CN110211422A - Car owner based on markov decision process seeks parking stall policy recommendation method - Google Patents
Car owner based on markov decision process seeks parking stall policy recommendation method Download PDFInfo
- Publication number
- CN110211422A CN110211422A CN201910574860.9A CN201910574860A CN110211422A CN 110211422 A CN110211422 A CN 110211422A CN 201910574860 A CN201910574860 A CN 201910574860A CN 110211422 A CN110211422 A CN 110211422A
- Authority
- CN
- China
- Prior art keywords
- berth
- decision process
- markov decision
- parking stall
- grid
- 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
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
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
Abstract
The present invention proposes that a kind of car owner based on markov decision process seeks parking stall policy recommendation method, belongs to wisdom parking lot.Technical solution of the present invention main points are as follows: data are pre-processed, the pretreated content includes: suppressing exception record, divides track by block and calculate the berth probability of success and the expected time of each block；It is modeled using markov decision process to berth optimal policy problem is looked for；Calculate the optimal solution for the model established；Consider load balancing and real-time road, remove the route of heavy congestion, and provides and optimal seek berth strategy.The present invention utilizes the model of markov decision process, to accomplish in Generalization bounds not only comprising seeking berth route, also there is equal berths strategy, it can use load-balancing algorithm simultaneously and solve multiple vehicles in same place while race problem caused by filing a request, and avoid that vehicle is allowed to drive into congested link by real-time road condition information.
Description
Technical field
The present invention relates to wisdom stopping technicals, in particular to the car owner based on markov decision process seeks parking stall strategy and pushes away
Recommend the technology of method.
Background technique
As motor vehicle and driver's quantity increase rapidly, while bringing convenient to people's production and living, while also existing
A series of problems is brought in varying degrees, the Urban Traffic Planning in China only payes attention to the construction of urban road facility mostly at present,
The reserved and construction for having ignored parking stall, so that city parking position is more and more nervous, with the hair at full speed of wireless sensor technology
Exhibition, GPS device is widely used by vehicle, while producing many wheelpath information.These information have existed at present
There is application in many fields, such as city calculates and route planning.Different from travelling bus and ground according to route daily
Iron, each car owner, which has, oneself to be looked for parking stall strategy to carry out expectation maximization saving time and oily expense.Existing recommended method can divide
For following two categories: (1) recommending next stop.These methods are mostly based on probabilistic type, can recommend the vehicle of most possible free time
Position, but the specific route where going to next parking stall cannot be provided.(2) recommend specifically to seek parking stall route.These recommend system
System can recommend one to seek parking stall route to car owner, to help them to reach parking stall as early as possible.On the whole, these systems all
It finds the best problem for seeking parking stall strategy and has been defined as an optimization problem.Although existing method realizes this target,
But still have some limitations.
Summary of the invention
The object of the present invention is to provide a kind of car owners based on markov decision process to seek parking stall policy recommendation method, energy
It is enough to realize quickly parking.
The present invention solves its technical problem, the technical solution adopted is that: the car owner based on markov decision process seeks vehicle
Position policy recommendation method, includes the following steps:
Step 1 pre-processes data, and the pretreated content includes: suppressing exception record, by block division rail
Mark and the berth probability of success and the expected time for calculating each block；
Step 2 is modeled using markov decision process to berth optimal policy problem is looked for；
Step 3, the optimal solution for calculating the model established；
Step 4 considers load balancing and real-time road, removes the route of heavy congestion, and provides and optimal seek berth plan
Slightly.
Particularly, in step 1, the data prediction includes the following steps:
Step 101, in order to remove GPS offset and equipment fault caused by error, filter out adjacent point-to-point transmission every more than 3 points
Clock or spacing are more than the sampled point of 3km；
City is divided into grid of corresponding size by step 102；
The place that step 103, detection can park；
The step 103 specifically comprises the following steps:
Step 1031 extracts that all there are the tracks of parking behavior from GPS track；
If step 1032, parking trajectory nearby do not have in information point or track there is no from free berth to berth hurry
State conversion, then it is assumed that the parking trajectory is ignored near nearest berth place；
Step 1033 finds all parking trajectories near the place of berth, calculates the geometric center conduct of every track
The point of parking stall is waited, and these points are clustered, obtains all places that can wait idle berth.
Further, in step 2, the markov decision process is come using four-tuple (S, A, P (s ' | s, a), R)
Indicate, wherein S is state set, and A is behavior aggregate, for s ∈ S, s' ∈ S, P (s' │ s, a) be when movement a ∈ A be performed after,
State is transferred to the transition probability of s' from s, and (s a) is to act the synthesis of the time and mileage that obtain after a in state s execution to R.
Particularly, when city being divided into grid of corresponding size, if S is corresponding grid set, A (s) is corresponding in s
The fair play set that grid can be taken, and P (s ' | s is a) when execution acts a, and the transfer from grid s to grid s ' is general
Rate, optimal policy is related with the time, and by R, (s a) is defined as R (s, a, t), i.e. time t is obtained after grid s performs movement a
Expected time.
Further, the movement includes that upper and lower, left and right, upper left, lower-left, upper right, bottom right and original place wait berth.
The invention has the advantages that seeking parking stall policy recommendation side by the above-mentioned car owner based on markov decision process
Method will wait berths to be considered as one kind near parking lot and effectively seek berth strategy, and be added into markov decision process
Model to accomplish in Generalization bounds not only also to have comprising seeking berth route equal berths strategy, while can use load
It accounts method and solves multiple vehicles in same place while race problem caused by filing a request, and avoided by real-time road condition information
Vehicle is allowed to drive into congested link.
Detailed description of the invention
Fig. 1 is that the present invention is based on the flow charts that the car owner of markov decision process seeks parking stall policy recommendation method.
Specific embodiment
With reference to the accompanying drawing, the technical schemes of the invention are described in detail.
Car owner of the present invention based on markov decision process seeks parking stall policy recommendation method, and flow chart is referring to figure
1, wherein this method comprises the following steps:
Step 1 pre-processes data, and the pretreated content includes: suppressing exception record, by block division rail
Mark and the berth probability of success and the expected time for calculating each block.
Wherein, data prediction includes the following steps:
Step 101, in order to remove GPS offset and equipment fault caused by error, filter out adjacent point-to-point transmission every more than 3 points
Clock or spacing are more than the sampled point of 3km；
City is divided into grid of corresponding size by step 102；
The place that step 103, detection can park；
The step 103 specifically comprises the following steps:
Step 1031 extracts that all there are the tracks of parking behavior from GPS track；
If step 1032, parking trajectory nearby do not have in information point or track there is no from free berth to berth hurry
State conversion, then it is assumed that the parking trajectory is ignored near nearest berth place；
Step 1033 finds all parking trajectories near the place of berth, calculates the geometric center conduct of every track
The point of parking stall is waited, and these points are clustered, obtains all places that can wait idle berth.
Step 2 is modeled using markov decision process to berth optimal policy problem is looked for.
Wherein, markov decision process can be indicated using four-tuple (S, A, P (s ' | s, a), R), wherein S is shape
State collection, A are behavior aggregate, and for s ∈ S, s' ∈ S, P, (s' │ s is a) after movement a ∈ A is performed, and state is transferred to s' from s
Transition probability, (s a) is to act the synthesis of time and mileage obtained after a in state s execution to R.
When city is divided into grid of corresponding size, if S is corresponding grid set, A (s) is to be corresponding in s grid institute energy
The fair play set taken, and P (s ' | s is a) when execution acts a, from grid s to the transition probability of grid s ', optimal plan
Slightly related with the time, by R, (s a) is defined as R (s, a, t), i.e. time t is when grid s performs the expectation obtained after movement a
Between.
Movement can preferably include upper and lower, left and right, upper left, lower-left, upper right, bottom right and original place and wait berth.
Step 3, the optimal solution for calculating the model established.
Wherein it is possible to learn to solve optimal solution using Q-.
Step 4 considers load balancing and real-time road, removes the route of heavy congestion, and provides and optimal seek berth plan
Slightly.
Wherein, consider load balancing and real-time road, provide and optimal seek parking stall strategy.After receiving recommendation request, benefit
K optimal policy before being calculated with markov decision process model；Scheduling system according to same grid in a time window its
The request situation of his vehicle carries out load balancing to every route, and removes the route of heavy congestion in conjunction with real-time road, to give
It is most suitable out to seek berth strategy.
Claims (5)
1. the car owner based on markov decision process seeks parking stall policy recommendation method, which comprises the steps of:
Step 1 pre-processes data, the pretreated content include: suppressing exception record, by block divide track with
And calculate the berth probability of success and the expected time of each block；
Step 2 is modeled using markov decision process to berth optimal policy problem is looked for；
Step 3, the optimal solution for calculating the model established；
Step 4 considers load balancing and real-time road, removes the route of heavy congestion, and provides and optimal seek berth strategy.
2. the car owner according to claim 1 based on markov decision process seeks parking stall policy recommendation method, feature
It is, in step 1, the data prediction includes the following steps:
Step 101, in order to remove error caused by GPS offset and equipment fault, filter out adjacent point-to-point transmission every more than 3 minutes or
Person's spacing is more than the sampled point of 3km；
City is divided into grid of corresponding size by step 102；
The place that step 103, detection can park；
The step 203 specifically comprises the following steps:
Step 1031 extracts that all there are the tracks of parking behavior from GPS track；
If step 1032, parking trajectory do not have nearby in information point or track, there is no the states hurried from free berth to berth
Conversion, then it is assumed that the parking trajectory is ignored near nearest berth place；
Step 1033 finds all parking trajectories near the place of berth, calculates the geometric center of every track as waiting
The point of parking stall, and these points are clustered, obtain all places that can wait idle berth.
3. the car owner according to claim 1 based on markov decision process seeks parking stall policy recommendation method, feature
It is, in step 2, the markov decision process is indicated using four-tuple (S, A, P (s ' | s, a), R), wherein S is shape
State collection, A are behavior aggregate, and for s ∈ S, s' ∈ S, P, (s' │ s is a) after movement a ∈ A is performed, and state is transferred to s' from s
Transition probability, (s a) is to act the synthesis of time and mileage obtained after a in state s execution to R.
4. the car owner according to claim 3 based on markov decision process seeks parking stall policy recommendation method, feature
It is, when city is divided into grid of corresponding size, if S is corresponding grid set, A (s) can be taken for correspondence in s grid
Fair play set, P (s ' | s, a) be when execution acts a, from grid s to the transition probability of grid s ', optimal policy and
Time is related, and by R, (s a) is defined as R (s, a, t), i.e. time t performs the expected time obtained after movement a in grid s.
5. the car owner according to claim 4 based on markov decision process seeks parking stall policy recommendation method, feature
It is, the movement includes that upper and lower, left and right, upper left, lower-left, upper right, bottom right and original place wait berth.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910574860.9A CN110211422A (en) | 2019-06-28 | 2019-06-28 | Car owner based on markov decision process seeks parking stall policy recommendation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910574860.9A CN110211422A (en) | 2019-06-28 | 2019-06-28 | Car owner based on markov decision process seeks parking stall policy recommendation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110211422A true CN110211422A (en) | 2019-09-06 |
Family
ID=67795360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910574860.9A Pending CN110211422A (en) | 2019-06-28 | 2019-06-28 | Car owner based on markov decision process seeks parking stall policy recommendation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110211422A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103700281A (en) * | 2013-12-20 | 2014-04-02 | 柳州市盛景科技有限公司 | Application method of parking guide system based on dynamic statistics |
| CN103838846A (en) * | 2014-03-06 | 2014-06-04 | 中国科学院软件研究所 | Emergency guiding method and emergency guiding system for individual on basis of big data |
| CN104408967A (en) * | 2014-11-26 | 2015-03-11 | 浙江中南智能科技有限公司 | Cloud-computing-based parking lot management system |
| CN106228848A (en) * | 2016-09-29 | 2016-12-14 | 北京百度网讯科技有限公司 | A kind of parking navigation method and apparatus |
| CN107832882A (en) * | 2017-11-03 | 2018-03-23 | 上海交通大学 | A kind of taxi based on markov decision process seeks objective policy recommendation method |
-
2019
- 2019-06-28 CN CN201910574860.9A patent/CN110211422A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103700281A (en) * | 2013-12-20 | 2014-04-02 | 柳州市盛景科技有限公司 | Application method of parking guide system based on dynamic statistics |
| CN103838846A (en) * | 2014-03-06 | 2014-06-04 | 中国科学院软件研究所 | Emergency guiding method and emergency guiding system for individual on basis of big data |
| CN104408967A (en) * | 2014-11-26 | 2015-03-11 | 浙江中南智能科技有限公司 | Cloud-computing-based parking lot management system |
| CN106228848A (en) * | 2016-09-29 | 2016-12-14 | 北京百度网讯科技有限公司 | A kind of parking navigation method and apparatus |
| CN107832882A (en) * | 2017-11-03 | 2018-03-23 | 上海交通大学 | A kind of taxi based on markov decision process seeks objective policy recommendation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Energy-optimal speed control for electric vehicles on signalized arterials | |
| CN105279955B (en) | A kind of share-car method and apparatus | |
| CN107832882A (en) | A kind of taxi based on markov decision process seeks objective policy recommendation method | |
| CN105225472A (en) | A kind of share-car method and apparatus | |
| CN106816009B (en) | Highway real-time traffic congestion road conditions detection method and its system | |
| CN104750963B (en) | Intersection delay duration method of estimation and device | |
| CN110751366A (en) | Method for calculating running capacity matching degree of main railway and urban rail transit | |
| US20160055743A1 (en) | Congo: System and Method of Transportation Using Carrier Vehicles and Personal Transport Vehicles | |
| CN112036757B (en) | Mobile phone signaling and floating car data-based parking transfer parking lot site selection method | |
| CN105206038B (en) | Bus trip response system and response method based on car networking technology | |
| CN102800212B (en) | Intelligent parking system under a kind of environment of internet of things | |
| CN101833855A (en) | Planar intelligent traffic system of taxi get-on stations in large-scale traffic hub | |
| CN104778832B (en) | A kind of multimode Urban Road Cross-section resource coordination collocation method | |
| CN107844603A (en) | Electric automobile charging pile inquiry guiding system based on cloud computing | |
| CN106530695A (en) | Urban trunk road vehicle travel time real-time prediction method based on Internet of vehicles | |
| CN105046983A (en) | Traffic flow prediction system and method based on vehicle-road cooperation | |
| CN111859718A (en) | Method and system for calculating congestion coefficient of regional multi-standard rail transit station | |
| CN113140108B (en) | Cloud traffic situation prediction method in internet-connected intelligent traffic system | |
| CN114971136A (en) | Bus and tour bus scheduling method | |
| CN110364003A (en) | Intersection double-circuit tramcar signal priority control method based on car networking | |
| CN103226666B (en) | Sophisticated functions intercommunication road crossing system efficacy assessments method | |
| CN110426969B (en) | Intelligent network connection micro bus-based control system, reservation and response control method | |
| CN110211422A (en) | Car owner based on markov decision process seeks parking stall policy recommendation method | |
| Sargious et al. | Forecasting demand for the park-and-ride mode and determining the optimal location of stations | |
| CN111785065B (en) | Intelligent remote autonomous parking method based on user travel cost and travel efficiency |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190906 |