CN107844853A - It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price - Google Patents
It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price Download PDFInfo
- Publication number
- CN107844853A CN107844853A CN201710992208.XA CN201710992208A CN107844853A CN 107844853 A CN107844853 A CN 107844853A CN 201710992208 A CN201710992208 A CN 201710992208A CN 107844853 A CN107844853 A CN 107844853A
- Authority
- CN
- China
- Prior art keywords
- price
- net
- riding
- region
- predictability
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Databases & Information Systems (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention is a kind of commending system by bus predicted using dynamic price and reduce net about fare, and for user, sometime the net in place about car is ridden demand in city, there is provided the recommendation of stroke Cost Forecast and less expense riding scheme.First, city is subdivided into multiple regions by the system, and the dynamic price predictability in region is calculated using entropy and Fa Nuo inequality.Then, system is directed to the region of different predictability sizes, selects Markov Chain Forecast device or neural network prediction device to carry out dynamic price prediction.Finally, system prediction goes out riding fee and recommends the riding scheme of reduction expense, and such as user original place a period of time or a mobile segment distance can obtain lower price by bus.Test result indicates that the prediction result of the system is substantially coincide with survey data, make user reduce because riding fee it is uncertain caused by anxiety, and save trip spending.
Description
Technical field
The present invention relates to trip spatiotemporal data structure, commending system field, particularly net about car order spatiotemporal data structure.
Background technology
The nets such as Uber and Didi about car platform is of increasing concern as emerging learning on demand service (RoD), as biography
To unite a kind of supplements of tax services, the features such as RoD services are to clean, facilitate, flexibility and dynamic price are reasonable, attracts passenger,
On the other hand those have been attracted to want to be not desired to the driver to apply for a license using idle automobile.The core and uniqueness of RoD services are special
Sign is Dynamic Pricing, and it reflects the supply (vehicle fleet size) of the Service controll ad-hoc location and arranging for demand (number of requests)
Apply, supply and demand is reached balanced.In traditional tax services, trip price is as stroke length changes, and passenger can be with
Stroke expense is substantially judged according to the experience of life, but in RoD services, the stroke expense floating of passenger is larger, is sometimes sending out
When raw major issue or bad weather, dynamic price can become very high (up to 5 to 10 times of normal price).This to use
Family can produce burden and anxiety inwardly, if can be with when a forecast price of RoD service platforms return is obtained
Walk a segment distance or wait a period of time to obtain lower price, specifically, where walk how farDeng how longMesh
Pointed out in preceding multiple researchs, in city the movement of people there is stronger regularity, people are always in locality special time
Trip, and dynamic price is associated by force with supply/demand by bus, and this prediction just for regional dynamics price provides possibility.Together
When, reply above mentioned problem requires that the system can be that the about automobile-used family of net provides information, including the certain model in pick-up point periphery in real time
All riding fees of all areas within this time point and following a period of time in enclosing are used, and are finally recommended to user optional
And the riding scheme that expense is lower.By the system, user can reduce the uncertain and caused anxiety to price of riding,
Trip spending is also saved simultaneously.
The content of the invention
The present invention is a kind of commending system by bus predicted using dynamic price and reduce net about fare, for user in city
Sometime the net in place about car is ridden demand in city, there is provided stroke Cost Forecast and is reduced the riding scheme of expense and is recommended.
Specifically, when user sent on RoD service platforms ask by bus when, the system is according to the time place of request, in advance
Measure including a range of all areas in place periphery, within this time point and following a period of time it is all by bus
Expense (riding fee use=original prices * dynamic prices multiplying power), finally the riding scheme of information of forecasting and reduction expense is provided
To user.First, city is subdivided into multiple grid spaces by the system, is being found the stage of region characteristic, is being introduced more generation
The functional areas of table are analyzed, such as Fig. 2.By the analysis to the about car historical data of the net on functional areas, RoD services are drawn
Dynamic price shows different variation characteristics, such as Fig. 3 with the time in different regions, on the strong and weak basis of the periodicity of its change
On, calculate some region of price predictability using true entropy and Fa Nuo inequality.
Then, system is directed to the region of different price predictability sizes, chooses Markov Chain Forecast device and neutral net is pre-
Survey the multiplying power prediction that device carries out regional dynamics price.The net of Markov Chain Forecast device using area about car History Order data are entered
Row training, neural network prediction device are then trained using the weather data of net about car historical data and place of corresponding time.
Finally, system prediction goes out riding fee and recommends the riding scheme of reduction expense, i.e., the user original place etc. a period of time or
Person, which moves a segment distance, can obtain lower price by bus.System details are shown in Fig. 1.
Specifically, it is of the invention, a kind of side that the commending system by bus for reducing net about fare is predicted using dynamic price
Case is:
It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price, it is characterized in that:Urban area valency
Lattice predictability module, fallout predictor training module and riding scheme recommending module;
Described urban area price predictability module, it is to be based on urban geography data and net about car historical data, for difference
City subregion calculate the predictability of its dynamic price, so as to be the suitable fallout predictor of the regional choice;
Described fallout predictor training module, it is the price predictability size according to regional, chooses different training respectively
Device model, and corresponding fallout predictor model is trained according to the about car historical data of the net on the region;
Described riding scheme recommending module, it is to provide time of specific demand by bus in the case of place, by pre-
Survey device prediction include a range of all areas in ride site periphery at the moment and subsequent time period riding fee use, pass through
Integrated comparative, user can select most suitable riding scheme, such as walk a segment distance or a period of time.
A kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1,
Described urban area price predictability module comprises the following steps:
Step 1:Mark representative functional areas in city, while according to urban geography data be subdivided into city multiple
Deng grid spaces;
Step 2:Using the about car historical data of the net on the regional marked in step 1, the true entropy of dynamic price is calculated;
Step 3:The entropy calculated in step 2 according to each region, the dynamic in the region is calculated using method promise inequality
Price predictability.
A kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1,
Fallout predictor can predict 7 multiplying powers of the dynamic price in this section of 1.0-1.6, described fallout predictor training module include with
Lower step:
Step 1:Markov Chain Forecast device is built, is instructed using the net about car historical data in the high region of price predictability
Practice;
Step 2:Collect the weather data that each region corresponds to the time;
Step 3:Neural network prediction device is built, uses the net about car historical data and correspondingly in the relatively low region of price predictability
Weather data be trained.
A kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1,
Described riding scheme recommending module comprises the following steps:
Step 1:The time place and Weather information selection fallout predictor sent according to user's demand of riding, predicts regional dynamics valency
Lattice multiplying power;
Step 2:Calculate a range of all areas in pick-up point periphery at the moment and subsequent time period riding fee
With;
Step 3:More economical riding scheme is sorted out, such as in region a period of time or is moved to another region of riding,
User is allowed voluntarily to select suitable scheme.
Brief description of the drawings
Fig. 1 is the system block diagram of the present invention;
Fig. 2 is that city major function distinguishes Butut;
Fig. 3 is dynamic price variation tendency of three kinds of functional areas in one month;
Fig. 4 is the distribution situation of zone price predictability under different entropy computational methods and different Annual distributions;
Fig. 5 is the curve that the sMAPE error amounts of two kinds of fallout predictors change with zone price predictability.
Embodiment
The present invention is further described below in conjunction with the accompanying drawings.
The system framework of the system mainly includes three modules:Urban area price predictability module, fallout predictor training
Module and riding scheme recommending module, shown in Fig. 1 is the system block diagram of the present invention, wherein:
Urban area price predictability module, the starting point of the module are the dynamic price change week in different regions
Phase property intensity is different, and corresponding prediction difficulty is also just different, such as Fig. 3, so classification of the output as region of this module, with suitable
With different fallout predictors, comprise the following steps:
Step 1:Urban function region, such as Fig. 2 are marked off, while is rectangle grid spaces with mesh generation by city, it is each to make
Enough data can be obtained in grid size, each grid size is to be arranged to 420*300 rice;
Step 2:According to the about car historical data of the net on each region, the true entropy of dynamic price is calculated, true entropy can be anti-
While mirroring the confusion degree of data, it is contemplated that relevance of the data in time series,
Step 3:Using method promise inequality according to the dynamic price entropy in step 2 on regional, corresponding region is calculated
Dynamic price predictability, as Fig. 4 disclose dynamic price under different entropy computational methods and different Annual distributions can
Predictive distribution situation.
Fallout predictor training module, 7 multiplying powers that can predict dynamic price in this section of 1.0-1.6 are trained, wrapped
Include following steps:
Step 1:Use the net about car historical data (every 1 hour) in the higher region of dynamic price predictability, training pace
For 3 Markov Chain Forecast device;
Step 2:Collect the weather data of regional, including daily precipitation and the temperature of each hour;
Step 3:A three-layer neural network fallout predictor is built, including the use of 4 features (which on the same day, which in one week the hour
One day, intra day ward, the mean temperature of current hour) input layer, ReLU excitation layers and Softmax function output layers, use
The weather data (every 1 hour) in the net in the relatively low region of price predictability about car historical data and place of corresponding time is carried out
Training, Fig. 5 are the curve that the sMAPE error amounts of two kinds of fallout predictors change with zone price predictability.
Riding scheme recommending module comprises the following steps:
Step 1:The time place and Weather information selection fallout predictor sent according to user's demand of riding, predicts regional dynamics valency
Lattice multiplying power;
Step 2:Calculate a range of all areas in pick-up point periphery at the moment and subsequent time period riding fee
With;
Step 3:More economical riding scheme is sorted out, such as in region a period of time or is moved to another region of riding,
User is allowed voluntarily to select suitable scheme.
Claims (4)
1. a kind of predict the commending system by bus for reducing net about fare using dynamic price, it is characterized in that:Urban area price
Predictability module, fallout predictor training module and riding scheme recommending module;
Described urban area price predictability module, it is to be based on urban geography data and net about car historical data, for difference
City subregion calculate its dynamic price predictability, so as to being the suitable fallout predictor of the regional choice;
Described fallout predictor training module, it is the price predictability size according to regional, chooses different training respectively
Device model, and corresponding fallout predictor is trained according to the about car historical data of the net on the region;
Described riding scheme recommending module, it is to provide time of specific demand by bus in the case of place, by pre-
Survey device prediction include a range of all areas in ride site periphery at the moment and subsequent time period riding fee use, pass through
Integrated comparative, user can select most suitable riding scheme, such as walk a segment distance or a period of time.
2. a kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1, its
It is characterized in:Described urban area price predictability module comprises the following steps:
Step 1:Mark representative functional areas in city, while according to urban geography data be subdivided into city multiple
Deng grid spaces;
Step 2:Using the about car historical data of the net on the regional marked in step 1, the true entropy of dynamic price is calculated;
Step 3:The entropy calculated in step 2 according to each region, the dynamic in the region is calculated using method promise inequality
Price predictability.
3. a kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1, its
It is characterized in, fallout predictor can predict 7 multiplying powers of the dynamic price in this section of 1.0-1.6, and training module includes following step
Suddenly:
Step 1:Markov Chain Forecast device is built, is instructed using the net about car historical data in the high region of price predictability
Practice;
Step 2:Collect the weather data that each region corresponds to the time;
Step 3:Neural network prediction device is built, uses the net about car historical data and correspondingly in the relatively low region of price predictability
Weather data be trained.
4. a kind of commending system by bus that reduction net about fare is predicted using dynamic price according to claim 1, its
It is characterized in, described riding scheme recommending module comprises the following steps:
Step 1:The time place and Weather information selection fallout predictor sent according to user's demand of riding, predicts regional dynamics valency
Lattice multiplying power;
Step 2:Calculate the pick-up point neighboring area at the moment and subsequent time period riding fee use;
Step 3:More economical riding scheme is sorted out, such as in region a period of time or is moved to another region of riding,
User is allowed voluntarily to select suitable scheme.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710992208.XA CN107844853A (en) | 2017-10-23 | 2017-10-23 | It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710992208.XA CN107844853A (en) | 2017-10-23 | 2017-10-23 | It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107844853A true CN107844853A (en) | 2018-03-27 |
Family
ID=61662720
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710992208.XA Pending CN107844853A (en) | 2017-10-23 | 2017-10-23 | It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107844853A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108985475A (en) * | 2018-06-13 | 2018-12-11 | 厦门大学 | Net based on deep neural network about vehicle car service needing forecasting method |
CN110189182A (en) * | 2019-06-28 | 2019-08-30 | 重庆长安新能源汽车科技有限公司 | A kind of mileage anxiety management method based on car networking |
CN111667083A (en) * | 2020-06-11 | 2020-09-15 | 北京白龙马云行科技有限公司 | Network appointment vehicle pre-estimation determining method and device |
EP3716162A1 (en) * | 2019-03-29 | 2020-09-30 | Bayerische Motoren Werke Aktiengesellschaft | Neural-network-based dynamic matching |
CN112017004A (en) * | 2020-08-28 | 2020-12-01 | 北京嘀嘀无限科技发展有限公司 | Information display method and device and electronic equipment |
CN112101998A (en) * | 2020-09-15 | 2020-12-18 | 北京嘀嘀无限科技发展有限公司 | Residual value determining method, model obtaining method, device, equipment and storage medium |
CN113408877A (en) * | 2021-06-07 | 2021-09-17 | 北京百度网讯科技有限公司 | Network appointment information processing method, device, equipment and computer storage medium |
WO2021232203A1 (en) * | 2020-05-18 | 2021-11-25 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for online to offline services |
-
2017
- 2017-10-23 CN CN201710992208.XA patent/CN107844853A/en active Pending
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108985475A (en) * | 2018-06-13 | 2018-12-11 | 厦门大学 | Net based on deep neural network about vehicle car service needing forecasting method |
EP3716162A1 (en) * | 2019-03-29 | 2020-09-30 | Bayerische Motoren Werke Aktiengesellschaft | Neural-network-based dynamic matching |
CN110189182A (en) * | 2019-06-28 | 2019-08-30 | 重庆长安新能源汽车科技有限公司 | A kind of mileage anxiety management method based on car networking |
WO2021232203A1 (en) * | 2020-05-18 | 2021-11-25 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for online to offline services |
CN111667083A (en) * | 2020-06-11 | 2020-09-15 | 北京白龙马云行科技有限公司 | Network appointment vehicle pre-estimation determining method and device |
CN112017004A (en) * | 2020-08-28 | 2020-12-01 | 北京嘀嘀无限科技发展有限公司 | Information display method and device and electronic equipment |
CN112101998A (en) * | 2020-09-15 | 2020-12-18 | 北京嘀嘀无限科技发展有限公司 | Residual value determining method, model obtaining method, device, equipment and storage medium |
CN113408877A (en) * | 2021-06-07 | 2021-09-17 | 北京百度网讯科技有限公司 | Network appointment information processing method, device, equipment and computer storage medium |
WO2022257357A1 (en) * | 2021-06-07 | 2022-12-15 | 北京百度网讯科技有限公司 | Online car-hailing information processing method and apparatus, and device and computer storage medium |
EP4123527A4 (en) * | 2021-06-07 | 2023-08-30 | Beijing Baidu Netcom Science Technology Co., Ltd. | Online car-hailing information processing method and apparatus, and device and computer storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107844853A (en) | It is a kind of that the commending system by bus for reducing net about fare is predicted using dynamic price | |
Shaheen et al. | Shared ride services in North America: definitions, impacts, and the future of pooling | |
CN105303882B (en) | Parking management system and method based on Spatial dimensionality dynamic Car park payment strategy | |
Mavi et al. | Bus rapid transit (BRT): A simulation and multi criteria decision making (MCDM) approach | |
Henderson | EVs are not the answer: a mobility justice critique of electric vehicle transitions | |
Fagnant et al. | Dynamic ride-sharing and fleet sizing for a system of shared autonomous vehicles in Austin, Texas | |
Pojani et al. | Policy design for sustainable urban transport in the global south | |
CN102867408B (en) | Method for selecting bus trip route | |
May | Singapore: The development of a world class transport system | |
DE112018007300T5 (en) | ROUTE GUIDANCE WITH PERCEPTION OF THE SURROUNDINGS | |
Makarova et al. | Ensuring sustainability of public transport system through rational management | |
CN110210664B (en) | Deep learning method for short-term prediction of using behaviors of multiple individual vehicles | |
Kontou et al. | Reducing ridesourcing empty vehicle travel with future travel demand prediction | |
Chiou et al. | Sustainable consumption, production and infrastructure construction for operating and planning intercity passenger transport systems | |
CN113642758B (en) | Supply and demand prediction and scheduling method and device for shared electric automobile | |
Cornejo et al. | An approach to comprehensively evaluate potential park and ride facilities | |
CN114970944B (en) | Order matching and vehicle repositioning method based on multi-agent reinforcement learning | |
Tzamourani et al. | Exploring the social acceptance of transforming urban arterials to multimodal corridors. The case of Panepistimiou Avenue in Athens | |
CN116324838A (en) | System and method for scheduling shared rides through a taxi calling platform | |
Wang et al. | A user-based adaptive joint relocation model combining electric car-sharing and bicycle-sharing | |
Tajaddini et al. | Modeling the long-term regional impacts of autonomous vehicles: A case study of Victoria, Australia | |
Ribeiro Pimenta et al. | Long-term effects of autonomous vehicles on the built environment: a systematic scoping review towards conceptual frameworks | |
Li et al. | Short-term traffic demand prediction using graph convolutional neural networks | |
Xi | Data-driven optimization technologies for MaaS | |
Kong et al. | Optimal storage and loading zones within surface parking facilities for privately owned automated vehicles |
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: 20180327 |
|
WD01 | Invention patent application deemed withdrawn after publication |