CN110211379A - A kind of public transport method for optimizing scheduling based on machine learning - Google Patents

A kind of public transport method for optimizing scheduling based on machine learning Download PDF

Info

Publication number
CN110211379A
CN110211379A CN201910459783.2A CN201910459783A CN110211379A CN 110211379 A CN110211379 A CN 110211379A CN 201910459783 A CN201910459783 A CN 201910459783A CN 110211379 A CN110211379 A CN 110211379A
Authority
CN
China
Prior art keywords
agent model
disaggregation
time
website
training
Prior art date
Legal status (The legal status 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 status listed.)
Pending
Application number
CN201910459783.2A
Other languages
Chinese (zh)
Inventor
孙哲人
黄玉划
蔡昕烨
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Aeronautics and Astronautics
Original Assignee
Nanjing University of Aeronautics and Astronautics
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 Nanjing University of Aeronautics and Astronautics filed Critical Nanjing University of Aeronautics and Astronautics
Priority to CN201910459783.2A priority Critical patent/CN110211379A/en
Publication of CN110211379A publication Critical patent/CN110211379A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/20—Monitoring the location of vehicles belonging to a group, e.g. fleet of vehicles, countable or determined number of vehicles
    • G08G1/202—Dispatching vehicles on the basis of a location, e.g. taxi dispatching

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Development Economics (AREA)
  • Analytical Chemistry (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a kind of public transport method for optimizing scheduling based on machine learning, first according to historical data, agent model is established using the method for machine learning, then initial solution is generated, and it is predicted with agent model, scheme is preferably retained according to optimization aim, then the excellent solution of reservation is sampled, and it is predicted with agent model, the optimization of scheduling scheme is realized according to optimization aim and assessment result.This method, which can be realized to dispatch public transport, carries out Automatic Optimal, so that the scheduling for public transport provides certain auxiliary, provides more convenience for the trip of citizen.

Description

A kind of public transport method for optimizing scheduling based on machine learning
Technical field
The present invention relates to a kind of public transport method for optimizing scheduling based on machine learning, belongs to operational research field.
Background technique
With the promotion at full speed of economic level, the quantity rapid increase of private car, self-driving becomes the main side of trip Formula, still, consequent are congestion and the wretched insufficiency on parking stall of road, cause to occur in recent years many associated Dispute, for example road anger, fight for parking stall etc..It under such a background, while being also to answer environmental requirement, energy-saving and emission-reduction, drop The discharge of low temperature chamber gas, widelys popularize transit trip, including public transport, subway etc., just there is critically important meaning.
Existing public transport scheduling is manually to arrange an order according to class and grade in advance, is dispatched a car according to timetable, by bus rush hour section and non-height Peak time section is the rule of thumb selected period, and in rush hour section and off-peak hours section, dispatching a car is according to same Interval time.Since number of passengers is dynamic change, for example in the technical dates as festivals or holidays, number of passengers is even It is also especially more in off-peak hours section, but present artificial vehicle scheduling often considers this point without surplus energy. Many times, manned number can be excessive or very few in fact for vehicle, and due to not having safety belt, and most of passengers stand, such as Manned number is excessive in fact for fruit vehicle, and there will be biggish security risks, and if vehicle manned number is very few in fact, be to public resource A kind of waste.
Summary of the invention
In view of problem above, the present invention provides a kind of public transport method for optimizing scheduling of base machine learning, this method energy Enough functions of automatically generating and optimize of realizing to public transport scheduling scheme, so that the scheduling for public transport provides centainly Auxiliary, provide more convenience for the trip of citizen.
To achieve the above object, the technical solution adopted by the present invention are as follows:
A kind of public transport method for optimizing scheduling based on machine learning, comprising the following steps:
Step 1, using machine learning method, training prediction agent model；
Step 2, disaggregation scale is set, initial disaggregation is generated；
Step 3, using the agent model of step 1 training, disaggregation is predicted, obtains the prediction of each website manned number in fact Value collection；
Step 4, scheme is preferably retained according to optimization aim；
Step 5, judge whether to meet the number of iterations of setting or meet dispatching requirement, go to step 7 if meeting, otherwise Continue；
Step 6, the excellent solution retained step 5 samples, and generates new disaggregation, and go to step 3；
Step 7, implement optimal case, and utilize the real data renewal agency model for implementing to obtain.
Preferably,
Wherein disaggregation, including initial disaggregation, coding form are as follows:
X={ x1, x2..., xn}
Wherein, X represents disaggregation, xi(i=1,2 ..., n) represents the single solution of Xie Jizhong, and n represents disaggregation scale, that is, solves The number of the solution of concentration.
Preferably,
Wherein solve the single solution of concentration, xi(i=1,2 ..., n), coding form it is as follows:
xi={ t0, t1, t2..., tm}
Wherein, t0Represent the time of departure of the first regular bus in one day, ti(i=1 2 ..., m) represents beginning of i+1 regular bus The hair station time of departure subtracts starting station time of departure of the i-th regular bus, i.e. their interval times for dispatching a car, m+1 is total on the route The class's of dispatching a car number, m represent the departure interval number at the starting station in one day.
Preferably,
Wherein training agent model in step 1, comprising the following steps:
Step 11, according to the difference of agent model function, feature is screened, historical data is divided into according to feature by 2 training Sample；
Step 12, using machine learning method, the different corresponding agent models of training sample training.
Preferably,
Wherein training reaches the feature that each website time prediction agent model uses, comprising: the date, what day, when set out in advance to make arrangements, Uplink/downlink, the starting station time of departure, whether technical dates (holiday), arrival time；
Each website of the training feature that manned number prediction agent model uses in fact, comprising: the date, what day, when set out in advance to make arrangements, on Row/downlink, whether technical dates (holiday), arrival time, real manned number.
Preferably,
The wherein agent model in step 1, comprising: reach each website time prediction agent model, each website manned number in fact Predict agent model.
Preferably,
The method of initial disaggregation is wherein generated in step 2, comprising: random to generate, issue according to the existing vehicle starting station Timetable generate, empirically generate.
Preferably,
Wherein prediction result is obtained in step 3, comprising the following steps:
Step 31, using each website time prediction agent model is reached to disaggregation progress regression forecasting, each coastiong is obtained Arrive different websites time；
Step 32, using each website, manned number predicts agent model in fact, according to the prediction result that step 31 obtains, obtains every Reality manned number of one coastiong in different websites.
Preferably,
Wherein in step 4 optimization aim include: maximize (the manned numbers of total reality of all websites of all vehicles in one day), Minimize (manned number subtracts the absolute value of ideal value to every each website of coastiong in fact).
Preferably,
The wherein method of sampling in step 6, comprising: neighbor seaching, cross and variation, stochastical sampling.
Preferably,
Wherein machine learning algorithm, comprising: linear regression, weighted linear regression, neural network, support vector machine.
Preferably,
For the training sample being newly added, Incremental Learning Algorithm more new model is utilized.
The present invention compared with prior art, has the advantages that
This method optimizes public transport scheduling scheme by using truthful data, to meet the need of more aspect It asks.First according to historical data, agent model is established using the method for machine learning, including reaches each website time prediction agency Manned number predicts agent model in fact for model, each website.Then initial solution is generated, and it is predicted with agent model, according to optimization Target preferably retains scheme.Then the excellent solution of reservation is sampled, and it is predicted with agent model, according to optimization aim The optimization of scheduling scheme is realized with assessment result.Above step is repeated, until meeting preset the number of iterations or meeting scheduling Demand.This method can be realized to public transport dispatch carry out Automatic Optimal, thus for public transport scheduling provide it is certain Auxiliary provides more convenience for the trip of citizen.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention.
Specific embodiment
In the following with reference to the drawings and specific embodiments, the present invention is furture elucidated, it should be understood that these examples are merely to illustrate this It invents rather than limits the scope of the invention, after the present invention has been read, those skilled in the art are to of the invention various The modification of equivalent form falls within the application range as defined in the appended claims.
A kind of public transport method for optimizing scheduling of base machine learning, as shown in Figure 1, comprising the following steps:
Step 1, using machine learning method, training prediction agent model；
Training agent model process is as follows:
Step 11, according to the difference of agent model function, feature is screened, historical data is divided into according to feature by 2 training Sample；
Step 12, using machine learning method, the different corresponding agent models of training sample training.
Machine learning algorithm, including but not limited to: linear regression, weighted linear regression, neural network, support vector machine.
Trained agent model includes but is not limited to: reaching each website time prediction agent model, each website manned number in fact Predict agent model.
Training reaches the feature that each website time prediction agent model uses, comprising: the date, what day, when set out in advance to make arrangements, on Row/downlink, the starting station time of departure, whether technical dates (holiday), arrival time.Each website of training in fact act on behalf of by manned number prediction The feature that model uses, comprising: the date, what day, when next stop, uplink/downlink, whether technical dates (holiday), arrival time, Real manned number.
It is worth noting that in addition to feature listed above, can also increase vehicle size, when next stop state of trade, when Next stop residential building scale etc., or Partial Feature is suitably deleted, feature increases or decreases the accuracy for influencing whether model.
Reach each website time prediction agent model be for according to the date, what day, when next stop, uplink/downlink, originate Stand the time of departure, whether technical dates etc. carrys out the arrival time at each station in line of prediction road.And manned number predicts generation to each website in fact Reason model be for according to the date, what day, when next stop, uplink/downlink, whether technical dates (holiday), arrival time etc. come The manned number of reality in each CFS to CFS rear vehicle in line of prediction road.
Step 2, disaggregation scale is set, initial disaggregation is generated；
Disaggregation includes initial disaggregation, and coding form is as follows:
X={ x1, x2..., xn}
Wherein, X represents disaggregation, xi(i=1,2 ..., n) represents the single solution of Xie Jizhong, and n represents disaggregation scale, that is, solves The number of the solution of concentration.
And xi(i=1,2 ..., n), coding form it is as follows:
xi={ t0, t1, t2..., tm}
Wherein, t0Represent the time of departure of the first regular bus in one day, ti(i=1 2 ..., m) represents beginning of i+1 regular bus The hair station time of departure subtracts starting station time of departure of the i-th regular bus, i.e. their interval times for dispatching a car, m+1 is total on the route The class's of dispatching a car number, m represent the departure interval number at the starting station in one day.
For example, route A shares 5 regular buses in the original plan in one day, the first 8 points of regular bus is issued, and the second 9 points of regular bus issues, third 10 points of regular bus sendings, the 4th regular bus 10: 30 issue, and the 5th 11 points of regular bus issues, then corresponding x1=480,60,60,30, 30 }, unit is minute.Changing the time of departure, the first 10 points of regular bus issues, and the second regular bus 10: 30 issues, and 11 points of third regular bus 30 issue, and the 4th regular bus 12: 30 issues, and the 5th 13 points of regular bus issues, then corresponding x2={ 600,30,60,60,30 }, unit For minute.To present disaggregation X={ x1, x2}。
The method for generating initial disaggregation, including but not limited to: it is random generate, according to the existing vehicle starting station issue when Between table generate, empirically generate.
It can generate at random, it is as follows: the first car starting station time of departure [time in one day is set first0, time1], so Interval time [lower, upper] is set afterwards, then i=1,2 ..., n generate initial solution in turn.
For example, assuming that disaggregation scale n=3, always dispatch a car several m+1=5, then m=4, [time0, time1]=[480,540], [lower, upper]=[15,30].Generate x1Process are as follows: t0Take random number [480,540], it is assumed that 480, t1, t2, t3, t4Respectively Take random number [15,30], it is assumed that 15,20,30,25, x1={ 480,15,20,30,25 }.x2, x3So generate.
The timetable that can also be issued according to the existing vehicle starting station generates, i.e., is generated just according to the existing plan of arranging an order according to class and grade Begin solution.
It can also empirically generate, i.e., initial solution be generated according to the scheduling experience of related dispatcher.
Generation initial solution can also be used in mixed way with several method.
Step 3, using the agent model of step 1 training, disaggregation is predicted, obtains the prediction of each website manned number in fact Value collection；
The process predicted disaggregation is as follows:
Step 31, using each website time prediction agent model is reached to disaggregation progress regression forecasting, each coastiong is obtained Arrive different websites time；
Step 32, using each website, manned number predicts agent model in fact, according to the prediction result that step 31 obtains, obtains every Reality manned number of one coastiong in different websites.
It is worth noting that can also be established single or multiple in addition to above establishes two agent model prediction processes Agent model is predicted according to the starting station time of departure.
Step 4, scheme is preferably retained according to optimization aim；
Optimization aim can be maximization (the manned numbers of total reality of all websites of all vehicles in one day), be also possible to most Smallization (manned number subtracts the absolute value of ideal value to every each website of coastiong in fact) establishes optimization aim according to the actual situation.
Such as optimization aim is set as maximizing (the manned numbers of total reality of all websites of all vehicles in one day), for disaggregation In each solution, the manned number of reality of each website can be obtained in step 3, it is easy to ground, can calculate it is each solution it is corresponding The manned number of total reality of all websites of all vehicles in one day, according to the manned number of total reality of all websites of all vehicles in one day Descending sort is carried out, retains disaggregation scale n that the manned number of total reality of all websites of all vehicles in one day sorts forward Solution.
Step 5, judge whether to meet the number of iterations of setting or meet dispatching requirement, go to step 7 if meeting, otherwise Continue；
Stop after can satisfy the number of iterations of setting, general the number of iterations is set as 50,100,500,1000 etc., iteration The more big then effect of number is better, but has a upper limit, is maintained for being basically unchanged more than the upper limit, generally can be set to 500.
Also stop after can satisfy dispatching requirement, dispatching requirement in general be based on optimization aim propose, such as according to (the manned numbers of total reality of all websites of all vehicles in one day) this optimization aim is maximized, dispatching requirement can be in one day The manned number of total reality of all websites of all vehicles reaches a fixed value.
Step 6, the excellent solution retained step 5 samples, and generates new disaggregation, and go to step 3；
The method of sampling includes but is not limited to: neighbor seaching, cross and variation, stochastical sampling.
Neighbor seaching can be used, as follows: disaggregation scale n solution is selected from the excellent solution of reservation at random and repeatably, Then for each solution selected, each has the probability of p to be generated at random, i.e., for each solved for one, A random number is generated, generates this value at random if being less than p, it is otherwise constant.
For example, one of solution xj={ 480,15,20,30,25 }, it is assumed that p=0.6, for xj, 5 random numbers are generated, It is assumed that respectively 0.8,0.4,0.1,0.9,0.2, the first, fourth invariant position, remaining position regenerate, it is assumed that be generated as 23 respectively, 18,28, then new xj={ 480,23,18,30,28 }, xjFor the neighbours solution of original solution.
Cross and variation can also be used, it is as follows: to select two solution x from the excellent solution of reservation at random and repeatablyj1, xj2, For xj1, each has pcProbability intersected, i.e. xj1Each for, generate a random number, if be less than pcThen With xj2This value is exchanged, it is otherwise constant；For xj1, each has pmProbability make a variation, i.e., for xj1Each For, a random number is generated, if being less than pmThis value is then generated at random, it is otherwise constant.
For example, two of them solution xj1={ 480,15,20,30,25 }, xj2={ 500,17,22,20,27 }, it is assumed that pc= 0.5, for xj1, generate 5 random numbers, it is assumed that and it is respectively 0.6,0.3,0.1,0.9,0.8, first, fourth, five invariant positions, remaining Position and xj2Cross-over value, then xj1={ 480,17,22,30,25 }, it is assumed that pm=0.5, for xj1, then generate 5 random numbers, it is assumed that Respectively 0.2,0.6,0.7,0.3,0.8, second and third, five invariant positions, remaining position regenerates, it is assumed that is respectively 490,21, then New xj1={ 490,17,22,21,25 }.
Stochastical sampling can also be used, i.e., it is random again to generate disaggregation scale n solution.
Step 7, implement optimal case, and utilize the real data renewal agency model for implementing to obtain.
Implement optimal case, obtains model using Incremental Learning Algorithm more new model according to the truthful data of optimal case Dynamically to update.
Then historical data of the present invention by analysis public transport, training agent model are formulated initial scheme and are used Agent model is predicted, sampling operation is carried out on the basis of preferred embodiment, and mould is acted on behalf of in new departure use that sampling generates Type is predicted, is preferably retained new and old scheme, based on optimization aim to public transport scheduling scheme iteration optimization.We Method, which can be realized to dispatch public transport, carries out Automatic Optimal, so that the scheduling for public transport provides certain auxiliary, it is city The trip of the people provides more convenience.
The above is only a preferred embodiment of the present invention, it should be pointed out that: for the ordinary skill people of the art For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered It is considered as protection scope of the present invention.

Claims (10)

1. a kind of public transport method for optimizing scheduling based on machine learning, which comprises the following steps:
Step 1, using machine learning method, training prediction agent model；
Step 2, disaggregation scale is set, initial disaggregation is generated；
Step 3, using the agent model of step 1 training, disaggregation is predicted, obtains the predicted value of each website manned number in fact Collection；
Step 4, scheme is preferably retained according to optimization aim；
Step 5, judge whether meet setting the number of iterations or meet dispatching requirement, if meet if go to step 7, otherwise after It is continuous；
Step 6, the excellent solution retained step 5 samples, and generates new disaggregation, and go to step 3；
Step 7, implement optimal case, and utilize the real data renewal agency model for implementing to obtain.
2. according to claim 1, it is characterised in that: the disaggregation, including initial disaggregation, coding form are as follows:
X={ x1, x2..., xn}
Wherein, X represents disaggregation, xi(i=1,2 ..., n) represents the single solution of Xie Jizhong, and n represents disaggregation scale, i.e. solution is concentrated The number of solution.
3. according to claim 2, it is characterised in that: the single solution that the solution is concentrated, xi(i=1,2 ..., n), coding Form is as follows:
xi={ t0, t1, t2..., tm}
Wherein, t0Represent the time of departure of the first regular bus in one day, ti(i=1 2 ..., m) represents the starting station of i+1 regular bus The time of departure subtracts the starting station time of departure of the i-th regular bus, i.e. their interval times for dispatching a car, m+1 is always to dispatch a car on the route Class's number, m represent the departure interval number at the starting station in one day.
4. according to claim 1, it is characterised in that: training agent model in the step 1, comprising the following steps:
Step 11, according to the difference of agent model function, feature is screened, historical data is divided into according to feature by 2 training samples；
Step 12, using machine learning method, the different corresponding agent models of training sample training.
5. according to claim 4, it is characterised in that:
The training reaches the feature that each website time prediction agent model uses, comprising: the date, what day, when set out in advance to make arrangements, on Row/downlink, the starting station time of departure, whether technical dates (holiday), arrival time；
Each website of the training feature that manned number prediction agent model uses in fact, comprising: the date, what day, when set out in advance to make arrangements, uplink/under Row, whether technical dates (holiday), arrival time, real manned number.
6. according to claim 1, it is characterised in that: the agent model in the step 1, comprising: reach each website time Predict agent model, each website manned number prediction agent model in fact.
7. according to claim 1, it is characterised in that: generate the method for initial disaggregation in the step 2, comprising: random raw It generates at, the timetable that is issued according to the existing vehicle starting station, empirically generate.
8. according to claim 1, it is characterised in that: obtain prediction result in the step 3, comprising the following steps:
Step 31, it using each website time prediction agent model is reached to disaggregation progress regression forecasting, obtains each coastiong and arrives The time of different websites；
Step 32, using each website, manned number predicts agent model in fact, according to the prediction result that step 31 obtains, obtains each time Reality manned number of the vehicle in different websites.
9. according to claim 1, it is characterised in that: optimization aim includes: and maximizes (to own in one day in the step 4 The manned number of total reality of all websites of vehicle), minimize (manned number subtracts the absolute of ideal value to every each website of coastiong in fact Value).
10. according to claim 1, it is characterised in that: the method for sampling in the step 6, comprising: neighbor seaching intersects change Different, stochastical sampling.
CN201910459783.2A 2019-05-27 2019-05-27 A kind of public transport method for optimizing scheduling based on machine learning Pending CN110211379A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910459783.2A CN110211379A (en) 2019-05-27 2019-05-27 A kind of public transport method for optimizing scheduling based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910459783.2A CN110211379A (en) 2019-05-27 2019-05-27 A kind of public transport method for optimizing scheduling based on machine learning

Publications (1)

Publication Number Publication Date
CN110211379A true CN110211379A (en) 2019-09-06

Family

ID=67789440

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910459783.2A Pending CN110211379A (en) 2019-05-27 2019-05-27 A kind of public transport method for optimizing scheduling based on machine learning

Country Status (1)

Country Link
CN (1) CN110211379A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111539133A (en) * 2020-07-09 2020-08-14 深圳市都市交通规划设计研究院有限公司 Single-line public vehicle scheduling method combining expert experience and operational research optimization
CN112580909A (en) * 2019-09-29 2021-03-30 珠海格力电器股份有限公司 Dynamic regular bus arrangement method, computer readable storage medium and terminal
CN113450557A (en) * 2020-03-24 2021-09-28 支付宝(杭州)信息技术有限公司 Method and device for updating prediction model for vehicle passenger flow
CN114066032A (en) * 2021-11-08 2022-02-18 郑州天迈科技股份有限公司 Automatic compiling method for bus timetable
US11593729B2 (en) 2020-03-13 2023-02-28 International Business Machines Corporation Cognitive tuning of scheduling constraints

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101354763A (en) * 2008-09-09 2009-01-28 中兴智能交通***（北京）有限公司 Method and system for scheduling resource
CN102074124A (en) * 2011-01-27 2011-05-25 山东大学 Dynamic bus arrival time prediction method based on support vector machine (SVM) and H-infinity filtering
CN105224992A (en) * 2014-05-28 2016-01-06 国际商业机器公司 To waiting for the method and system predicted of ridership and evaluation method and system
CN106530689A (en) * 2016-10-19 2017-03-22 南昌大学 Bus arrival time real-time predication method based on genetic algorithm and driving data
CN107092976A (en) * 2017-03-28 2017-08-25 东南大学 A kind of method of multi-objective Model collaboration optimization a plurality of public bus network departure interval
CN107845259A (en) * 2017-10-24 2018-03-27 东南大学 Public transport operation situation real-time feedback system and public transport real-time running data processing method
CN108491958A (en) * 2018-02-11 2018-09-04 浙江工业大学 A kind of bus passenger flow string invariant prediction technique in short-term
CN108629469A (en) * 2017-03-17 2018-10-09 上海苍烨智能科技有限公司 A kind of public transport operation management dispatching method and system
CN109035770A (en) * 2018-07-31 2018-12-18 上海世脉信息科技有限公司 The real-time analyzing and predicting method of public transport passenger capacity under a kind of big data environment

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101354763A (en) * 2008-09-09 2009-01-28 中兴智能交通***（北京）有限公司 Method and system for scheduling resource
CN102074124A (en) * 2011-01-27 2011-05-25 山东大学 Dynamic bus arrival time prediction method based on support vector machine (SVM) and H-infinity filtering
CN105224992A (en) * 2014-05-28 2016-01-06 国际商业机器公司 To waiting for the method and system predicted of ridership and evaluation method and system
CN106530689A (en) * 2016-10-19 2017-03-22 南昌大学 Bus arrival time real-time predication method based on genetic algorithm and driving data
CN108629469A (en) * 2017-03-17 2018-10-09 上海苍烨智能科技有限公司 A kind of public transport operation management dispatching method and system
CN107092976A (en) * 2017-03-28 2017-08-25 东南大学 A kind of method of multi-objective Model collaboration optimization a plurality of public bus network departure interval
CN107845259A (en) * 2017-10-24 2018-03-27 东南大学 Public transport operation situation real-time feedback system and public transport real-time running data processing method
CN108491958A (en) * 2018-02-11 2018-09-04 浙江工业大学 A kind of bus passenger flow string invariant prediction technique in short-term
CN109035770A (en) * 2018-07-31 2018-12-18 上海世脉信息科技有限公司 The real-time analyzing and predicting method of public transport passenger capacity under a kind of big data environment

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
黄婧婧: "中小城市公交需求特性分析与线网设计方法研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112580909A (en) * 2019-09-29 2021-03-30 珠海格力电器股份有限公司 Dynamic regular bus arrangement method, computer readable storage medium and terminal
US11593729B2 (en) 2020-03-13 2023-02-28 International Business Machines Corporation Cognitive tuning of scheduling constraints
CN113450557A (en) * 2020-03-24 2021-09-28 支付宝(杭州)信息技术有限公司 Method and device for updating prediction model for vehicle passenger flow
CN111539133A (en) * 2020-07-09 2020-08-14 深圳市都市交通规划设计研究院有限公司 Single-line public vehicle scheduling method combining expert experience and operational research optimization
CN114066032A (en) * 2021-11-08 2022-02-18 郑州天迈科技股份有限公司 Automatic compiling method for bus timetable

Similar Documents

Publication Publication Date Title
CN110211379A (en) A kind of public transport method for optimizing scheduling based on machine learning
Kuppusamy et al. Deep learning based energy efficient optimal timetable rescheduling model for intelligent metro transportation systems
CN104239484B (en) One kind is based on table preparation method at the time of multi-mode public transport combined schedule
CN105740556B (en) The automatic preparation method of route map of train based on passenger flow demand
CN105575108B (en) A kind of intelligent public transportation dispatching method for running
Chen et al. Uncertainty in urban mobility: Predicting waiting times for shared bicycles and parking lots
CN107358319A (en) Flow Prediction in Urban Mass Transit method, apparatus, storage medium and computer equipment
CN111452669B (en) Intelligent bus charging system and method and medium
Zhao et al. An integrated approach of train scheduling and rolling stock circulation with skip-stopping pattern for urban rail transit lines
CN110458456B (en) Demand response type public transportation system scheduling method and system based on artificial intelligence
CN109558985A (en) A kind of bus passenger flow amount prediction technique based on BP neural network
CN107274665A (en) Bus transport capacity resource method and system for planning
CN111340289A (en) Genetic algorithm-based bus departure and speed adjustment optimization method and system
CN106652434A (en) Bus dispatching method based on rail transit coordination
Li et al. Deep learning based parking prediction on cloud platform
CN102737356A (en) Intelligent bus scheduling calculation method
CN102324122A (en) Can generate the intelligent riding system of the scheme of riding and ability passenger flow forecast amount
CN104477217A (en) Operation and circulation of trains on city rail with single end station
CN115456257A (en) Airport passenger distributing and transporting system based on modular automatic driving bus connection
CN112990536A (en) Bus arrival time prediction method based on quantile convolution network
Zhao et al. Robust Optimization of Mixed-Load School Bus Route Based on Multi-Objective Genetic Algorithm.
CN113033928A (en) Design method, device and system of bus shift scheduling model based on deep reinforcement learning
CN111785015A (en) Public transport real-time regulation and control information system and scheduling method based on genetic algorithm
CN113393029B (en) Method and equipment for predicting rail transit passenger flow
CN106845721A (en) The optimization method of many beat synthetic operations of train based on Cross-Entropy Algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20190906

WD01 Invention patent application deemed withdrawn after publication