency of the instruction during transmission;

[0085] After receiving the instruction, the execution and feedback module performs the corresponding operation, generates status information or execution results during the execution process, and feeds back the information to the upper level;

[0086] The monitoring and evaluation module monitors the execution process of the instruction to ensure that the execution effect meets expectations, and evaluates and optimizes the performance of the instruction transmission system based on the execution results and feedback information.

[0087] Compared with the prior art, the present invention has the following beneficial effects:

[0088] (1) The large passenger flow prediction based on the long short-term memory network model of the present invention is a method that uses deep learning technology to predict the changes in passenger flow in the future. It is particularly good at processing long-term dependencies in time series data. Therefore, it has significant advantages in the field of passenger flow prediction and can more scientifically arrange traffic capacity, optimize road network layout, and improve resource utilization efficiency, thereby improving travel experience and enhancing overall traffic operation efficiency.

[0089] (2) The present invention is based on the capacity coordination analysis of the traffic large model. Through intelligent means, taxis, online car-hailing vehicles, buses and other transport vehicles are reasonably dispatched according to different passenger flow levels to optimize the allocation of traffic resources and improve traffic operation efficiency.

[0090] (3) The present invention comprehensively uses modern information technology and management methods to achieve coordination and linkage in multiple aspects such as passenger flow monitoring, capacity scheduling, emergency response, and passenger services, which can effectively improve the operational efficiency of transportation hubs and the travel experience of passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is a flow chart of the transportation capacity joint coordination and allocation method based on passenger flow prediction of the present invention;

[0092] Figure 2 It is a flow chart of instruction transmission of the instruction transmission system of the present invention;

[0093] Figure 3 A flow chart of large passenger flow prediction using the long short-term memory network (LSTM) model of the present invention;

[0094] Figure 4 The flow chart of the transport capacity coordination based on the traffic big model (Traffic GPT) of the present invention;

[0095] Figure 5 A passenger flow data variation diagram of a certain passenger transport hub in a certain period of time of the present invention;

[0096] Figure 6 This is a comparison chart of the LSTM test set predicted value and actual value data of the present invention;

[0097] Figure 7 This is a screenshot of the main interface of the traffic operation management system of the present invention;

[0098] Figure 8 This is a screenshot of the parking monitoring interface of the traffic operation management system of the present invention. DETAILED DESCRIPTION

[0099] The exemplary embodiments, features and aspects of the present invention will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0100] The embodiment of the present invention provides a method for joint dispatching and allocating transportation capacity based on passenger flow prediction, such as Figure 1 As shown in the figure, the urban traffic passenger flow data is obtained, and the factors affecting the passenger flow are analyzed using the entropy weight method; a long short-term memory network passenger flow prediction model is constructed to predict the urban traffic passenger flow data; according to the passenger flow prediction data, the traffic capacity allocation based on the traffic large model is realized; according to the traffic capacity allocation results, the capacity is dynamically adjusted and coordinated. It includes:

[0101] Step S1: Obtain the urban traffic passenger flow data to obtain the actual value p of the urban traffic passenger flow data in , and use the entropy weight method to analyze the factors affecting passenger flow, specifically:

[0102] Summarize the passenger transport hub data, railway passenger transport data, airport passenger transport data and traffic checkpoint statistics, and use the above data to superimpose to obtain the real value p of urban traffic passenger flow data in .

[0103] Determine the factors affecting passenger flow, including external factors and travel mode factors; external factors include weather parameters, temperature parameters and date parameters; travel mode factors include route direction, station coverage, departure interval, operating speed, vehicle facilities and service operation parameters. Table 1 shows the weights assigned to each indicator influencing factor.

[0104] Table 1 Weights of factors affecting each indicator

[0105]

[0106] like Figure 5 The figure shows the passenger flow data change diagram of a certain time period of a certain passenger transport hub of the present invention; the experimental data uses the daily passenger flow data of a certain popular business district in a certain city, the data of the first time period from 11:00 to 13:00 is the training set, and the data of the second time period from 11:00 to 13:00 is the test set. The passenger flow data of the second time period is predicted by using the data of the first time period. The embodiment of the present invention lists the passenger flow data of August 1st for the whole day as follows: Figure 5 shown.

[0107] The entropy weight method is used to analyze the impact of passenger flow influencing factors on urban traffic passenger flow data, specifically:

[0108] Step S11: Standardize the factors affecting passenger flow. The method for determining the factors affecting positive passenger flow is:

[0109]

[0110] Among them, x ij is the original value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; x' ij is the standardized value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; min is the minimum value function; max is the maximum value function; n is the n-city traffic passenger flow data in the evaluation system; i is the city traffic passenger flow data number.

[0111] The method for determining the factors affecting negative passenger flow is as follows:

[0112]

[0113] Step S12: Determine the proportion of each passenger flow influencing factor in the urban traffic passenger flow data; calculate the proportion of the j-th passenger flow influencing factor in the i-th city traffic passenger flow data, and regard it as the probability used in the relative entropy calculation:

[0114]

[0115] Among them, y ij is the proportion of the jth passenger flow influencing factor in the traffic passenger flow data of the ith city; m is the m passenger flow influencing factors of the traffic passenger flow data of each city in the evaluation system, and j is the passenger flow influencing factor number.

[0116] Step S13: Calculate the information entropy of each passenger flow influencing factor; for the jth passenger flow influencing factor, the calculation formula of its information entropy is:

[0117]

[0118] Among them, e j is the information entropy value of the j-th passenger flow influencing factor.

[0119] Step S14: Determine the influence weights of the urban traffic passenger flow data and each passenger flow influencing factor as follows:

[0120]

[0121] Among them, w ij is the influence weight of the jth passenger flow influencing factor in the traffic passenger flow data of the i-th city.

[0122] Step S15: Comprehensively determine the weights of the factors affecting passenger flow:

[0123]

[0124] Among them, S j is the weight of the jth passenger flow influencing factor.

[0125] Step S2: Construct a long short-term memory network passenger flow prediction model to predict urban traffic passenger flow data, which specifically includes the following sub-steps:

[0126] Step S21: construct a long short-term memory network passenger flow prediction model including a forget gate, an input gate and an output gate, obtain the urban traffic passenger flow data in step S1, and input the long short-term memory network passenger flow prediction model according to the time series; Figure 3 The figure shows a flow chart of large passenger flow prediction using the long short-term memory network (LSTM) model of the present invention; the long short-term memory network passenger flow prediction model is a recurrent neural network; the forget gate controls the information of the unit state at the previous moment that needs to be forgotten; the input gate controls the information of the candidate state at the current moment that needs to be saved; the output gate controls the information of the internal state at the current moment that needs to be output to the outside.

[0127] Step S22: Calculate the output data of the forget gate and the input gate respectively to generate a candidate passenger flow data memory unit, which specifically includes the following sub-steps:

[0128] Step S221: Use the forget gate to control the factors affecting passenger flow in the current unit state. The output of the forget gate is a value between 0 and 1, 1 means using the information, and 0 means not using the information. Calculate the forget gate output:

[0129] f t =σ(W f ×[h t-1 , x t ]+b f );

[0130] Among them, f t is the output of the forget gate; σ is the sigmoid activation function; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous moment; x t is the current input connection; b f is the bias term of the forget gate.

[0131] Step S222: Using the input gate to control the influence of the influencing factors affecting the passenger flow data, including two parts: the first layer of activation function sigmoid layer determines the passenger flow data information that needs to be updated, and the second layer of activation function tanh layer generates new candidate passenger flow data memory units; the input gate output is calculated as:

[0132] i t =σ(W i ×[h t-1 , x t ]+b i );

[0133] Among them, i t is the output of the input gate; W i is the weight matrix of the input gate; b i is the bias term of the input gate.

[0134] Step S223: Generate candidate passenger flow data memory unit as:

[0135]

[0136] in, is a new candidate memory unit; W c is the weight matrix of the candidate memory unit; b c is the bias term of the candidate memory unit.

[0137] Step S224: Use the forget gate and the input gate to generate the current passenger flow data memory unit state:

[0138]

[0139] Among them, C t is the memory unit state at the current moment; f t is the output of the forget gate; C t-1 is the memory unit state at the previous moment; i t is the output of the input gate; is a new candidate memory cell.

[0140] Step S23: Calculate the output data of the output gate, and use the output gate to control the output of the memory unit state at the current moment:

[0141] O t =σ(W o ×[h t-1 , x t ]+b o );

[0142] Among them, O t is the output of the output gate; W o is the weight matrix of the output gate; b o is the bias of the output gate.

[0143] Determine the hidden state of the current passenger flow data memory unit as:

[0144] h t =O t ×tanh(C t );

[0145] Among them, h t is the hidden state at the current moment; O t is the output of the output gate; tanh is the second layer activation function.

[0146] Step S24: according to the hidden state of the passenger flow data memory unit, update the weight matrix and bias inside the long short-term memory network passenger flow prediction model; use the loss function of the long short-term memory network passenger flow prediction model to calculate the predicted value p of urban traffic passenger flow data out and the actual value p of the urban traffic passenger flow data in step S1 in The gap is:

[0147] Q=||p out -p in || 2 ≤Sq;

[0148] Where Q is the loss function of the long short-term memory network passenger flow prediction model; p out is the predicted value of urban traffic passenger flow data; p in is the true value of urban traffic passenger flow data, which is obtained using the long short-term memory network passenger flow prediction model; Sq is the loss function indicator of the long short-term memory network passenger flow prediction model.

[0149] like Figure 6 The figure shows the comparison of the predicted value and the actual value of the LSTM test set of the present invention. After the data is normalized, the parameters of the LSTM neural network are set, and the time step is set to 1, the prediction time is 2 hours, and the number of iterations is 100. Using the LSTM neural network, the comparison of the predicted value and the actual value is obtained. Figure 6 As shown, by comparison, it can be seen that the line graphs of the predicted value and the actual value are similar, proving that the prediction results of this scheme are accurate.

[0150] Step S25: Divide the passenger flow level into four levels according to the predicted value of the urban traffic passenger flow data: low peak period Le1, flat peak period Le2, peak period Le3 and extreme peak period Le4.

[0151] Step S3: Based on the urban traffic passenger flow forecast data in step S2, traffic capacity allocation based on the traffic big model is realized.

[0152] like Figure 4 The figure shows the flow chart of the transport capacity coordination based on the traffic big model (Traffic GPT) of the present invention; based on the traffic big model, the traffic capacity coordination analysis is performed, and according to the predicted value of the urban traffic passenger flow data in step S2, according to the passenger flow level and the passenger flow influencing factors in step S1, taxis, online car-hailing vehicles and buses are dispatched and routed. The formula of the traffic big model is specifically:

[0153] (t,w,b,p,l,d)=TG(p out , Ld, Ca, Le, S);

[0154] Among them, t is the number of taxis; w is the number of online-hailing taxis; b is the number of buses; p is the configured station; l is the traffic route; d is the departure frequency; TG is the traffic model built based on Traffic GPT; Ld is the real-time traffic data of the road network; Ca is the capacity analysis result of the vehicle; Le is the passenger flow level; and S is the weight of the factors affecting passenger flow.

[0155] The capacity analysis result Ca of transportation tools can ensure sufficient capacity for the transportation capacity allocation result based on the large traffic model, including the required number of taxis A1, the equilibrium number of online car-hailing orders A2, and the number of vehicles N that should be configured on the bus line.

[0156] The method for obtaining the transportation capacity analysis result Ca of the transportation tool is as follows:

[0157] Ca=A1+A2+N;

[0158] Among them, A1 is the required number of taxis; A2 is the equilibrium number of online-hailing taxis; and N is the number of vehicles that should be allocated on the bus route.

[0159] The number of taxis A1 required is:

[0160] A1=Qa÷(q1×T d ×N d ÷24);

[0161] Among them, Qa is the total passenger flow; q1 is the average daily carrying capacity of each taxi; T d N is the total daily operating time of the taxi; d is the average number of passengers per trip.

[0162] The equilibrium number of online ride-hailing orders A2 is determined by the revenue per vehicle and the average revenue per order, specifically:

[0163] A2=Sa÷S d ;

[0164] Among them, Sa is the income of online car-hailing bicycles; S d It is the average revenue per online ride-hailing vehicle.

[0165] The number of vehicles N that should be configured for a bus route is determined by the length of the bus route, the average operating speed of the buses, the proportion of rest time at the first and last stops, the departure frequency during peak hours, and the surplus coefficient of vehicles that cannot be operated due to maintenance. Specifically:

[0166] N=(l N ×h N ×γ N )÷[v×(1-t N )];

[0167] Among them, l N is the length of the bus line; h N is the frequency of departure during peak hours; N is the surplus coefficient; v is the average operating speed of public transport vehicles; t N It is the proportion of rest time waiting for departure at the first and last stations.

[0168] Step S4: According to the results of traffic capacity allocation, dynamic adjustment and coordinated scheduling of capacity are realized: according to the output results of the traffic model in step S3, specifically: the number of taxis t, the number of online-hailing cars w and the number of buses b, the departure frequency d is controlled, the traffic routes l are planned and the stations p are configured; taxis, online-hailing cars and buses are scheduled to timely reduce the peak period Le3 and the extreme peak period Le4 of the passenger flow level to the flat peak period Le2 or the low peak period Le1. The specific strategies adopted are:

[0169] When the passenger flow level is at the peak period Le3, guide empty taxis to the congested area, increase the number of taxis t, and increase the supply of transportation capacity; improve the transportation capacity of the online car-hailing platform, attract drivers to go online, increase the number of online car-hailing vehicles w, and shorten the waiting time for passengers; increase the density of bus trips, shorten the departure interval, and increase the number of buses b to cope with sudden passenger flow;

[0170] When the passenger flow level reaches the extreme peak period t4, the taxi emergency dispatch mechanism is activated, the number of taxis is increased t, and the travel of passengers with emergency and special needs is prioritized; the online car-hailing carpooling mode is started to increase the vehicle passenger load factor, encourage drivers to go online, increase the number of online car-hailing cars w, and increase the platform capacity; temporary bus routes are opened or the operating hours of existing routes are extended, and big data analysis is used to predict passenger flow, increase the number of buses b, and deploy capacity in advance;

[0171] The control response strategy of increasing the number of taxis t, the number of online-hailing taxis w, and the number of buses b reduces the peak period Le3 and the extreme peak period Le4 of the passenger flow level to the flat period Le2 or the low peak period Le1 to complete the joint coordination and allocation of traffic capacity.

[0172] The second aspect of the embodiment of the present invention proposes an instruction transmission system for a traffic capacity joint dispatching and allocation method based on passenger flow prediction, such as Figure 2 It is a flow chart of instruction transmission of the instruction transmission system of the present invention; it includes: an instruction input module, an instruction parsing module, an instruction allocation module, an instruction transmission module, an execution and feedback module and a monitoring and evaluation module.

[0173] The command input module inputs specific instructions for passenger evacuation based on passenger flow forecasts and capacity coordination analysis results, including: traffic signal control and vehicle scheduling.

[0174] After receiving the instruction, the instruction parsing module performs parsing processing, identifies the type, target, and parameter information of the instruction, and verifies the legality and integrity of the instruction to ensure the correct execution of the instruction.

[0175] The instruction allocation module allocates the instruction to the corresponding subsystem or execution unit according to the parsed instruction information.

[0176] The instruction transmission module uses the communication network to transmit instructions to the execution device and encrypts and compresses the instruction data to ensure the security and efficiency of the instructions during transmission.

[0177] After receiving the instruction, the execution and feedback module performs the corresponding operation. During the execution process, it generates status information or execution results and feeds this information back to the upper level.

[0178] The monitoring and evaluation module monitors the execution process of instructions to ensure that the execution effect meets expectations, and evaluates and optimizes the performance of the instruction transmission system based on the execution results and feedback information.

[0179] like Figure 7 Shown is a screenshot of the main interface of the traffic operation management system of the present invention. The traffic operation management system developed based on the traffic capacity coordination and allocation method and the command transmission system based on passenger flow prediction is used for comprehensive management and command coordination of the urban traffic system, to achieve optimal allocation of traffic capacity resources and improved management of traffic operation efficiency.

[0180] like Figure 8 The screenshot shown is a parking monitoring interface of the traffic operation management system of the present invention, which is a screenshot of the parking monitoring interface module of the traffic operation management system, and is mainly used to monitor the current vehicle usage and parking status of urban traffic, so as to facilitate the dispatch of vehicles in emergency situations.

[0181] The beneficial effects of the present invention are as follows: the present invention provides a method for joint dispatching and allocating transportation capacity based on passenger flow prediction and a command transmission system, which constructs a transportation capacity joint dispatching model for multiple modes of transportation, combines the dynamic adjustment and coordinated dispatching of taxis, online car-hailing vehicles, buses, and tourist buses, and realizes the optimal allocation of transportation resources and the improvement of transportation operation efficiency through means such as real-time data analysis, intelligent dispatching system, and cross-platform collaboration. The embodiment of the present invention uses the long-short-term memory network passenger flow prediction model and the traffic big model to perform intelligent optimization of transportation capacity joint dispatching, which can realize real-time monitoring and prediction of traffic flow, and the prediction results of this method are proved to be accurate through analysis of actual data. By using intelligent dispatching strategies, transportation vehicles such as taxis, online car-hailing vehicles, buses, and tourist buses are reasonably dispatched according to different passenger flow levels, optimizing the allocation of transportation resources and improving transportation operation efficiency. The transportation operation management system developed by the embodiment of the present invention has a good use effect in optimizing the allocation of transportation capacity resources and improving transportation operation efficiency.

[0182] The embodiments described above are only descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for joint dispatching and allocating transportation capacity based on passenger flow prediction, characterized in that: It includes: S1: Obtain the urban traffic passenger flow data to obtain the actual value p of the urban traffic passenger flow data in , and use the entropy weight method to analyze the factors affecting passenger flow, specifically: Summarize the passenger transport hub data, railway passenger transport data, airport passenger transport data and traffic checkpoint statistics, and use the above data to superimpose to obtain the real value p of urban traffic passenger flow data in ; Determine the factors affecting passenger flow and use the entropy weight method to analyze the impact of passenger flow factors on urban traffic passenger flow data; S2: Construct a long short-term memory network passenger flow prediction model to predict urban traffic passenger flow data, which includes the following sub-steps: S21: constructing a long short-term memory network passenger flow prediction model including a forget gate, an input gate and an output gate, obtaining the urban traffic passenger flow data in step S1, and inputting the long short-term memory network passenger flow prediction model according to the time series; S22: Calculate the output data of the forget gate and the input gate respectively, generate candidate passenger flow data memory units, and generate the current passenger flow data memory unit state as follows: Among them, C t is the memory unit state at the current moment; f t is the output of the forget gate; C t-1 is the memory unit state at the previous moment; i t is the output of the input gate; is a new candidate memory unit; S23: Calculate the output data of the output gate and determine the hidden state of the current passenger flow data memory unit: h t =O t ×tanh(C t ) Among them, h t is the hidden state at the current moment; O t is the output of the output gate; tanh is the second layer activation function; S24: Update the weight matrix and bias inside the long short-term memory network passenger flow prediction model according to the hidden state of the passenger flow data memory unit; calculate the predicted value p of the urban traffic passenger flow data using the loss function of the long short-term memory network passenger flow prediction model out and the actual value p of the urban traffic passenger flow data in step S1 in The gap is: Q=||p out -p in || 2 ≤Sq Where Q is the loss function of the long short-term memory network passenger flow prediction model; p out is the predicted value of urban traffic passenger flow data; pi n is the real value of urban traffic passenger flow data, obtained by using the long short-term memory network passenger flow prediction model; Sq is the loss function indicator of the long short-term memory network passenger flow prediction model; S25: Divide the passenger flow level into four levels according to the predicted value of urban traffic passenger flow data: low peak period Le1, flat peak period Le2, peak period Le3 and extreme peak period Le4; S3: according to the urban traffic passenger flow forecast data in step S2, realize the traffic capacity allocation based on the traffic big model; Based on the traffic big model, traffic capacity joint coordination analysis is carried out. According to the predicted value of urban traffic passenger flow data in step S2, according to the passenger flow level and the passenger flow influencing factors in step S1, taxis, online car-hailing vehicles and buses are dispatched and routed. The formula of the traffic big model is as follows: (t、w、b、p、l、d)=TG(pou t 、Ld、Ca、Le、S) Among them, t is the number of taxis; w is the number of online-hailing cars; b is the number of buses; p is the configured station; l is the traffic route; d is the departure frequency; TG is the traffic model built based on Traffic GPT; Ld is the real-time traffic data of the road network; Ca is the capacity analysis result of the transportation tool; Le is the passenger flow level; S is the weight of the factors affecting passenger flow; The transport capacity analysis result Ca of the transport tool in step S3 can ensure sufficient transport capacity based on the transport capacity allocation result of the traffic large model, including the required number of taxis A1, the equilibrium number of online car-hailing orders A2, and the number of vehicles N that should be configured for the bus line; The method for obtaining the transportation capacity analysis result Ca of the transportation tool is: Ca=A1+A2+N Among them, A1 is the number of taxis required; A2 is the equilibrium number of online ride-hailing orders; N is the number of vehicles that should be configured for the bus line: The required number of taxis A1 is: A1=Qa÷(q1×T d ×N d ÷24) Among them, Qa is the total passenger flow; q1 is the average daily carrying capacity of each taxi; T d N is the total operating time of the taxi every day; d is the average number of passengers carried per trip; The equilibrium number of online ride-hailing orders A2 is determined by the revenue per vehicle and the average revenue per order, specifically: <h2 style=";text-align:left;direction:ltr">A2=Sa÷S<h2 style=";text-align:left;direction:ltr"> d Among them, Sa is the income of online car-hailing bicycles; S d The average revenue per online ride-hailing vehicle; The number of vehicles N that should be deployed on the bus line is determined by the length of the bus line, the average operating speed of the buses, the proportion of rest time at the first and last stops, the departure frequency during peak hours, and the surplus coefficient of vehicles that cannot be operated due to maintenance, specifically: N=(l N ×h N ×γ N )÷[v×(1-t N )] Among them, l N is the length of the bus line; h N is the frequency of departure during peak hours; N is the surplus coefficient; v is the average operating speed of public transport vehicles; t N The proportion of rest time waiting for departure at the first and last stops; S4: According to the results of traffic capacity allocation, dynamic adjustment and coordinated scheduling of capacity are realized: according to the output results of the traffic model in step S3, specifically: the number of taxis t, the number of online-hailing cars w and the number of buses b, the departure frequency d is controlled, the traffic routes l are planned and the stations p are configured; taxis, online-hailing cars and buses are dispatched to timely reduce the peak period Le3 and extreme peak period Le4 of passenger flow level to flat period Le2 or low peak period Le1, and complete the joint coordination and allocation of traffic capacity.

2. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: In step S1, factors influencing passenger flow are determined, including: external factors and travel mode influencing factors; The external factors include: weather parameters, temperature parameters and date parameters; The factors affecting travel modes include: route direction, station coverage, departure interval, operating speed, vehicle facilities and service operation parameters.

3. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: The entropy weight method is used in step S1 to analyze the impact of passenger flow influencing factors on urban traffic passenger flow data, specifically: S11: Standardize the factors affecting passenger flow; the method for determining the factors affecting positive passenger flow is: Among them, x ij is the original value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; x′ ij is the standardized value of the jth passenger flow influencing factor of the i-th city traffic passenger flow data; min is the minimum value function; max is the maximum value function; n is the n-city traffic passenger flow data in the evaluation system; i is the number of the city traffic passenger flow data; The method for determining the factors affecting negative passenger flow is as follows: S12: Determine the proportion of each passenger flow influencing factor in the urban traffic passenger flow data; calculate the proportion of the j-th passenger flow influencing factor in the i-th city traffic passenger flow data, and regard it as the probability used in the relative entropy calculation: Among them, y ij is the proportion of the jth passenger flow influencing factor in the traffic passenger flow data of the i-th city; m is the m passenger flow influencing factors of the traffic passenger flow data of each city in the evaluation system, and j is the passenger flow influencing factor number; S13: Calculate the information entropy of each passenger flow influencing factor; for the jth passenger flow influencing factor, the calculation formula of its information entropy is: Among them, e j is the information entropy value of the j-th passenger flow influencing factor; S14: Determine the urban traffic passenger flow data and the influence weights of each passenger flow influencing factor: Among them, w ij is the influence weight of the jth passenger flow influencing factor in the traffic passenger flow data of the i-th city; S15: Comprehensively determine the weights of factors affecting passenger flow: Among them, S j is the weight of the jth passenger flow influencing factor.

4. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: The long short-term memory network passenger flow prediction model in step S2 is a recurrent neural network; the forget gate controls the information of the unit state at the previous moment that needs to be forgotten; the input gate controls the information of the candidate state at the current moment that needs to be saved; the output gate controls the information of the internal state at the current moment that needs to be output to the outside.

5. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: In step S22, the output data of the forget gate and the input gate are calculated to generate a candidate passenger flow data memory unit, which specifically includes the following sub-steps: S221: Use the forget gate to control the information of factors affecting passenger flow in the current unit state; the output of the forget gate is a value between 0 and 1, 1 means using this information, 0 means not using this information, calculate the forget gate output: f t =ρ(W f ×[h t-1 ,x t ]+b f ) Among them, f t is the output of the forget gate; σ is the sigmoid activation function; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous moment; x t is the current input connection; b f is the bias term of the forget gate; S222: Using the input gate to control the influence of the factors affecting the passenger flow data, including two parts: the first layer of activation function sigmoid layer determines the passenger flow data information that needs to be updated, and the second layer of activation function tanh layer generates new candidate passenger flow data memory units; the input gate output is calculated as: i t =σ(W i ×[h t-1 ,x t ]+b i ) Among them, i t is the output of the input gate; W i is the weight matrix of the input gate; b i is the bias term of the input gate; S223: Generate candidate passenger flow data memory unit as: in, is a new candidate memory unit; W c is the weight matrix of the candidate memory unit; b c is the bias term of the candidate memory unit; S224: Generate the current passenger flow data memory unit state using the forget gate and the input gate.

6. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: In step S23, the output data of the output gate is calculated, specifically: The output gate is used to control the output of the memory cell state at the current moment: The t =ρ(W o ×[h t-1 ,x t ]+b o ) Among them, O t is the output of the output gate; W o is the weight matrix of the output gate; b o is the bias of the output gate.

7. The method for joint dispatching and allocating transportation capacity based on passenger flow prediction according to claim 1 is characterized in that: In step S4, the strategy adopted to reduce the peak period Le3 and the extreme peak period Le4 of the passenger flow level to the flat peak period Le2 or the low peak period Le1 is as follows: When the passenger flow level is at the peak period Le3, guide empty taxis to the congested area, increase the number of taxis t, and increase the supply of transportation capacity; improve the transportation capacity of the online car-hailing platform, attract drivers to go online, increase the number of online car-hailing vehicles w, and shorten the waiting time for passengers; increase the density of bus trips, shorten the departure interval, and increase the number of buses b to cope with sudden passenger flow; When the passenger flow level is at the extreme peak period Le4, the taxi emergency dispatch mechanism is activated, the number of taxis t is increased, and the travel of passengers with emergency and special needs is prioritized; the online car-hailing carpooling mode is activated to increase the vehicle passenger rate, encourage drivers to go online, increase the number of online car-hailing cars w, and increase the platform capacity; Open temporary bus routes or extend the operating hours of existing routes, use big data analysis to predict passenger flow, increase the number of buses, and deploy transportation capacity in advance; The control response strategy of increasing the number of taxis t, the number of online-hailing cars w, and the number of buses b reduces the peak and extreme peak periods of passenger flow levels to flat peak periods Le2 or low peak periods Le1.

8. A command transmission system for the traffic capacity coordination and allocation method based on passenger flow prediction according to claim 1, characterized in that: It includes: Instruction input module, instruction parsing module, instruction allocation module, instruction transmission module, execution and feedback module and monitoring and evaluation module; The command input module inputs specific commands for passenger evacuation based on passenger flow forecast and transport capacity joint coordination analysis results, including: traffic signal control and vehicle dispatch; After receiving the instruction, the instruction parsing module performs parsing processing, identifies the type, target, and parameter information of the instruction, and verifies the legitimacy and integrity of the instruction to ensure that the instruction is executed correctly; The instruction distribution module distributes the instruction to the corresponding subsystem or execution unit according to the parsed instruction information: The instruction transmission module transmits the instruction to the execution device via the communication network, and encrypts and compresses the instruction data to ensure the security and efficiency of the instruction during transmission; After receiving the instruction, the execution and feedback module performs the corresponding operation, generates status information or execution results during the execution process, and feeds back the information to the upper level; The monitoring and evaluation module monitors the execution process of the instruction to ensure that the execution effect meets expectations, and evaluates and optimizes the performance of the instruction transmission system based on the execution results and feedback information.

Images