WO2022116447A1 - 公交车调度方法、装置、计算机设备及介质 - Google Patents

公交车调度方法、装置、计算机设备及介质 Download PDF

Info

Publication number
WO2022116447A1
WO2022116447A1 PCT/CN2021/084302 CN2021084302W WO2022116447A1 WO 2022116447 A1 WO2022116447 A1 WO 2022116447A1 CN 2021084302 W CN2021084302 W CN 2021084302W WO 2022116447 A1 WO2022116447 A1 WO 2022116447A1
Authority
WO
WIPO (PCT)
Prior art keywords
bus
information
station
subway
passenger
Prior art date
Application number
PCT/CN2021/084302
Other languages
English (en)
French (fr)
Inventor
王健宗
李泽远
Original Assignee
平安科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2022116447A1 publication Critical patent/WO2022116447A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/123Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams
    • GPHYSICS
    • G07CHECKING-DEVICES
    • G07CTIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C9/00Individual registration on entry or exit
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02TCLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00Road transport of goods or passengers
    • Y02T10/10Internal combustion engine [ICE] based vehicles
    • Y02T10/40Engine management systems

Definitions

  • the present application belongs to the technical field of intelligent transportation, and in particular, relates to a bus scheduling method, device, computer equipment and readable storage medium.
  • Public transportation is an important part of urban infrastructure. It has an increasing influence in social life, urban construction and economic development. Ensuring the smooth flow of public transportation is of great significance to safeguarding people's death and promoting regional development. At present, the inventor realizes that most areas cannot reasonably schedule public transportation according to the flow of people, resulting in insufficient bus capacity during peak hours, crowded vehicles, and even many passengers unable to get on the bus. If it is too high, there will be a waste of resources.
  • One of the purposes of the embodiments of the present application is to provide a bus scheduling method, device, computer equipment and readable storage medium, aiming to solve the technical problems of insufficient transport capacity of public transport vehicles during peak periods and high no-load rate during idle periods.
  • a first aspect of the embodiments of the present application provides a bus scheduling method, including:
  • bus station information includes location information of multiple bus stops
  • subway station information includes location information of multiple subway stations
  • the target vehicle number of the bus station is determined according to the passenger flow information of the bus station.
  • a second aspect of the embodiments of the present application provides a bus scheduling device, including:
  • a station information acquisition module configured to acquire bus station information and subway station information, where the bus station information includes location information of multiple bus stations, and the subway station information includes location information of multiple subway stations;
  • a station determination module configured to determine a target subway station associated with the bus station according to the position information of the bus station and the position information of the subway station;
  • a passenger information acquisition module configured to acquire the bus passenger information of the bus station and the subway passenger information of the target subway station;
  • a passenger flow determination module configured to determine the passenger flow information of the bus station according to the bus passenger information and the subway passenger information based on the trained passenger flow prediction model
  • the scheduling module is configured to determine the number of public transport vehicles bound for the bus station according to the passenger flow information of the bus station.
  • a third aspect of the embodiments of the present application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program When realized:
  • bus station information includes location information of multiple bus stops
  • subway station information includes location information of multiple subway stations
  • the target vehicle number of the bus station is determined according to the passenger flow information of the bus station.
  • a fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement:
  • bus station information includes location information of multiple bus stops
  • subway station information includes location information of multiple subway stations
  • the target vehicle number of the bus station is determined according to the passenger flow information of the bus station.
  • a fifth aspect of the embodiments of the present application further provides a computer program product, when the computer program product is executed on a computer device, the computer device is executed to realize:
  • bus station information includes location information of multiple bus stops
  • subway station information includes location information of multiple subway stations
  • the target vehicle number of the bus station is determined according to the passenger flow information of the bus station.
  • the embodiments of the present application include the following advantages:
  • the bus station information includes location information of multiple bus stops, and the subway station information includes location information of multiple subway stations; according to the location of the bus station information and the location information of the subway station to determine the target subway station associated with the bus station; obtain the bus passenger information of the bus station and the subway passenger information of the target subway station; based on the trained passenger flow prediction model, according to The bus passenger information and the subway passenger information determine the passenger flow information of the bus station; determine the target vehicle number of the bus station according to the passenger flow information of the bus station, so that the bus dispatch center can
  • the passenger flow information dispatches buses to solve the problem of insufficient capacity of buses during peak hours and too high no-load rate of buses during off-peak hours.
  • FIG. 1 is a schematic flowchart of a bus scheduling method provided by an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a method for training a passenger flow prediction model provided by an embodiment of the present application
  • FIG. 3 is a schematic diagram of a usage scenario of a method for training a passenger flow prediction model provided by an embodiment of the present application
  • FIG. 4 is a schematic structural block diagram of a bus dispatching device provided by an embodiment of the present application.
  • FIG. 5 is a schematic block diagram of the structure of an apparatus for training a passenger flow prediction model provided by an embodiment of the present application
  • FIG. 6 is a schematic structural block diagram of a computer device provided by an embodiment of the present application.
  • Embodiments of the present application also provide a bus scheduling method, apparatus, computer device, and computer-readable storage medium. It is used to predict the passenger flow of the bus based on the passenger flow prediction model, so as to determine the target vehicle of the bus, reduce the problem of the no-load rate and/or overload rate of the bus, and improve the capacity of the urban public transport.
  • the bus scheduling method can be used for a server, and of course also for a terminal, wherein the terminal can be an electronic device such as a tablet computer, a notebook computer, a desktop computer, etc.; the server can be, for example, a single server or a server cluster.
  • the following embodiments will be described in detail with the bus scheduling applied to the server.
  • FIG. 1 is a schematic flowchart of a bus scheduling method provided by an embodiment of the present application.
  • the bus scheduling method may include the following steps S110-S150.
  • Step S110 Acquire bus station information and subway station information, where the bus station information includes location information of multiple bus stops, and the subway station information includes location information of multiple subway stations.
  • bus station information and subway station information may be obtained from systems such as an urban construction system, a bus operation system, and a subway operation system.
  • bus station information and subway station information can also be obtained through the Internet of Things in a smart city.
  • the bus station information also includes the location information of the bus station, the name of the bus station, the longitude and latitude of the bus station, and it can be understood that the subway station information includes the position information of the subway station, the name of the subway station, and the latitude and longitude of the subway station.
  • the bus station information includes aa avenue and the main entrance station of bb park, the longitude and latitude are 112 degrees east longitude and 26 degrees north latitude, and the subway station information includes, aa avenue, bb park station, the longitude and latitude are 112 degrees east longitude and 26 degrees north latitude.
  • Step S120 Determine a target subway station associated with the bus station according to the position information of the bus station and the position information of the subway station, and the target subway station may be one or more.
  • the target subway station is determined according to the position information of the bus station and the position information of the subway station, and the subway station is associated with the bus station, that is, the passengers have a certain probability that they will be at the associated bus station after exiting the subway station. Take the bus.
  • the target subway station associated with the bus station may be determined according to the station name information of the bus station and the station name information of the subway station.
  • the station name information of a bus station is bb Park Main Gate Station
  • the station name information of a subway station is bb Park Station
  • the determining the target subway station associated with the bus station according to the position information of the bus station and the position information of the subway station includes: according to the position information of the bus station and the subway station The location information of the station is used, and a subway station located within a preset range of the geographic location of the bus station is determined as a target subway station associated with the bus station.
  • the location information of the bus station for example, at the main entrance of bb park, it may be determined that all subway stations within 1 km are the target subway stations.
  • the location information of the bus station it is also possible to determine, for example, five subway stations that are closest to the bus station as the target subway station.
  • Step S130 Acquire bus passenger information of the bus station and subway passenger information of the target subway station.
  • the passenger information may include the passenger number, the passenger transportation card number, the passenger's customary boarding time, the name of the boarding station, and the like.
  • the passenger's boarding time, boarding station, etc. can be determined according to the number of the traffic card that the passenger swipes when he takes the bus.
  • the passenger information may also be determined according to a monitoring camera device in the station.
  • passenger information may also be obtained at the bus operation center and/or the subway operation center.
  • Step S140 Based on the trained passenger flow prediction model, determine the passenger flow information of the bus station according to the bus passenger information and the subway passenger information.
  • the trained footfall prediction model can be stored in the blockchain node.
  • blockchain is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.
  • the passenger flow information of the bus station is predicted according to the trained passenger flow prediction model.
  • the passenger flow information of the bus station is predicted according to the historical passenger flow information of the bus station and the subway station.
  • the trained passenger flow prediction model includes: a forget gate, an input gate and an output gate; the trained passenger flow prediction model is based on the bus passenger information and the subway passenger information Determine the passenger flow information of the bus station, including:
  • the passenger flow information of the bus station is determined according to the passenger flow of the first and second bus stops and the passenger flow of the first and second subway stations.
  • the bus passenger information and the subway passenger information are input into the model, the bus passenger information and the subway passenger information are screened based on the forget gate, and the passenger flow of the first bus station and the passenger flow of the first subway station are determined.
  • the passenger information includes the boarding time and the boarding location, the passenger flow of the bus station and the subway station at the corresponding time is determined, and the passenger flow is carried out based on the forget gate. Filter to get the passenger flow of the first bus station and the passenger flow of the first subway station.
  • the expected predicted passenger flow information of the bus station is the passenger flow in the cc residential area during the off-peak hours, based on the forgotten gate, the passenger information for the bus station and the subway station, the rush-hour boarding time, and the rush-hour boarding time in the passenger information of the bus station and subway station.
  • the pick-up location in the cc residential area is deleted, and only the information of passengers who get on and off at the cc residential area is retained during the rush hour after get off work.
  • the bus passenger information be x k
  • the subway passenger information be P K
  • X K includes a series of historical time records ⁇ f(i,t),f(i,t-1),... ⁇ , so
  • the t is a parameter related to time
  • the estimated passenger flow is set to f(i, t+1)
  • P K and X K are input to the forget gate, and the screening operation is performed in the forget gate, and an output between 0 and 1 is output.
  • a value between 0 and 1 is output to determine how much information is retained, and 1 is used to represent all retention.
  • the information retained after the calculation is the passenger flow of the first bus station and the passenger flow of the first subway station.
  • an activation update operation is performed according to the passenger information to obtain the passenger flow of the second bus station and the passenger flow of the second subway station.
  • the useless information in the passenger information and the candidate information that needs to be added can be determined by using the Hadamard product.
  • the Hadamard product where f t ⁇ C t-1 represents the forgotten information, Indicates the candidate information to be added.
  • the state of the current computing unit is updated according to the forgotten information and the added candidate information.
  • the passenger flow information of the bus station is predicted based on the passenger flow of the first and second bus stations and the passenger flow of the first and second subway stations according to the updated unit state.
  • the passenger flow information of the bus station is determined by calculating the passenger flow of the first and second bus stations and the passenger flow of the first and second subway stations according to the tanh function.
  • tanh is the activation function, given by:
  • the predicted passenger flow information of the bus station is determined according to the calculation of the passenger flow of the first and second bus stations and the passenger flow of the first and second subway stations output by the output gate to the forget gate and the input gate.
  • Step S150 Determine the number of target vehicles at the bus station according to the passenger flow information of the bus station.
  • the bus dispatch center may predict the passenger flow of the bus station according to the predicted passenger flow information of the bus station, so as to determine the target number of vehicles at the bus station.
  • the passenger flow of the bus station during the off-peak period on Friday is significantly higher than the passenger flow during the off-peak period on Wednesday, and it is determined that the target number of vehicles on Friday is greater than the target number of vehicles on Wednesday.
  • determining the target number of vehicles at the bus station according to the passenger flow of the bus station can improve the utilization efficiency of the bus, and avoid the problems of too many passengers being unable to get on the bus and the high idling rate of the bus during the peak period.
  • the bus scheduling method further comprises: obtaining the passenger flow prediction model by training according to the training data.
  • FIG. 2 is a schematic flowchart of a training method for a passenger flow prediction model provided in an embodiment of the present application.
  • the training method is used for obtaining the passenger flow prediction model by training according to the training data.
  • the training method of the passenger flow prediction model includes steps S210-S250.
  • Step S210 Acquire training data, where the training data includes bus passenger information of a bus station and subway passenger information of a subway station associated with the geographic location of the bus station.
  • training data for training the passenger flow prediction model is obtained, where the training data includes information of bus passengers and information of subway passengers of a subway station associated with the geographic location of the bus station.
  • the training data may be passenger information of a bus station and passenger information of subway stations within 2 kilometers of the bus station.
  • the passenger information includes a boarding station and time, and according to the boarding station and time, the passenger flow of the corresponding station within a period of time can be determined.
  • X K represents the input feature of the training data, that is, passenger information of a bus station and/or a subway station, and X K includes historical information of passengers, such as historical bus stops and corresponding times.
  • Step S220 Based on the first passenger flow prediction model, screen, activate, update, and scale the bus passenger information to determine a first weight matrix and a first deviation.
  • the first passenger flow prediction model is used to screen, activate, update, and scale the passenger information of the bus station, and the first passenger flow prediction model is constructed based on a standard LSTM network.
  • an operation is performed on the bus passenger information based on a first passenger flow prediction model, and a first weight matrix and a first deviation are determined.
  • the first passenger flow prediction model includes a forget gate, an input gate, and an output gate; and based on the first passenger flow prediction model, the bus passenger information is screened, activated, updated, and scaled to determine the first passenger flow.
  • a weight matrix and the first deviation include: sub-steps S221-step S224.
  • Step S221 screen the bus passenger information based on the forget gate, determine the passenger flow of the first bus stop, and obtain the first forget gate weight matrix and the first forget gate deviation.
  • the first weight matrix and the first deviation are determined according to the passenger information of the bus stop.
  • filter information from the passenger information of the bus station according to the Forgotten Gate that is, retain the relevant historical information and stations.
  • the passenger information of the bus station includes the bb park main entrance station and the cc residential area station, and the boarding time is 9:00 a.m. , 6:00 pm, if the predicted passenger flow information of the bus is the passenger flow of the cc residential area during the off-duty peak period, then keep the cc residential area station and the information at 6:00 pm.
  • W f is determined as the first forgetting gate weight matrix
  • b f is determined as the first forgetting gate bias
  • Step S222 based on the input gate, perform activation and update processing on the bus passenger information, determine the passenger flow of the second bus stop, and obtain the first input gate weight matrix and the first input gate deviation,
  • the bus passenger information is scaled to determine the first output gate weight matrix and the first output gate deviation.
  • Step S223 based on the output gate, perform scaling processing on the bus passenger information, and determine the first output gate weight matrix and the first output gate deviation.
  • an activation update process is performed on the bus passenger information, the passenger flow of the second bus stop is determined, and the first input gate weight matrix and the first input gate deviation are obtained.
  • the passenger flow of the second bus station is determined according to the output of the forget gate, that is, the historical information extracted from the bus information and the bus passenger information after the input gate update processing, and the first input gate weight matrix can be obtained during the calculation process. and the first input gate bias.
  • the passenger flow of the second bus station is determined according to the output of the forget gate, that is, the historical information extracted from the bus information and the bus passenger information after the input gate update processing, and the first input gate weight matrix can be obtained in the calculation process. and the first input gate bias.
  • the sigmoid function can be used to determine the information that needs to be retained for activation, and the output value is 0 to 1, and 1 means to retain all the information, where the sigmoid function is
  • Exemplary, exemplary, tanh is the activation function, given by:
  • the weight matrix of the first input gate is Wi and W c
  • the deviations of the first input gate are b i and b c .
  • Step S224 determining the first weight matrix according to the first forget gate weight matrix, the first input gate weight matrix, and the first output gate weight matrix, and according to the first forget gate deviation, the The first input gate bias, the first output gate bias determine the first bias.
  • the first weight matrix and the first deviation are determined according to the weight matrix and deviation of each operation gate.
  • a weight matrix and a weight value corresponding to the deviation may be assigned to each operation gate, so as to determine the first weight matrix and the first deviation according to the weight value.
  • the weight matrix of the forget gate and the deviation of the input gate have the greatest impact on the operation result, and the weight value assigned to the weight matrix of the forget gate is higher than the weight value of the weight matrix of the input gate and the output gate. , and the bias given to the input gate is given a higher weight value than the bias of the forget gate and the output gate.
  • the weight matrix and the bias are determined to train the model, so that the prediction result of the model is closer to the actual situation.
  • Step S230 based on the second passenger flow prediction model, screen, activate, update, and scale the subway passenger information of the subway station to determine a second weight matrix and a second deviation.
  • the second passenger flow prediction model is used to screen, activate, update, and scale the passenger information of the subway station, and the second passenger flow prediction model is constructed based on a standard LSTM network.
  • each subway station has its own second passenger flow prediction model, and according to the respective second passenger flow prediction model, the passenger information in the respective subway station is screened, activated, updated, and scaled to determine the second weight matrix and the first weight matrix. Two deviations.
  • determining the second weight matrix and the second deviation based on the second passenger flow prediction model according to the subway passenger information screening, activation update, and scaling processing of the subway station includes: sub-steps S231 to S234.
  • Step S231 Perform screening processing on the subway passenger information based on the forget gate, determine the passenger flow of the first subway station, and obtain the second forget gate weight matrix and the second forget gate deviation.
  • Step S232 Activating and updating the subway passenger information based on the input gate, determining the passenger flow of the second subway station, and obtaining the second input gate weight matrix and the second input gate deviation.
  • Step S233 based on the output gate, perform scaling processing on the subway passenger information, and determine the second output gate weight matrix and the second output gate deviation.
  • Step S234 determining the second weight matrix according to the second forget gate weight matrix, the second input gate weight matrix, and the second output gate weight matrix, and according to the second forget gate deviation, the The second input gate bias, the second output gate bias determine the second bias.
  • Step S240 Determine update parameters according to the first weight matrix, the second weight matrix, the first deviation, and the second deviation.
  • the first and second weight matrices and the first and second deviations are determined according to the times of training.
  • the weight matrix and/or the deviation can be determined according to the training data with high correlation with factors that need to predict passenger flow, such as time, location, etc., and a higher weight can be assigned, so that when performing iterative calculation, focus on the correlation degree. higher training data.
  • the more training times the more accurate the determined first and second weight matrices and the first and second deviations, and the more accurate the determined update parameters.
  • the update parameter may also be determined according to the first and second weight matrices, the first and second deviations, and the distance between the target subway station and the bus station.
  • An update parameter is determined according to the distance.
  • FIG. 3 is a schematic diagram of a usage scenario of a method for training a passenger flow prediction model provided by an embodiment of the present application.
  • the determining the update parameter according to the first and second weight matrices and the first and second deviations includes: acquiring a second weight matrix determined by each subway station based on a second passenger flow prediction model and the second deviation; perform aggregation processing on the second weight matrix and the second deviation of each subway station to obtain the subway aggregation parameter; determine the update parameter according to the first weight matrix, the first deviation and the subway aggregation parameter.
  • the second weight matrix and the second deviation determined by each subway station are obtained. It can be understood that the passenger information between the subway stations is not publicly exchanged. Two deviations are calculated.
  • the second weight matrix and the second deviation may be aggregated according to the distance between the subway station and the bus station.
  • the second weight matrix and the second deviation account for a small proportion.
  • the second weight matrix and the second deviation may be aggregated according to whether there are buildings, such as shopping malls, office buildings, etc., between the subway station and the bus station.
  • the metro aggregation parameters are given by the formula OK, where, Also included is the first weight matrix and the first bias.
  • the update parameter of the bus station i is determined as:
  • Step S250 Adjust the parameters of the first passenger flow prediction model according to the updated parameters, and determine that the adjusted first passenger flow prediction model is the passenger flow prediction model.
  • the parameters of the first passenger flow prediction model are adjusted, and it is determined that the adjusted first passenger flow prediction model is the passenger flow prediction model.
  • different update parameters may be returned to perform parameter adjustment on the first passenger flow prediction model and the second passenger flow prediction model.
  • the parameter returned to the second passenger flow prediction model may be The second traffic prediction model is adjusted to make the second weight matrix and the second bias determined from the second traffic prediction model more accurate.
  • the training data includes bus passenger information of a bus station and subway passenger information of a subway station associated with the geographic location of the bus station;
  • the passenger information is screened, activated, updated, and scaled to determine the first weight matrix and the first deviation;
  • the subway passenger information of the subway station is screened, activated, updated, and scaled to determine the second weight matrix and the first deviation.
  • Two deviations determine update parameters according to the first and second weight matrices and the first and second deviations; adjust the parameters of the first passenger flow prediction model according to the update parameters, and determine the adjusted first
  • the passenger flow prediction model is the passenger flow prediction model.
  • the first passenger flow prediction model can be trained according to different passenger information, and the trained first passenger flow prediction model can be used as the passenger flow prediction model to more accurately predict the passenger flow of the bus station, so that the bus scheduling can be improved.
  • the center can schedule based on the predicted passenger flow to reduce problems such as bus overload and/or no-load.
  • FIG. 4 is a schematic diagram of a bus scheduling apparatus provided by an embodiment of the present application.
  • the bus scheduling apparatus may be configured in a server or a terminal to execute the aforementioned bus scheduling method.
  • the bus scheduling device includes: a station information acquisition module 110 , a station determination module 120 , a passenger information acquisition module 130 , a passenger flow determination module 140 , and a scheduling module 150 .
  • the station information acquisition module 110 is configured to acquire bus station information and subway station information, where the bus station information includes location information of multiple bus stops, and the subway station information includes location information of multiple subway stations.
  • the station determination module 120 is configured to determine the subway station as the target subway station associated with the bus station according to the position information of the bus station and the position information of the subway station.
  • the passenger information acquisition module 130 is configured to acquire the bus passenger information of the bus station and the subway passenger information of the target subway station.
  • the passenger flow determination module 140 is configured to determine the passenger flow information of the bus station according to the bus passenger information and the subway passenger information based on the trained passenger flow prediction model.
  • the scheduling module 150 is configured to determine the number of bus vehicles bound for the bus station according to the passenger flow information of the bus station.
  • the site determination module 120 includes an associated site determination sub-module.
  • An associated station determination submodule configured to determine a subway station located within a preset range of the geographic location of the bus station as being associated with the bus station according to the position information of the bus station and the position information of the subway station target subway station.
  • the passenger flow determination module 140 includes a forget gate submodule, an input gate submodule, and an output gate submodule.
  • the forget gate sub-module is used for screening the bus passenger information and the subway passenger information to obtain the passenger flow of the first bus station and the passenger flow of the first subway station.
  • the input gate sub-module is used to activate and update the bus passenger information and the subway passenger information, and determine the passenger flow and the first bus passenger flow of the second bus station according to the activated and updated bus passenger information and the subway passenger information. The passenger flow of the second subway station.
  • the output gate sub-module is used to determine the passenger flow of the bus station according to the passenger flow of the first bus station, the passenger flow of the second bus station, the passenger flow of the first subway station, and the passenger flow of the second subway station. Traffic information.
  • the bus dispatching device further includes: a passenger flow prediction model training device, which is configured to train the passenger flow prediction model according to the training data.
  • FIG. 5 is a schematic diagram of an apparatus for training a passenger flow prediction model according to an embodiment of the present application.
  • the apparatus for training a passenger flow prediction model may be configured in a server or a terminal to perform the aforementioned
  • a passenger flow prediction model training method is used to train the passenger flow prediction model according to the training data.
  • the passenger flow prediction model training device includes: a data acquisition module 210 , a first weight matrix and deviation determination module 220 , a second weight matrix and deviation determination module 230 , a parameter determination module 240 , and a parameter adjustment module 250 .
  • the data acquisition module 210 is configured to acquire training data, where the training data includes bus passenger information of a bus station and subway passenger information of a subway station associated with the geographic location of the bus station.
  • the first weight matrix and deviation determination module 220 is configured to screen, activate, update, and scale the bus passenger information based on the first passenger flow prediction model to determine the first weight matrix and the first deviation.
  • the second weight matrix and deviation determination module 230 is configured to determine, based on the second passenger flow prediction model, a second weight matrix and a second deviation for screening, activating, updating, and scaling the subway passenger information of the subway station.
  • a parameter determination module 240 configured to determine update parameters according to the first weight matrix, the second weight matrix, the first deviation, and the second deviation.
  • the parameter adjustment module 250 is configured to adjust the parameters of the first passenger flow prediction model according to the updated parameters, and determine that the adjusted first passenger flow prediction model is the passenger flow prediction model.
  • the first weight matrix and bias determination module 220 includes a forget gate submodule, an input gate submodule, an output gate submodule, and a synchronization determination module.
  • the forget gate sub-module is used to screen the bus passenger information, determine the passenger flow of the first bus station, and obtain the first forget gate weight matrix and the first forget gate deviation.
  • the input gate sub-module is used to activate and update the bus passenger information, determine the passenger flow of the second bus station, and obtain the first input gate weight matrix and the first input gate deviation.
  • the output gate sub-module is used for scaling the bus passenger information, and determining the first output gate weight matrix and the first output gate deviation.
  • a synchronization determination module configured to determine the first weight matrix according to the first forget gate weight matrix, the first input gate weight matrix, and the first output gate weight matrix, and determine the first weight matrix according to the first forget gate deviation , the first input gate deviation and the first output gate deviation determine the first deviation.
  • the second weight matrix and bias determination module 230 also includes a forget gate submodule, an input gate submodule, an output gate submodule, and a synchronization determination module.
  • the forget gate sub-module is used to screen the subway passenger information, determine the passenger flow of the first subway station, and obtain the second forget gate weight matrix and the second forget gate deviation.
  • the input gate sub-module is used to activate and update the subway passenger information, determine the passenger flow of the second subway station, and obtain the second input gate weight matrix and the second input gate deviation.
  • the output gate sub-module is used for scaling the subway passenger information, and determining the weight matrix of the second output gate and the deviation of the second output gate.
  • a synchronization determination module configured to determine the second weight matrix according to the second forget gate weight matrix, the second input gate weight matrix, and the second output gate weight matrix, and determine the second weight matrix according to the second forget gate bias , the second input gate deviation, and the second output gate deviation to determine the second deviation.
  • the parameter determination module 240 includes an aggregation parameter determination sub-module.
  • an aggregation parameter determination submodule configured to perform aggregation processing on the second weight matrix and the second deviation of each of the subway stations to obtain subway aggregation parameters; according to the first weight matrix, the first deviation and the The metro aggregation parameters described above determine the update parameters.
  • the methods and apparatus of the present application may be used in numerous general purpose or special purpose computing system environments or configurations.
  • the above-mentioned method and apparatus can be implemented in the form of a computer program, and the computer program can be executed on a computer device as shown in FIG. 6 .
  • FIG. 6 is a schematic diagram of a computer device provided by an embodiment of the present application.
  • the computer device can be a server or a terminal.
  • the computer device includes a processor, a memory, and a network interface connected through a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.
  • the nonvolatile storage medium can store operating systems and computer programs.
  • the computer program includes program instructions, which, when executed, can cause the processor to execute any bus scheduling method and/or any passenger flow prediction model training method.
  • the processor is used to provide computing and control capabilities to support the operation of the entire computer equipment.
  • the internal memory provides an environment for running a computer program in a non-volatile storage medium.
  • the computer program When executed by the processor, it can cause the processor to execute any bus scheduling method and/or any kind of passenger flow prediction model training. method.
  • the network interface is used for network communication, such as sending assigned tasks.
  • the structure of the computer device is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. More or fewer components are shown in the figures, either in combination or with different arrangements of components.
  • the processor may be a central processing unit (Central Processing Unit, CPU), and the processor may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated circuits) Circuit, ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor can be a microprocessor or the processor can also be any conventional processor or the like.
  • the processor is configured to run a computer program stored in the memory, so as to implement the following steps: acquiring bus stop information and subway station information, the bus stop information includes location information of multiple bus stops, The subway station information includes location information of a plurality of subway stations; the target subway station associated with the bus station is determined according to the location information of the bus station and the location information of the subway station; the bus station of the bus station is obtained Passenger information and subway passenger information of the target subway station; based on the trained passenger flow prediction model, determine the passenger flow information of the bus station according to the bus passenger information and the subway passenger information; The passenger flow information determines the target number of vehicles at the bus stop.
  • a subway station located within a preset range of the geographic location of the bus station is determined as a target subway station associated with the bus station.
  • the trained passenger flow prediction model includes: a forget gate, an input gate and an output gate; when the processor executes the computer program, it also implements:
  • the bus passenger information and the subway passenger information are screened to obtain the passenger flow of the first bus station and the passenger flow of the first subway station; based on the input gate, the bus passenger information and the subway The passenger information is activated and updated, and the passenger flow of the second bus station and the passenger flow of the second subway station are determined according to the activated and updated bus passenger information and the subway passenger information; based on the output gate, according to the first , the passenger flow of the second bus station and the passenger flow of the first and second subway stations to determine the passenger flow information of the bus station.
  • the training data includes bus passenger information of a bus station and subway passenger information of a subway station associated with the geographic location of the bus station; based on the first passenger flow prediction model, the bus passenger information is screened, The activation update and scaling process determine the first weight matrix and the first deviation; based on the second passenger flow prediction model, the subway passenger information of the subway station is screened, activated and updated, and the scaling process determines the second weight matrix and the second deviation; The first and second weight matrices and the first and second deviations are used to determine update parameters; the parameters of the first passenger flow prediction model are adjusted according to the update parameters, and the adjusted first passenger flow prediction model is determined as The passenger flow prediction model.
  • the first passenger flow prediction model includes: the forget gate, the input gate, and the output gate; when the processor executes the computer program, the processor further implements:
  • the bus passenger information is screened, the passenger flow of the first bus station is determined, and the first forget gate weight matrix and the first forget gate deviation are obtained; based on the input gate, the bus passenger information is activated and updated processing, determine the passenger flow of the second bus station and obtain the first input gate weight matrix and the first input gate deviation; based on the output gate, perform scaling processing on the bus passenger information, and determine the first output gate weight matrix and the first output gate gate bias; determine the first weight matrix according to the first forget gate weight matrix, the first input gate weight matrix, and the first output gate weight matrix, and determine the first weight matrix according to the first forget gate bias, the The first input gate bias, the first output gate bias determine the first bias.
  • the second passenger flow prediction model includes: the forget gate, the input gate, and the output gate; when the processor executes the computer program, the processor further implements:
  • the subway passenger information is screened, the passenger flow of the first subway station is determined, and the second forget gate weight matrix and the second forget gate deviation are obtained; based on the input gate, the subway passenger information is activated and updated Process, determine the passenger flow of the second subway station and obtain the second input gate weight matrix and the second input gate deviation; based on the output gate, perform scaling processing on the subway passenger information to determine the second output gate weight matrix and the second output gate gate bias; determine the second weight matrix according to the second forget gate weight matrix, the second input gate weight matrix, and the second output gate weight matrix, and determine the second weight matrix according to the second forget gate bias, the The second input gate bias, the second output gate bias determine the second bias.
  • a computer-readable storage medium where a computer program is stored in the computer-readable storage medium, the computer program includes program instructions, and the processor executes the program instructions to implement any one of the public transportation provided by the embodiments of the present application Vehicle scheduling method and/or any passenger flow prediction model training method. Specifically, when the computer program is executed by the processor, it realizes:
  • bus station information includes location information of multiple bus stops
  • subway station information includes location information of multiple subway stations
  • the target vehicle number of the bus station is determined according to the passenger flow information of the bus station.
  • a subway station located within a preset range of the geographic location of the bus station is determined as a target subway station associated with the bus station.
  • the trained passenger flow prediction model includes: a forget gate, an input gate and an output gate; when the computer program is executed by the processor, it also implements:
  • an activation and update process is performed on the bus passenger information and the subway passenger information, and the passenger flow and The passenger flow of the second subway station;
  • the passenger flow of the bus station is determined according to the passenger flow of the first bus station, the passenger flow of the second bus station, the passenger flow of the first subway station, and the passenger flow of the second subway station. Traffic information.
  • the training data includes bus passenger information of a bus station and subway passenger information of a subway station associated with the geographic location of the bus station;
  • the subway passenger information of the subway station is screened, activated, updated, and scaled to determine a second weight matrix and a second deviation;
  • the parameters of the first passenger flow prediction model are adjusted according to the updated parameters, and the adjusted first passenger flow prediction model is determined as the passenger flow prediction model.
  • the first passenger flow prediction model and the second passenger flow prediction model both include: the forget gate, the input gate, and the output gate; when the computer program is executed by the processor, it also implements:
  • the bus passenger information is screened to determine the passenger flow of the first bus stop, and the first forget gate weight matrix and the first forget gate deviation are obtained; based on the input gate, the bus passengers are The information is activated and updated, the passenger flow of the second bus station is determined, and the weight matrix of the first input gate and the deviation of the first input gate are obtained; based on the output gate, the bus passenger information is scaled and processed to determine the first output gate.
  • a weight matrix and a first output gate bias; the first weight matrix is determined according to the first forget gate weight matrix, the first input gate weight matrix, and the first output gate weight matrix, and the first weight matrix is determined according to the first A forget gate bias, the first input gate bias, the first output gate bias determine the first bias; and,
  • the subway passenger information is screened, the passenger flow of the first subway station is determined, and the second forget gate weight matrix and the second forget gate deviation are obtained; based on the input gate, the subway passenger The information is activated and updated, the passenger flow of the second subway station is determined, and the second input gate weight matrix and the second input gate deviation are obtained; based on the output gate, the subway passenger information is scaled and processed to determine the second output gate.
  • a weight matrix and a second output gate bias; the second weight matrix is determined according to the second forget gate weight matrix, the second input gate weight matrix, and the second output gate weight matrix, and the second weight matrix is determined according to the second The forget gate bias, the second input gate bias, and the second output gate bias determine the second bias.
  • An update parameter is determined according to the first weight matrix, the first deviation, and the metro aggregation parameter.
  • the computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as a hard disk or a memory of the computer device.
  • the computer-readable storage medium may be non-volatile or volatile.
  • the computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) ) card, Flash Card, etc.

Abstract

一种公交车调度方法、装置、计算机设备及可读存储介质,涉及智能交通技术领域。通过获取公交站信息和地铁站信息,公交站信息包括多个公交站的位置信息,地铁站信息包括多个地铁站的位置信息(S110);根据公交站的位置信息和地铁站的位置信息确定与公交站相关联的目标地铁站(S120);获取公交站的公交乘客信息和目标地铁站的地铁乘客信息(S130);基于训练好的客流量预测模型,根据公交乘客信息和地铁乘客信息确定公交站的客流量信息(S140);根据公交站的客流量信息确定公交站的目标车辆数目(S150),以解决高峰时期公交的运力不足且平峰时期公交的空载率过高的问题。还涉及区块链技术,训练好的客流量预测模型可以储存在区块链中。

Description

公交车调度方法、装置、计算机设备及介质
本申请要求于2020年12月01日在中国专利局提交的、申请号为202011389837.1、发明名称为“公交车调度方法、装置、计算机设备及介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请属于智能交通技术领域,尤其涉及一种公交车调度方法、装置、计算机设备及可读存储介质。
背景技术
公共交通运输是城市基础建设的重要组成部分,无论在社会生活、城市建设还是经济发展方面都有加大的影响力,保障公共交通的畅通,对于保障民生,促进地区发展具有重要意义。目前,发明人意识到,大多数的地区都无法根据人流量很合理地调度公共交通,导致在高峰时期公交运力不足,车内拥挤,甚至很多乘客无法上车,而在空闲时期公交空载率过高,出现资源浪费的问题。
技术问题
本申请实施例的目的之一在于:提供一种公交车调度方法、装置、计算机设备及可读存储介质,旨在解决公交车辆高峰时期运力不足且空闲时期空载率过高的技术问题。
技术解决方案
为解决上述技术问题,本申请实施例采用的技术方案是:
本申请实施例的第一方面提供了一种公交车调度方法,包括:
获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
本申请实施例的第二方面提供了一种公交车调度的装置,包括:
站点信息获取模块,用于获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
站点确定模块,用于根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
乘客信息获取模块,用于获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
客流量确定模块,用于基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
调度模块,用于根据所述公交站的客流量信息确定开往所述公交站的公交车辆数目。
本申请实施例的第三方面提供了一种计算机设备,包括:存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序,所述处理器执行所述计算机程序时实现:
获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述 公交站的客流量信息;
根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
本申请实施例的第四方面提供了一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行时实现:
获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
本申请实施例的第五方面还提供了一种计算机程序产品,当所述计算机程序产品在计算机设备上运行时,使得所述计算机设备执行时实现:
获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
有益效果
与现有技术相比,本申请实施例包括以下优点:
本申请实施例,通过获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;根据所述公交站的位置信息和所述地铁站的位置信息确定与公交站相关联的目标地铁站;获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;根据所述公交站的客流量信息确定所述公交站的目标车辆数目,以使公交车调度中心能根据预测的客流量信息调度公交车,以解决高峰时期公交的运力不足且平峰时期公交的空载率过高的问题。
附图说明
为了更清楚地说明本申请实施例中的技术方案,下面将对实施例或示范性技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其它的附图。
图1是本申请一实施例提供的一种公交车调度方法的流程示意图;
图2是本申请一实施例提供的一种客流量预测模型训练方法的流程示意图;
图3是本申请一实施例提供的一种客流量预测模型训练方法的使用场景示意图;
图4是本申请一实施例提供的一种公交车调度装置的结构示意框图;
图5是本申请一实施例提供的一种客流量预测模型训练装置的结构示意框图;
图6是本申请一实施例提供的一种计算机设备的结构示意框。
本发明的实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本申请一部分实施例,而不是全部的实施例。基于本申请 中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
附图中所示的流程图仅是示例说明,不是必须包括所有的内容和操作/步骤,也不是必须按所描述的顺序执行。例如,有的操作/步骤还可以分解、组合或部分合并,因此实际执行的顺序有可能根据实际情况改变。另外,虽然在装置示意图中进行了功能模块的划分,但是在某些情况下,可以以不同于装置示意图中的模块划分。
本申请的实施例还提供了一种公交车调度方法、装置、计算机设备及计算机可读存储介质。用于基于客流量预测模型预测公交车的客流量,以确定所述公交车的目标车辆,减少公交的空载率和/或过载率问题,提升城市公共交通的运力。
其中,该公交车调度方法可以用于服务器,当然也可以用于终端,其中,终端可以是平板电脑、笔记本电脑、台式电脑等电子设备;服务器例如可以为单独的服务器或服务器集群。但为了便于理解,以下实施例将以应用于服务器的公交车调度进行详细介绍。
下面结合附图,对本申请的一些实施方式作详细说明。在不冲突的情况下,下述的实施例及实施例中的特征可以相互组合。
请参阅图1,图1是本申请实施例提供的一种公交车调度方法的示意流程图。
如图1所示,该公交车调度方法可以包括以下步骤S110-步骤S150。
步骤S110、获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息。
示例性的,可以从城市建设系统、公交车运营系统、地铁运营系统等系统中获取公交站信息和地铁站信息。
示例性的,还可以通过智慧城市的物联网获取公交站信息和地铁站信息。
示例性的,公交站信息还包括公交站的位置信息,公交站名称,公交站的经度以及纬度,可以理解的,地铁站信息包括地铁站的位置信息,地铁站名称,地铁站的经纬度。
例如,公交站信息包括,aa大道,bb公园正门站,经纬度为东经112度北纬26度,地铁站信息包括,aa大道,bb公园站,经纬度为东经112度,北纬26度。
步骤S120、根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站,目标地铁站可以为一个或多个。
示例性的,根据公交站的位置信息和地铁站的位置信息确定目标地铁站,所述地铁站与所述公交站相关联,即乘客从地铁站出来后有一定的概率会在关联的公交站搭乘公交。
示例性的,可以根据公交站的站名信息与地铁站的站名信息确定与所述公交站相关联的目标地铁站。
例如,公交站的站名信息为bb公园正门站,地铁站的站名信息为bb公园站,可以确定该地铁站为与所述公交站关联的目标地铁站。
在一些实施例中,所述根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站包括:根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
示例性的,可以根据公交站的位置信息,例如在bb公园正门,确定1公里内的地铁站都为所述目标地铁站。
示例性的,还可以根据公交站的位置信息,确定最接近所述公交站的个如5个地铁站为目标地铁站。
步骤S130、获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息。
示例性的,所述乘客信息可以包括乘客编号、乘客交通卡编号、乘客的习惯上车时间、上车车站名称等。
可以根据乘客乘车时,刷的交通卡的编号确定乘客的上车时间、上车车站等。
示例性的,还可以根据站内的监控摄像装置,确定所述乘客信息。
示例性的,还可以在公交运营中心和/或地铁运营中心获取乘客信息。
步骤S140、基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息。
在一些实施方式中,训练好的客流量预测模型可以储存在区块链节点中。其中,区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
示例性的,根据训练好的客流量预测模型,预测公交站的客流量信息。
示例性的,根据公交站和地铁站的历史客流量信息预测公交站的客流量信息。
在一些实施例中,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;所述基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息,包括:
基于遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理得到第一公交站的客流量和第一地铁站的客流量;
基于输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息确定第二公交站的客流量和第二地铁站的客流量;
基于输出门,根据所述第一、第二公交站的客流量和所述第一、第二的地铁站客流量确定所述公交站的客流量信息。
示例性的,根据公交乘客信息和地铁乘客信息输入模型后,基于遗忘门对公交乘客信息和地铁乘客信息进行筛选,确定第一公交站的客流量和第一地铁站的客流量。
示例性的,根据多个公交乘客信息和地铁乘客信息,所述乘客信息包括上车时间和上车地点,确定对应时间的公交站和地铁站的客流量,基于遗忘门对所述客流量进行筛选得到第一公交站的客流量和第一地铁站的客流量。
例如,期望预测的公交站的客流量信息为,下班高峰期时间在cc住宅区的客流量,基于遗忘门,对公交站和地铁站的乘客信息中的上班高峰期上车时间,以及不位于cc住宅区的上车地点给予删除,只保留下班高峰期时间以及在cc住宅上下车的乘客信息。
示例性的,设公交乘客信息为x k,地铁乘客信息为P K,X K中包括一系列的历史时间记录{f(i,t),f(i,t-1),…},所述t为与时间相关的参数,将要预估的客流设为f(i,t+1),P K和X K输入遗忘门,在遗忘门中进行筛选运算,输出一个0到1之间的数值,以确定是否保留X K和P K中的信息,其中,筛选运算公式为f t=σ(W f[h t-1,X t]+b f,其中h t-1表示前一时间的输出,X t是当前时间的输入,即根据历史的输出记录以及当前输入确定需要筛选的信息。
示例性的,筛选运算后,输出一个0至1之间的数值以确定保留多少信息,1用于表示全部保留。运算之后保留的信息即为第一公交站的客流量和第一地铁站的客流量。
示例性的,基于输入门,根据乘客信息进行激活更新运算得到第二公交站的客流量和第二地铁站的客流量。
示例性的,基于输入门,对乘客信息进行激活和更新运算,所述激活运算为i t=σ(W i[h t-1,X t]+b i);所述更新运算为
Figure PCTCN2021084302-appb-000001
所述激活运算是为了确定乘客信息中的信息来进行更新,所述更新运算是根据乘客的历史时间信息产生新的待选 信息,根据所述乘客信息中的信息和待选信息确定第二公交站的客流量和第二地铁站的客流量。
示例性的,基于遗忘门和输入门,可以确定乘客信息中的无用信息以及需要添加的待选信息,利用哈达玛积(Hadamard乘积)确定。
示例性的,所述哈达玛积:
Figure PCTCN2021084302-appb-000002
其中,f t⊙C t-1表示遗忘的信息,
Figure PCTCN2021084302-appb-000003
表示添加的待选信息。
示例性的,根据遗忘的信息以及添加的待选信息更新当前计算单元的状态。
示例性的,根据更新后的单元状态对第一、第二公交站的客流量和第一、第二地铁站的客流量预测公交站的客流量信息。
示例性的,根据tanh函数对第一、第二公交站的客流量和第一、第二地铁站的客流量进行计算,确定所述公交站的客流量信息。
示例性的,根据o t=σ(W o[h t-1,X t]+b o),以及h t=o t⊙tanh(C t)进行运算,得到的h t即为预测的所述公交站的客流量信息。
示例性的,tanh为激活函数,由下式给出:
Figure PCTCN2021084302-appb-000004
示例性的,根据输出门对遗忘门以及输入门输出的第一、第二公交站的客流量和第一、第二地铁站的客流量运算,确定预测的所述公交站的客流量信息。
步骤S150、根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
示例性的,公交调度中心可以根据预测的所述公交站的客流量信息,预测公交站的客流量,以确定所述公交站的目标车辆数目。
例如,所述公交站周五的下班高峰期的客流明显高于周三的下班高峰期的客流量,确定周五的目标车辆数目大于周三的目标车辆数目。
示例性的,根据公交站的客流量确定所述公交站的目标车辆数目可以提高公交车的利用效率,避免乘客过多无法上车以及在平峰时期公交车空载率过高的问题。
在一些实施方式中,所述公交车调度方法还包括:根据训练数据训练得到所述客流量预测模型。
请结合前述实施例参阅图2,图2是本申请一实施例中提供的一种客流量预测模型的训练方法的流程示意图。所述训练方法用于根据训练数据训练得到所述客流量预测模型。
客流量预测模型的训练方法包括步骤S210-步骤S250。
步骤S210、获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息。
示例性的,获取训练所述客流量预测模型的训练数据,所述训练数据包括公交乘客的信息以及所述公交站地理位置相关联的地铁站的地铁乘客的信息。
示例性的,训练数据可以是,一个公交站的乘客信息,以及所述公交站在2公里内的地铁站的乘客信息。
示例性的,乘客信息包括上车站点和时间,根据所述上车站点和时间,可以确定对应站点在一段时间内的客流量。
示例性的,训练数据可以表示为D i={(x 1,y 1),(x 2,y 2),...,(x k,y k)},x k∈X i,y k∈Y i
示例性的,X K表示训练数据的输入特征,即公交站和/或地铁站的乘客信息,X K包含 乘客的历史信息,如历史乘车站点以及对应的时间。
步骤S220、基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差。
示例性的,第一客流量预测模型用于对公交站的乘客信息进行筛选、激活、更新、缩放处理,基于标准的LSTM网络构建所述第一客流量预测模型。
示例性的,根据公交乘客信息,基于第一客流量预测模型对所述公交乘客信息进行运算,确定第一权重矩阵和第一偏差。
在一些实施例中,所述第一客流量预测模型包括遗忘门、输入门和输出门;所述基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差,包括:子步骤S221-步骤S224。
步骤S221、基于遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差。
示例性的,基于遗忘门,根据公交站的乘客信息确定第一权重矩阵和第一偏差。
例如,根据遗忘门从公交站的乘客信息中筛选信息,即保留相关的历史信息和站点,如公交站的乘客信息中包括bb公园正门站、cc住宅区站,上车时间分别是上午9点、下午6点,若期望预测的公交的客流量信息为cc住宅区的下班高峰期的客流量,则保留cc住宅区站,下午6点的信息即可。
示例性的,根据运算公式f t=σ(W f[h t-1,X t]+b f确定所述第一遗忘门权重矩阵以及第一遗忘门偏差。
示例性的,在计算过程中,W f确定为第一遗忘门权重矩阵、b f确定为第一遗忘门偏差。
步骤S222、基于输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差、
示例性的,根据输出门,对公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差。
示例性的,根据tanh激活函数,对所述公交乘客信息进行缩放处理,所述缩放处理由o t=σ(W o[h t-1,X t]+b o确定,根据缩放处理计算中的W o确定为第一输出门权重矩阵,以及b o确定第一输出门偏差。
步骤S223、基于输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差。
示例性的,基于输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵和第一输入门偏差。
示例性的,根据遗忘门的输出,即在公交信息中提取的历史信息和输入门更新处理后的公交乘客信息确定第二公交站的客流量,在计算过程中可以得到第一输入门权重矩阵以及第一输入门偏差。
示例性的,根据遗忘门的输出,即在公交信息中提取的历史信息和输入门更新处理后的公交乘客信息确定第二公交站的客流量,在计算过程中可以得到第一输入门权重矩阵以及第一输入门偏差。
示例性的,激活处理为i t=σ(W i[h t-1,X t]+b i)以及更新
Figure PCTCN2021084302-appb-000005
示例性的,可以先通过sigmoid函数确定激活需要保留的信息,输出数值为0至1,1表示保留全部信息,其中sigmoid函数为
Figure PCTCN2021084302-appb-000006
示例性,示例性的,tanh为激活函数,由下式给出:
Figure PCTCN2021084302-appb-000007
示例性的,在计算过程中,根据激活处理和更新处理,确定第一输入门权重矩阵为W i和W c,第一输入门偏差为b i和b c
步骤S224、根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差。
示例性的,根据各个运算门的权重矩阵以及偏差确定第一权重矩阵以及第一偏差。
示例性的,可以赋予各个运算门的权重矩阵以及偏差对应的权重值,以根据权重值确定第一权重矩阵以及第一偏差。
示例性的,多次学习运算后,发现遗忘门的权重矩阵以及输入门的偏差对运算结果的影响最大,赋予遗忘门的权重矩阵的权重值高于输入门和输出门的权重矩阵的权重值,以及赋予输入门的偏差的权重值高于遗忘门和输出门的偏差的权重值。
示例性的,确定权重矩阵以及偏差对模型进行训练,以使模型预测结果更为接近实际情况。
步骤S230、基于第二客流量预测模型,对地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差。
示例性的,所述第二客流量预测模型用于对所述地铁站的乘客信息进行筛选、激活更新、缩放处理,基于标准的LSTM网络构建所述第二客流量预测模型。
示例性的,各地铁站都有各自的第二客流量预测模型,根据各自的第二客流量预测模型,对各自地铁站内的乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差。
在一些实施例中,所述基于第二客流量预测模型,根据地铁站的地铁乘客信息筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差,包括:子步骤S231-步骤S234。
步骤S231、基于遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差。
步骤S232、基于输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差。
步骤S233、基于输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差。
步骤S234、根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
示例性的,步骤S231-步骤S234的具体实施步骤可以参考上述步骤S221-步骤S224的具体实施步骤,在此不予撰述。
步骤S240、根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数。
示例性的,根据训练的次数确定第一、第二权重矩阵和第一、第二偏差。
例如,将若干次训练确定的所述第一权重矩阵、所述第二权重矩阵迭代计算,以及将若干次训练确定的所述第一偏差、第二偏差迭代计算,根据迭代计算的结果确定更新参数。
示例性的,可以根据与需要预测客流量的相关因素,如时间、地点等相关度较高的训练数据确定权重矩阵和/或偏差,赋予较高的权重,以进行迭代计算时,侧重相关度较高的训练数据。
示例性的,训练次数越多,确定的第一、第二权重矩阵以及第一、第二偏差会越准确,以确定的更新参数更精确。
示例性的,还可以根据所述第一、第二权重矩阵和所述第一、第二偏差以及所述目标地铁站与所述公交站的距离确定更新参数。
示例性的,若目标地铁站与公交站距离较远,乘客不去所述公交站换成公交的概率较大。根据所述距离确定更新参数。
如图3所示,图3是本申请一实施例提供的一种客流量预测模型训练方法的使用场景示意图。
在一些实施例中,所述根据所述第一、第二权重矩阵和所述第一、第二偏差确定更新参数,包括:获取各地铁站基于第二客流量预测模型确定的第二权重矩阵和第二偏差;对所述各地铁站的第二权重矩阵和第二偏差进行聚合处理得到地铁聚合参数;根据所述第一权重矩阵、第一偏差以及地铁聚合参数确定更新参数。
示例性的,获取各地铁站确定的第二权重矩阵以及第二偏差,可以理解的,各地铁站之间的乘客信息并不公开交换,因此需要通过服务器对地铁站的第二权重矩阵和第二偏差进行计算。
示例性的,可以根据地铁站与所述公交站的距离对第二权重矩阵以及第二偏差进行聚合处理,如若地铁站离所述公交站距离较远,则聚合处理时,所述地铁站的第二权重矩阵以及第二偏差占比较小。
示例性的,可以根据所述地铁站与所述公交站间有无建筑物如商场、写字楼等对第二权重矩阵以及第二偏差进行聚合处理。
示例性的,地铁聚合参数由公式
Figure PCTCN2021084302-appb-000008
确定,其中,
Figure PCTCN2021084302-appb-000009
还包括所述第一权重矩阵以及所述第一偏差。
示例性的,所述根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数,确定公交站i的更新参数为:
Figure PCTCN2021084302-appb-000010
步骤S250、根据所述更新参数对第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
示例性的,根据在步骤S240中确定的更新参数,对所述第一客流量预测模型的参数进行调整,以及确定调整后的第一客流量预测模型为所述客流量预测模型。
在一些实施例中,可以返回不同的更新参数对所述第一客流量预测模型和第二客流量预测模型进行参数调整。
示例性的,返回给第二客流量预测模型的参数可以是
Figure PCTCN2021084302-appb-000011
对第二客流量预测模型进行调整,以使从第二客流量预测模型确定的第二权重矩阵和第二偏差更准确。
示例性的,通过获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;基于第二客流量预测模型,对地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;根据所述第一、第二权重矩阵和所述第一、第二偏差确定更新参数;根据所述更新参数对第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。可以根据不同的乘客信息对第一客流量预测模型进行训练,并将训练后的第一客流量预测模型作为所述客流量预测模型,以更加准确地预测公交站的客流量, 以使公交调度中心可以根据预测的客流量进行调度,减少公交过载和/或空载等问题。
请参阅图4,图4是本申请一实施例提供的一种公交车调度装置的示意图,该公交车调度装置可以配置于服务器或终端中,用于执行前述的公交车调度方法。
如图4所示,该公交车调度装置,包括:站点信息获取模块110、站点确定模块120、乘客信息获取模块130、客流量确定模块140、调度模块150。
站点信息获取模块110,用于获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息。
站点确定模块120,用于根据所述公交站的位置信息和所述地铁站的位置信息将地铁站确定为与公交站相关联的目标地铁站。
乘客信息获取模块130,用于获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息。
客流量确定模块140,用于基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息。
调度模块150,用于根据所述公交站的客流量信息确定开往所述公交站的公交车辆数目。
示例性的,站点确定模块120包括关联站点确定子模块。
关联站点确定子模块,用于根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
示例性的,客流量确定模块140包括遗忘门子模块、输入门子模块、输出门子模块。
遗忘门子模块,用于对所述公交乘客信息和所述地铁乘客信息进行筛选处理得到第一公交站的客流量和第一地铁站的客流量。
输入门子模块,用于对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息确定第二公交站的客流量和第二地铁站的客流量。
输出门子模块,用于根据所述第一公交站的客流量、所述第二公交站的客流量和所述第一地铁站的客流量、第二地铁站的客流量确定所述公交站的客流量信息。
在一些实施方式中,所述公交车调度装置还包括:客流量预测模型训练装置,客流量预测模型训练装置用于根据训练数据训练得到所述客流量预测模型。
请结合前述实施例参阅图5,图5是本申请一实施例提供的一种客流量预测模型训练装置的示意图,该客流量预测模型训练装置可以配置于服务器或终端中,用于执行前述的客流量预测模型训练方法,以根据训练数据训练得到所述客流量预测模型。
如图5所示,该客流量预测模型训练装置,包括:数据获取模块210、第一权重矩阵与偏差确定模块220、第二权重矩阵与偏差确定模块230、参数确定模块240、参数调整模块250。
数据获取模块210,用于获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息。
第一权重矩阵与偏差确定模块220,用于基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理处理确定第一权重矩阵和第一偏差。
第二权重矩阵与偏差确定模块230,用于基于第二客流量预测模型,对所述地铁站的地铁乘客信息筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差。
参数确定模块240,用于根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数。
参数调整模块250,用于根据所述更新参数对所述第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
示例性的,第一权重矩阵与偏差确定模块220包括遗忘门子模块、输入门子模块、输 出门子模块以及同步确定模块。
遗忘门子模块,用于对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差。
输入门子模块,用于对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差。
输出门子模块,用于对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差。
同步确定模块,用于根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差。
示例性的,第二权重矩阵与偏差确定模块230也包括遗忘门子模块、输入门子模块、输出门子模块及同步确定模块。
遗忘门子模块,用于所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差。
输入门子模块,用于对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差。
输出门子模块,用于对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差。
同步确定模块,用于根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
示例性的,参数确定模块240包括聚合参数确定子模块。
聚合参数确定子模块,用于对各所述地铁站的所述第二权重矩阵和所述第二偏差进行聚合处理得到地铁聚合参数;根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数确定更新参数。
需要说明的是,所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的装置和各模块、单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
本申请的方法、装置可用于众多通用或专用的计算系统环境或配置中。例如:个人计算机、服务器计算机、手持设备或便携式设备、平板型设备、多处理器系统、基于微处理器的系统、机顶盒、可编程的消费电子设备、网络PC、小型计算机、大型计算机、包括以上任何系统或设备的分布式计算环境等等。
示例性地,上述的方法、装置可以实现为一种计算机程序的形式,该计算机程序可以在如图6所示的计算机设备上运行。
请参阅图6,图6是本申请实施例提供的一种计算机设备的示意图。该计算机设备可以是服务器或终端。
如图6所示,该计算机设备包括通过系统总线连接的处理器、存储器和网络接口,其中,存储器可以包括非易失性存储介质和内存储器。
非易失性存储介质可存储操作系统和计算机程序。该计算机程序包括程序指令,该程序指令被执行时,可使得处理器执行任意一种公交车调度方法和/或任意一种客流量预测模型训练方法。
处理器用于提供计算和控制能力,支撑整个计算机设备的运行。
内存储器为非易失性存储介质中的计算机程序的运行提供环境,该计算机程序被处理器执行时,可使得处理器执行任意一种公交车调度方法和/或任意一种客流量预测模型训练方法。
该网络接口用于进行网络通信,如发送分配的任务等。本领域技术人员可以理解,该 计算机设备的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备的限定,具体的计算机设备可以包括比图中所示更多或更少的部件,或者组合某些部件,或者具有不同的部件布置。
应当理解的是,处理器可以是中央处理单元(Central Processing Unit,CPU),该处理器还可以是其他通用处理器、数字信号处理器(Digital Signal Processor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现场可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。其中,通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
其中,在一些实施方式中,所述处理器用于运行存储在存储器中的计算机程序,以实现如下步骤:获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
示例性地,所述处理器执行所述计算机程序时还实现:
根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
示例性地,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;所述处理器执行所述计算机程序时还实现:
基于遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理得到第一公交站的客流量和第一地铁站的客流量;基于输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息确定第二公交站的客流量和第二地铁站的客流量;基于输出门,根据所述第一、第二公交站的客流量和所述第一、第二地铁站客流量确定所述公交站的客流量信息。
示例性的,所述处理器执行所述计算机程序时还实现:
获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;基于第二客流量预测模型,对地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;根据所述第一、第二权重矩阵和所述第一、第二偏差确定更新参数;根据所述更新参数对第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
示例性的,所述第一客流预测模型包括:所述遗忘门、所述输入门和所述输出门;所述处理器执行所述计算机程序时还实现:
基于遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差;基于输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差;基于输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差;根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差。
示例性的,所述第二客流预测模型包括:所述遗忘门、所述输入门和所述输出门;所述处理器执行所述计算机程序时还实现:
基于遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差;基于输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差;基 于输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差;根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
示例性的,所述处理器执行所述计算机程序时还实现:
获取各地铁站基于第二客流量预测模型确定的第二权重矩阵和第二偏差;对所述各地铁站的第二权重矩阵和第二偏差进行聚合处理得到地铁聚合参数;根据所述第一权重矩阵、第一偏差以及地铁聚合参数确定更新参数。
通过以上的实施方式的描述可知,本领域的技术人员可以清楚地了解到本申请可借助软件加必需的通用硬件平台的方式来实现。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在存储介质中,如ROM/RAM、磁碟、光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本申请各个实施例或者实施例的某些部分所述的方法,如:
一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序中包括程序指令,所述处理器执行所述程序指令,实现本申请实施例提供的任意一项公交车调度方法和/或任意一项客流量预测模型训练方法。具体的,计算机程序被处理器执行时实现:
获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
示例性的,所述计算机程序被处理器执行时还实现:
根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
示例性的,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;所述计算机程序被处理器执行时还实现:
基于所述遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理,得到第一公交站的客流量和第一地铁站的客流量;
基于所述输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息,确定第二公交站的客流量和第二地铁站的客流量;
基于所述输出门,根据所述第一公交站的客流量、所述第二公交站的客流量和所述第一地铁站的客流量、第二地铁站的客流量确定所述公交站的客流量信息。
示例性的,所述计算机程序被处理器执行时还实现:
获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;
基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;
基于第二客流量预测模型,对所述地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;
根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新 参数;
根据所述更新参数对所述第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
示例性的,所述第一客流预测模型和所述第二客流预测模型均包括:所述遗忘门、所述输入门和所述输出门;所述计算机程序被处理器执行时还实现:
基于所述遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差;基于所述输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差;基于所述输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差;根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差;以及,
基于所述遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差;基于所述输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差;基于所述输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差;根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
示例性的,所述计算机程序被处理器执行时还实现:
对各所述地铁站的所述第二权重矩阵和所述第二偏差进行聚合处理得到地铁聚合参数;
根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数确定更新参数。
其中,所述计算机可读存储介质可以是前述实施例所述的计算机设备的内部存储单元,例如所述计算机设备的硬盘或内存。所述计算机可读存储介质可以是非易失性,也可以是易失性。所述计算机可读存储介质也可以是所述计算机设备的外部存储设备,例如所述计算机设备上配备的插接式硬盘,智能存储卡(Smart Media Card,SMC),安全数字(Secure Digital,SD)卡,闪存卡(Flash Card)等。
以上所述,仅为本申请的具体实施方式,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到各种等效的修改或替换,这些修改或替换都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应以权利要求的保护范围为准。

Claims (20)

  1. 一种公交车调度方法,其中,包括以下步骤:
    获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
    根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
    获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
    基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
    根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
  2. 如权利要求1所述的公交车调度方法,其中,所述根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站,包括:
    根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
  3. 如权利要求1或2所述的公交车调度方法,其中,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;
    所述基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息,包括:
    基于所述遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理,得到第一公交站的客流量和第一地铁站的客流量;
    基于所述输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息,确定第二公交站的客流量和第二地铁站的客流量;
    基于所述输出门,根据所述第一公交站的客流量、所述第二公交站的客流量和所述第一地铁站的客流量、第二地铁站的客流量确定所述公交站的客流量信息。
  4. 如权利要求3所述的公交车调度方法,其中,在所述获取公交站信息和地铁站信息之前,所述公交车调度方法还包括:
    获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;
    基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;
    基于第二客流量预测模型,对所述地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;
    根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数;
    根据所述更新参数对所述第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
  5. 如权利要求4所述的公交车调度方法,其中,所述第一客流预测模型和所述第二客流预测模型均包括:所述遗忘门、所述输入门和所述输出门;
    所述基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差,包括:
    基于所述遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差;
    基于所述输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量 以及得到第一输入门权重矩阵以及第一输入门偏差;
    基于所述输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差;
    根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差;
    所述基于第二客流量预测模型,根据地铁站的地铁乘客信息筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差,包括:
    基于所述遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差;
    基于所述输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差;
    基于所述输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差;
    根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
  6. 如权利要求4所述的公交车调度方法,其中,所述根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数,包括:
    对各所述地铁站的所述第二权重矩阵和所述第二偏差进行聚合处理得到地铁聚合参数;
    根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数确定更新参数。
  7. 一种公交车调度的装置,其中,包括:
    站点信息获取模块,用于获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
    站点确定模块,用于根据所述公交站的位置信息和所述地铁站的位置信息将地铁站确定为与公交站相关联的目标地铁站;
    乘客信息获取模块,用于获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
    客流量确定模块,用于基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
    调度模块,用于根据所述公交站的客流量信息确定开往所述公交站的公交车辆数目。
  8. 如权利要求7所述的公交车调度的装置,其中,所述装置还包括如下模块:
    数据获取模块,用于获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;
    第一权重矩阵与偏差确定模块,用于基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理处理确定第一权重矩阵和第一偏差;
    第二权重矩阵与偏差确定模块,用于基于第二客流量预测模型,对所述地铁站的地铁乘客信息筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;
    参数确定模块,用于根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数;
    参数调整模块,用于根据所述更新参数对第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
  9. 一种计算机设备,其中,所述计算机设备包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序,所述处理器执行所述计算机程序时实现:
    获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地 铁站信息包括多个地铁站的位置信息;
    根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
    获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
    基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
    根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
  10. 根据权利要求9所述的计算机设备,其中,所述处理器执行所述计算机程序时还实现:
    根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
  11. 根据权利要求9或10所述的计算机设备,其中,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;所述处理器执行所述计算机程序时还实现:
    基于所述遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理,得到第一公交站的客流量和第一地铁站的客流量;
    基于所述输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息,确定第二公交站的客流量和第二地铁站的客流量;
    基于所述输出门,根据所述第一公交站的客流量、所述第二公交站的客流量和所述第一地铁站的客流量、第二地铁站的客流量确定所述公交站的客流量信息。
  12. 根据权利要求11所述的计算机设备,其中,所述处理器执行所述计算机程序时还实现:
    获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;
    基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;
    基于第二客流量预测模型,对所述地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;
    根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数;
    根据所述更新参数对所述第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
  13. 根据权利要求12所述的计算机设备,其中,所述第一客流预测模型和所述第二客流预测模型均包括:所述遗忘门、所述输入门和所述输出门;所述处理器执行所述计算机程序时还实现:
    基于所述遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差;以及,基于所述遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差;
    基于所述输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差;以及,基于所述输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差;
    基于所述输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差;以及,基于所述输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差;
    根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差;以及,根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
  14. 根据权利要求12所述的计算机设备,其中,所述处理器执行所述计算机程序时还实现:
    对各所述地铁站的所述第二权重矩阵和所述第二偏差进行聚合处理得到地铁聚合参数;
    根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数确定更新参数。
  15. 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,其中,所述计算机程序被处理器执行时实现:
    获取公交站信息和地铁站信息,所述公交站信息包括多个公交站的位置信息,所述地铁站信息包括多个地铁站的位置信息;
    根据所述公交站的位置信息和所述地铁站的位置信息确定与所述公交站相关联的目标地铁站;
    获取所述公交站的公交乘客信息和所述目标地铁站的地铁乘客信息;
    基于训练好的客流量预测模型,根据所述公交乘客信息和所述地铁乘客信息确定所述公交站的客流量信息;
    根据所述公交站的客流量信息确定所述公交站的目标车辆数目。
  16. 根据权利要求15所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时还实现:
    根据所述公交站的位置信息和所述地铁站的位置信息,将位于所述公交站地理位置的预设范围内的地铁站确定为与所述公交站相关联的目标地铁站。
  17. 根据权利要求15或16所述的计算机可读存储介质,其中,所述训练好的客流量预测模型,包括:遗忘门、输入门和输出门;所述计算机程序被处理器执行时还实现:
    基于所述遗忘门,对所述公交乘客信息和所述地铁乘客信息进行筛选处理,得到第一公交站的客流量和第一地铁站的客流量;
    基于所述输入门,对所述公交乘客信息和所述地铁乘客信息进行激活更新处理,以及根据激活更新后的所述公交乘客信息和所述地铁乘客信息,确定第二公交站的客流量和第二地铁站的客流量;
    基于所述输出门,根据所述第一公交站的客流量、所述第二公交站的客流量和所述第一地铁站的客流量、第二地铁站的客流量确定所述公交站的客流量信息。
  18. 根据权利要求17所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时还实现:
    获取训练数据,所述训练数据包括公交站的公交乘客信息和与所述公交站地理位置相关联的地铁站的地铁乘客信息;
    基于第一客流量预测模型,对所述公交乘客信息进行筛选、激活更新、缩放处理确定第一权重矩阵和第一偏差;
    基于第二客流量预测模型,对所述地铁站的地铁乘客信息进行筛选、激活更新、缩放处理确定第二权重矩阵和第二偏差;
    根据所述第一权重矩阵、所述第二权重矩阵和所述第一偏差、所述第二偏差确定更新参数;
    根据所述更新参数对所述第一客流量预测模型的参数进行调整,并确定调整后的第一客流量预测模型为所述客流量预测模型。
  19. 根据权利要求18所述的计算机可读存储介质,其中,所述第一客流预测模型和所 述第二客流预测模型均包括:所述遗忘门、所述输入门和所述输出门;所述计算机程序被处理器执行时还实现:
    基于所述遗忘门,对所述公交乘客信息进行筛选处理,确定第一公交站的客流量以及得到第一遗忘门权重矩阵以及第一遗忘门偏差;以及,基于所述遗忘门,对所述地铁乘客信息进行筛选处理,确定第一地铁站的客流量以及得到第二遗忘门权重矩阵以及第二遗忘门偏差;
    基于所述输入门,对所述公交乘客信息进行激活更新处理,确定第二公交站的客流量以及得到第一输入门权重矩阵以及第一输入门偏差;以及,基于所述输入门,对所述地铁乘客信息进行激活更新处理,确定第二地铁站的客流量以及得到第二输入门权重矩阵以及第二输入门偏差;
    基于所述输出门,对所述公交乘客信息进行缩放处理,确定第一输出门权重矩阵以及第一输出门偏差;以及,基于所述输出门,对所述地铁乘客信息进行缩放处理,确定第二输出门权重矩阵以及第二输出门偏差;
    根据所述第一遗忘门权重矩阵、所述第一输入门权重矩阵、所述第一输出门权重矩阵确定所述第一权重矩阵,以及根据所述第一遗忘门偏差、所述第一输入门偏差、所述第一输出门偏差确定所述第一偏差;以及,根据所述第二遗忘门权重矩阵、所述第二输入门权重矩阵、所述第二输出门权重矩阵确定所述第二权重矩阵,以及根据所述第二遗忘门偏差、所述第二输入门偏差、所述第二输出门偏差确定所述第二偏差。
  20. 根据权利要求18所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时还实现:
    对各所述地铁站的所述第二权重矩阵和所述第二偏差进行聚合处理得到地铁聚合参数;
    根据所述第一权重矩阵、所述第一偏差以及所述地铁聚合参数确定更新参数。
PCT/CN2021/084302 2020-12-01 2021-03-31 公交车调度方法、装置、计算机设备及介质 WO2022116447A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011389837.1A CN112562378B (zh) 2020-12-01 2020-12-01 公交车调度方法、装置、计算机设备及介质
CN202011389837.1 2020-12-01

Publications (1)

Publication Number Publication Date
WO2022116447A1 true WO2022116447A1 (zh) 2022-06-09

Family

ID=75047666

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/084302 WO2022116447A1 (zh) 2020-12-01 2021-03-31 公交车调度方法、装置、计算机设备及介质

Country Status (2)

Country Link
CN (1) CN112562378B (zh)
WO (1) WO2022116447A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115409295A (zh) * 2022-11-01 2022-11-29 深圳市城市交通规划设计研究中心股份有限公司 一种基于瓶颈分析的公交调度方法、电子设备及存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112562378B (zh) * 2020-12-01 2023-04-18 平安科技(深圳)有限公司 公交车调度方法、装置、计算机设备及介质
CN113284362A (zh) * 2021-05-19 2021-08-20 中国联合网络通信集团有限公司 公交车辆调度方法、装置、设备及存储介质
CN114693495B (zh) * 2022-05-24 2022-08-23 成都秦川物联网科技股份有限公司 智慧城市公共交通管理方法、物联网系统、装置及介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150348068A1 (en) * 2014-05-28 2015-12-03 International Business Machines Corporation Predicting waiting passenger count and evaluation
CN107358045A (zh) * 2017-07-12 2017-11-17 东南大学 一种评价地铁与常规公交换乘效率的流程与方法
CN110956328A (zh) * 2019-11-30 2020-04-03 天津市市政工程设计研究院 一种大客流影响的轨道交通站点公交接驳规模预测方法
CN110991794A (zh) * 2019-10-28 2020-04-10 上海城市交通设计院有限公司 一种城市轨道、公交两网融合水平评价方法
CN111079875A (zh) * 2019-12-17 2020-04-28 广州交通信息化建设投资营运有限公司 基于多源数据的公共交通客流监测方法、装置和存储介质
CN112562378A (zh) * 2020-12-01 2021-03-26 平安科技(深圳)有限公司 公交车调度方法、装置、计算机设备及介质

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090207049A1 (en) * 2008-02-15 2009-08-20 Chunghwa United Television Co., Ltd. Method for smart announcing of bus stop
US20130041941A1 (en) * 2010-04-09 2013-02-14 Carnegie Mellon University Crowd-Sourcing of Information for Shared Transportation Vehicles
US8306848B1 (en) * 2011-06-06 2012-11-06 International Business Machines Corporation Estimation of transit demand models for enhancing ridership
CN104318113A (zh) * 2014-10-29 2015-01-28 中国科学院深圳先进技术研究院 基于乘客换乘的时空特征推算乘客上车站点的方法及系统
CN105469602B (zh) * 2015-12-31 2017-08-11 北京航空航天大学 一种基于ic卡数据的公交乘客候车时间范围的预测方法
CN106652434B (zh) * 2016-12-02 2019-04-30 东南大学 一种基于轨道交通协调的公交调度方法
CN107273999B (zh) * 2017-04-27 2020-12-08 北京交通大学 一种突发事件下城市轨道交通客流预测方法
CN107545320B (zh) * 2017-07-03 2021-07-06 北京交通大学 一种基于图论的城市轨道交通乘客路径规划方法及系统
CN107529651B (zh) * 2017-08-18 2020-10-16 北京航空航天大学 一种基于深度学习的城市交通客流预测方法和设备
KR101974109B1 (ko) * 2017-12-21 2019-04-30 그제고스 말레비치 출발지 위치로부터 목적지 위치로의 여정에 대한 루트 또는 루트 소요 시간을 제공하기 위한 방법 및 컴퓨터 시스템
CN109887267B (zh) * 2019-03-21 2021-04-30 华侨大学 一种轨道交通共线段常规公交调整方法
CN111640294B (zh) * 2020-04-27 2022-02-11 河海大学 一种新建地铁线路影响下城市公交线路客流变化预测方法
CN111785017B (zh) * 2020-05-28 2022-04-15 博泰车联网科技(上海)股份有限公司 一种公交车调度方法、装置及计算机存储介质
CN111932925B (zh) * 2020-07-09 2023-06-09 中咨数据有限公司 一种公共交通站点的出行客流的确定方法、装置及系统
CN111985710B (zh) * 2020-08-18 2023-08-25 深圳诺地思维数字科技有限公司 一种公交乘客出行站点预测方法、存储介质及服务器

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150348068A1 (en) * 2014-05-28 2015-12-03 International Business Machines Corporation Predicting waiting passenger count and evaluation
CN107358045A (zh) * 2017-07-12 2017-11-17 东南大学 一种评价地铁与常规公交换乘效率的流程与方法
CN110991794A (zh) * 2019-10-28 2020-04-10 上海城市交通设计院有限公司 一种城市轨道、公交两网融合水平评价方法
CN110956328A (zh) * 2019-11-30 2020-04-03 天津市市政工程设计研究院 一种大客流影响的轨道交通站点公交接驳规模预测方法
CN111079875A (zh) * 2019-12-17 2020-04-28 广州交通信息化建设投资营运有限公司 基于多源数据的公共交通客流监测方法、装置和存储介质
CN112562378A (zh) * 2020-12-01 2021-03-26 平安科技(深圳)有限公司 公交车调度方法、装置、计算机设备及介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115409295A (zh) * 2022-11-01 2022-11-29 深圳市城市交通规划设计研究中心股份有限公司 一种基于瓶颈分析的公交调度方法、电子设备及存储介质
CN115409295B (zh) * 2022-11-01 2023-02-14 深圳市城市交通规划设计研究中心股份有限公司 一种基于瓶颈分析的公交调度方法、电子设备及存储介质

Also Published As

Publication number Publication date
CN112562378A (zh) 2021-03-26
CN112562378B (zh) 2023-04-18

Similar Documents

Publication Publication Date Title
WO2022116447A1 (zh) 公交车调度方法、装置、计算机设备及介质
CN109872535B (zh) 一种智慧交通通行预测方法、装置及服务器
US20200050938A1 (en) Systems and methods for improvement of index prediction and model building
US20150100238A1 (en) Large scale demand responsive transit framework
CN108830504B (zh) 用车需求预测方法、系统、服务器及计算机存储介质
US20150269491A1 (en) Predicting arrival times of vehicles based upon observed schedule adherence
CN110570678B (zh) 一种公交车辆从起点到终点总行驶时间预测方法及装置
US20180314998A1 (en) Resource Allocation in a Network System
CN113538067B (zh) 一种基于机器学习的城际网约车需求预测方法及系统
US11345251B2 (en) Priority-based energy transfer
US20230122069A1 (en) Distance-based energy transfer from a transport
US11790289B2 (en) Systems and methods for managing dynamic transportation networks using simulated future scenarios
CN116013059A (zh) 一种营运车辆的调度方法、装置、电子设备及存储介质
CN112669604B (zh) 城市交通调度方法及装置
US11507896B2 (en) Method and system for spatial-temporal carpool dual-pricing in ridesharing
CN111008729B (zh) 一种迁移预测的方法和装置
CN112230663A (zh) 车辆定位数据的监测方法、装置
CN112669603B (zh) 基于大数据的城市交通协同方法及装置
WO2021121206A1 (zh) 一种用于判定服务事故的责任的方法和系统
CN113269339A (zh) 一种网约车任务自动创建和分发的方法及系统
TWI838746B (zh) 電動載具的充電需求預估方法及其預估系統
CN114580787B (zh) 一种基于公交地铁接驳站点的客流量预测方法及装置
US20230145366A1 (en) Operation management apparatus, system, and operation management method
US20230017054A1 (en) Method and system for bundling data insights
WO2021136147A1 (zh) 一种共享设备的预设区域定位方法和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21899491

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21899491

Country of ref document: EP

Kind code of ref document: A1