[go: up one dir, main page]

CN110347937B - Intelligent passenger searching method for taxi - Google Patents

Intelligent passenger searching method for taxi Download PDF

Info

Publication number
CN110347937B
CN110347937B CN201910567942.0A CN201910567942A CN110347937B CN 110347937 B CN110347937 B CN 110347937B CN 201910567942 A CN201910567942 A CN 201910567942A CN 110347937 B CN110347937 B CN 110347937B
Authority
CN
China
Prior art keywords
passenger
taxi
time
searching
hotspot
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.)
Active
Application number
CN201910567942.0A
Other languages
Chinese (zh)
Other versions
CN110347937A (en
Inventor
王桐
沈昭晛
张乐君
李升波
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.)
Harbin Engineering University
Original Assignee
Harbin Engineering University
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 Harbin Engineering University filed Critical Harbin Engineering University
Priority to CN201910567942.0A priority Critical patent/CN110347937B/en
Publication of CN110347937A publication Critical patent/CN110347937A/en
Application granted granted Critical
Publication of CN110347937B publication Critical patent/CN110347937B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/025—Services making use of location information using location based information parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
    • H04W4/48—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for in-vehicle communication

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Databases & Information Systems (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Marketing (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Tourism & Hospitality (AREA)
  • Data Mining & Analysis (AREA)
  • General Business, Economics & Management (AREA)
  • Signal Processing (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Development Economics (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Mathematical Physics (AREA)
  • Entrepreneurship & Innovation (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

本发明公开了一种出租车智能寻客方法，包括：载客量预测：基于出租车历史轨迹数据对城市中的载客热点区域的载客量作预测，根据预测结果筛选出历史数据中与当天载客量相似的日期，将筛选出的日期生成时空Index；构建寻客指标数据库：基于出租车历史轨迹数据建立载客热点区域的寻客效率数据库和热点间驾驶时间数据库，所述寻客效率数据库包括热点的寻客时间、载客概率和载客收入；载客热点筛选：根据步骤一生成的时空Index从步骤二中筛选出对应日期的寻客效率数据库和热点间驾驶时间数据库，并对出租车前往不同热点的寻客效率进行均衡，筛选出最佳的热点区域。本发明的方法包含线上和线下两种处理过程，能够极大缩短推荐服务的计算时间。

Figure 201910567942

The invention discloses a method for intelligently searching for taxis. On the days with similar passenger load on the day, the selected dates will be used to generate a spatiotemporal index; build a customer-seeking index database: based on the historical taxi trajectory data, a customer-seeking efficiency database and a driving time database between hotspots are established based on the taxi historical trajectory data. The efficiency database includes the passenger-seeking time, passenger-carrying probability, and passenger-carrying income of hot spots; passenger-carrying hotspot screening: according to the spatiotemporal Index generated in step 1, filter out the passenger-seeking efficiency database and the driving time database between hot spots from step 2, and Balance the efficiency of taxis going to different hotspots, and screen out the best hotspots. The method of the present invention includes two processing processes, online and offline, and can greatly shorten the calculation time of the recommendation service.

Figure 201910567942

Description

Intelligent passenger searching method for taxi
Technical Field
The invention provides an intelligent taxi passenger searching method by combining historical taxi track data and real-time traffic information.
Background
In recent years, with the rapid development of communication technology and computer technology, an Intelligent Transportation System (ITS) is favored by many researchers, and Intelligent taxi technology is an important component of an Intelligent Transportation System and is receiving wide attention of researchers. Meanwhile, the demand of the passenger on the taxi is not only on waiting for the taxi, and more intelligent matching services are proposed successively, such as: taxi reservation service, carpooling service, optimal driving route service and the like, which are provided without analyzing the historical track data of the taxi. Since a large amount of taxi history track data is generated every day, rapid data analysis is urgent to meet the real-time performance of each service. Although the big data platform is an important means for increasing analysis efficiency, the analysis strategy of taxi historical track data is more critical, and a plurality of researchers adopt an offline historical track processing mode and an online service recommendation mode to meet the real-time performance of recommendation.
The first is to recommend the best passenger searching position for the passenger from the perspective of the passenger, so as to improve the passenger searching efficiency and provide the expected arrival time and the arrival probability of the taxi. And secondly, recommending an optimal passenger searching route for the driver from the perspective of the driver, analyzing a passenger searching hot spot area near the taxi, analyzing expected passenger searching time and passenger searching probability of the taxi going to different hot spots, and screening and recommending the passenger searching hot spots according to the two factors in order to save the passenger searching cost of the driver. Since taxi historical data generated every day are different, selection of different taxi historical data has a great influence on accuracy of recommended service, and currently, taxi historical data are divided into working day data and non-working day data in a relatively large manner in the field, but the division method is not fine enough. There are many factors affecting people's traveling, such as weather, holidays, time, etc., and in order to reasonably select historical data to realize recommendation service, the factors should be considered comprehensively, and reasonable historical data should be dynamically selected in real time, rather than considering one factor singly.
In the process of searching for passengers in a taxi, drivers with rich driving experiences can effectively find passenger carrying hot spot areas such as railway stations and movie theaters, and the number of the passengers in the areas changes along with time, so that the passenger carrying hot spot areas cannot be recommended to the drivers singly. And the change of the number of passengers in the passenger carrying hot spot and a passenger searching scheme which is not obvious to a taxi driver are hidden in the historical track data of the taxi. Therefore, the TCSFP passenger searching strategy is provided by integrating the offline processing process of the taxi historical track data and the online processing process of the real-time traffic information.
Disclosure of Invention
The invention aims to provide an intelligent Taxi Passenger searching method, which aims to solve the problem that a Taxi and a Passenger in a city are difficult to match and provides a Taxi Passenger searching Strategy (TCSFP) Based on Passenger capacity prediction by combining historical track data and real-time traffic information of the Taxi.
The invention is realized by the following technical scheme: an intelligent taxi passenger searching method comprises the following steps:
the method comprises the following steps: and (3) passenger capacity prediction: predicting passenger capacity of a passenger carrying hot spot area in a city based on historical taxi track data, screening out dates similar to the current passenger capacity in the historical data according to a prediction result, and generating a space-time Index for the screened dates;
step two: constructing a visitor finding index database: establishing a passenger searching efficiency database and a driving time database between hot spots of a passenger carrying hot spot area based on historical track data of the taxi, wherein the passenger searching efficiency database comprises passenger searching time, passenger carrying probability and passenger carrying income of the hot spots;
step three: passenger carrying hotspot screening: screening a passenger searching efficiency database and a driving time database between hot spots on corresponding dates from the second step according to the space-time Index generated in the first step, balancing passenger searching efficiency of taxies going to different hot spots, and screening out an optimal hot spot area, wherein the screening principle is shown as a formula (1):
Figure BDA0002110040750000021
wherein i represents the number of the area where the taxi is currently located, j represents the number of the jth hot spot area, m represents the mth time period of each day,
Figure BDA0002110040750000022
for the taxi to travel to the jth hotspot,
Figure BDA0002110040750000023
time, p, required for taxi to search for passenger at jth hot spotjmThe probability of carrying a passenger for the jth hotspot, rjmThe average passenger revenue for the jth hotspot.
Further, the step one comprises the following steps:
the method comprises the following steps: extracting all passenger boarding points from the historical data, and counting the number of passenger loads occurring in different time periods, as shown in formula (2):
Figure BDA0002110040750000024
wherein h isiDenotes the ith hotspot zone, djIndicating the jth day in the taxi history, L indicating that the day is divided into L time segments,
Figure BDA0002110040750000025
representing the number of passengers present during the kth time period;
the first step is: collecting hotspotsiAll passenger load data in the historical data, as shown in equation (3):
M(hi)＝{P(hi,d1),P(hi,d2),...,P(hi,dn)} (3)
assuming that the load of the mth time slot of the ith hot spot is predicted and the load of the first k time slots is used as the prediction feature, the machine learning feature and the label as shown in formula (4) and formula (5) can be constructed,
Figure BDA0002110040750000031
Figure BDA0002110040750000032
step one is three: the machine learning classification model is trained on the features and labels provided by equations (4) and (5). And finding out the passenger capacity close to the predicted result from the formula (5)
Figure BDA0002110040750000033
And selecting the most reasonable passenger searching index of the historical data analysis hotspot of the mth time period on the jth day.
Further, in the second step, specifically, the
Figure BDA0002110040750000034
pjk，rjkQuantization is performed, and the quantization result is shown in formula (6):
Figure BDA0002110040750000035
wherein
Figure BDA0002110040750000036
And
Figure BDA0002110040750000037
respectively are the passenger searching time, the passenger searching probability and the quantitative result of the passenger carrying income,
Figure BDA0002110040750000038
Figure BDA0002110040750000039
and
Figure BDA00021100407500000310
are all between 0 and k-1,
use of
Figure BDA00021100407500000311
Quantized results instead of seek time
Figure BDA00021100407500000312
Then, formula (7):
Figure BDA00021100407500000313
changing selection of targets to find
Figure BDA00021100407500000314
And
Figure BDA00021100407500000315
all with a large hot spot.
Further, the third step includes the following steps:
step three, firstly: setting weight for each hotspot, wherein the weight of the hotspot is the sum of passenger searching time, passenger searching probability and passenger searching income of the passenger carrying hotspot:
Figure BDA00021100407500000316
step three: sorting the weights of the passenger-carrying hotspots from big to small, and recording the sorting result as:
S(hi,m)＝(weight1,weight2,...,weightl)
(9)
and screening out the edge hot spots, and recommending the screened edge hot spots to a car rental driver.
The invention has the beneficial effects that: the invention starts from the perspective of recommending the best Passenger searching Strategy for the Taxi running empty, and provides a Passenger searching Strategy (Taxi cruise route Based on Passenger searching Passenger Volume, TCSFP) Based on Passenger capacity prediction by combining the historical track data of the Taxi and the online Passenger capacity information, wherein the Strategy is divided into two stages: a passenger capacity prediction stage and a taxi passenger carrying hot spot screening stage. And a space-time Index can be generated in the passenger capacity prediction stage, historical track data can be dynamically selected, and a passenger searching hotspot can be recommended for the taxi in the passenger carrying hotspot recommendation stage. Meanwhile, the strategy comprises two processing processes of online processing and offline processing, so that the calculation time of the recommendation service can be greatly shortened.
Drawings
FIG. 1 is a flow chart of a method for intelligently searching for a passenger in a taxi according to the present invention;
fig. 2 is a schematic diagram illustrating passenger-carrying hotspot screening.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the invention is realized by the following technical scheme: an intelligent taxi passenger searching method comprises the following steps:
the method comprises the following steps: and (3) passenger capacity prediction: predicting passenger capacity of a passenger carrying hot spot area in a city based on historical taxi track data, screening out dates similar to the current passenger capacity in the historical data according to a prediction result, and generating a space-time Index for the screened dates;
step two: constructing a visitor finding index database: establishing a passenger searching efficiency database and a driving time database between hot spots of a passenger carrying hot spot area based on historical track data of the taxi, wherein the passenger searching efficiency database comprises passenger searching time, passenger carrying probability and passenger carrying income of the hot spots;
step three: passenger carrying hotspot screening: screening a passenger searching efficiency database and a driving time database between hot spots on corresponding dates from the second step according to the space-time Index generated in the first step, balancing passenger searching efficiency of taxies going to different hot spots, and screening out an optimal hot spot area, wherein the screening principle is shown as a formula (1):
Figure BDA0002110040750000051
wherein i represents the number of the area where the taxi is currently located, j represents the number of the jth hot spot area, m represents the mth time period of each day,
Figure BDA0002110040750000052
for the taxi to travel to the jth hotspot,
Figure BDA0002110040750000053
time, p, required for taxi to search for passenger at jth hot spotjmThe probability of carrying a passenger for the jth hotspot, rjmThe average passenger revenue for the jth hotspot.
Specifically, in the first step, the processing result of dynamically selecting historical data from the database is realized through the prediction of passenger capacity; in the second step, the database is built, the passenger searching efficiency and the driving time of each hotspot are provided, and the recommendation efficiency of the third step is accelerated. The invention integrates the offline processing process of the taxi historical track data and the online processing process of the real-time traffic information, and provides a TCSFP passenger searching strategy, as shown in figure 1. The TCSFP passenger searching strategy is divided into an online processing process and an offline processing process, wherein the processing process above the dotted line in the figure 1 is the offline processing process of the taxi historical data, and the processing process below the dotted line is the online recommending process. The historical data processing is divided into two aspects, wherein a left-side dotted line frame represents a passenger capacity prediction stage, and a right-side dotted line frame represents a passenger hot spot recommendation stage. The taxi passenger searching time, the passenger carrying probability and the passenger carrying income are very key factors for screening the passenger carrying hot spots, the influence on the passenger searching performance is very large, and the factors are changed in real time, so that the factors need to be rapidly obtained and comprehensively evaluated.
Referring to fig. 1, in the present preferred embodiment, the first step includes the following steps:
the method comprises the following steps: extracting all passenger boarding points from the historical data, and counting the number of passenger loads occurring in different time periods, as shown in formula (2):
Figure BDA0002110040750000054
wherein h isiDenotes the ith hotspot zone, djIndicating the jth day in the taxi history, L indicating that the day is divided into L time segments,
Figure BDA0002110040750000055
representing the number of passengers present during the kth time period;
the first step is: collecting hotspotsiAll passenger load data in the historical data, as shown in equation (3):
M(hi)＝{P(hi,d1),P(hi,d2),...,P(hi,dn)} (3)
assuming that the load of the mth time slot of the ith hot spot is predicted and the load of the first k time slots is used as the prediction feature, the machine learning feature and the label as shown in formula (4) and formula (5) can be constructed,
Figure BDA0002110040750000056
Figure BDA0002110040750000062
step one is three: the machine learning classification model is trained on the features and labels provided by equations (4) and (5). And finding out the passenger capacity close to the predicted result from the formula (5)
Figure BDA0002110040750000063
The load capacity of the mth time period of the history data of the jth day is closest to the prediction result, so that the searching of the historical data analysis hot spot of the mth time period of the jth day is selectedThe guest index is most reasonable.
Specifically, as the travel of people is influenced by a plurality of factors, such as weather, holidays and the like, the selection of proper historical data for analysis is very critical in the process of finding the best passenger searching hot spot for the taxi. And (4) screening a date similar to the predicted passenger capacity result from the historical data, reasonably showing that the travel of the passengers on the date is similar to that of the current day, generating a space-time Index according to the screened date, and inputting the space-time Index into the step three to serve as a basis for selecting the historical data in the step three.
Referring to fig. 1, in the present preferred embodiment, in step two, specifically, the taxi is started from the starting position hiGo to the hotspot h for carrying passengersjThe passenger searching time, the passenger searching probability and the passenger carrying income are respectively different types of indexes, so the passenger searching time, the passenger searching probability and the passenger carrying income are required to be adjusted
Figure BDA0002110040750000064
pjk，rjkQuantization is performed, and the quantization result is shown in formula (6):
Figure BDA0002110040750000065
wherein
Figure BDA0002110040750000066
And
Figure BDA0002110040750000067
respectively the passenger searching time, the passenger searching probability and the passenger carrying income, and the values are between 0 and k-1,
the passenger searching time is as small as possible, and the passenger searching probability and the passenger carrying income are as high as possible, so the passenger searching method is used
Figure BDA0002110040750000068
The quantization result instead of the time for searching the passenger can still be ensured
Figure BDA0002110040750000069
Between 0 and k-1 and the selection target is changed to find
Figure BDA00021100407500000610
And
Figure BDA00021100407500000611
the hot spots are all large, and the formula (7) is shown:
Figure BDA00021100407500000612
as shown in FIG. 2, the "edge hotspots" far from the three coordinate axes should be selected as much as possible, and the "interior hotspots" near the far points should be deleted, so that the hotspots h are deleted2。
Specifically, because the recommendation of the taxi passenger searching scheme needs to meet the real-time performance, the processing of the historical track data is finished on line. The passenger carrying probability and the passenger carrying income of the taxi can directly correspond to the passenger carrying probability and the passenger carrying income of each hotspot, and can be relatively easily extracted from historical data, and the passenger searching time of the taxi consists of the time required by a driver to go to different hotspots and the average passenger searching time of each hotspot, so that the attributes of the hotspots are divided into: the hot spot passenger searching time, the passenger carrying probability and the passenger carrying income. The time required by the taxi to go to the hot spot can be obtained from a driving time database between the hot spots, and the passenger searching time of the taxi is the sum of the time required by the taxi to go to the hot spot and the average passenger searching time of the hot spot. This step belongs to the off-line treatment process.
Referring to fig. 1, in the preferred embodiment of this section, the following steps are included in the third step:
step three, firstly: setting weight for each hotspot, wherein the weight of the hotspot is the sum of passenger searching time, passenger searching probability and passenger searching income of the passenger carrying hotspot:
Figure BDA0002110040750000071
step three: sorting the weights of the passenger-carrying hotspots from big to small, and recording the sorting result as:
S(hi,m)＝(weight1,weight2,...,weightl) (9)
step three: for the sorted result S (h)iAnd m) judging whether each element in m) is an edge hotspot:
1. due to weight1For the maximum weight, it can be determined that the hotspot to which the hotspot belongs is an edge hotspot. The following was demonstrated:
if the hot spot haHas the greatest weight, and haBelonging to an "internal" hotspot, then there must be a hotspot hbThe passenger searching time, the passenger searching probability and the passenger carrying income are all larger than the hotspot haThus hot spot hbWeight (h) of (c)i,hbK) is greater than haWeight (h) of (c)i,haK), this is in combination with haWith the greatest weight contradiction, haThe result is an edge hot spot after the verification.
2. If weight2The visitor searching time, the visitor searching probability and the visitor carrying income of the hotspot are all less than weight1At the hot spot, then weight2The hot spot is an internal hot spot; otherwise, weight2The hot spots are edge hot spots, which are proved as follows:
if weight2The hot spot is an internal hot spot, and then a hot spot h existsbThe passenger searching time, the passenger searching probability and the passenger carrying income are all larger than weight2At a hot spot, therefore hbIs greater than weight2. And is greater than weight2Has weight only1This contradicts the hypothesis, and is confirmed.
By analogy to S (h)iAnd k) comparing each element in the k) with the selected edge hotspots, judging one by one, screening the edge hotspots, and recommending the screened edge hotspots to the car rental driver.
Specifically, as can be seen from fig. 1, the taxi passenger searching request only goes through one step, and other processing procedures are all completed on line or in advance, so that the time spent by the taxi passenger searching request can be greatly shortened.

Claims (4)

1. An intelligent passenger searching method for a taxi is characterized by comprising the following steps:
the method comprises the following steps: and (3) passenger capacity prediction: predicting passenger capacity of a passenger carrying hot spot area in a city based on historical taxi track data, screening out dates similar to the current passenger capacity in the historical data according to a prediction result, and generating a space-time Index for the screened dates;
step two: constructing a visitor finding index database: establishing a passenger searching efficiency database and a driving time database between hot spots of a passenger carrying hot spot area based on historical track data of the taxi, wherein the passenger searching efficiency database comprises passenger searching time, passenger carrying probability and passenger carrying income of the hot spots;
step three: passenger carrying hotspot screening: screening a passenger searching efficiency database and a driving time database between hot spots on corresponding dates from the second step according to the space-time Index generated in the first step, balancing passenger searching efficiency of taxies going to different hot spots, and screening out an optimal hot spot area, wherein the screening principle is shown as a formula (1):
Figure FDA0002110040740000011
wherein i represents the number of the area where the taxi is currently located, j represents the number of the jth hot spot area, m represents the mth time period of each day,
Figure FDA0002110040740000012
for the taxi to travel to the jth hotspot,
Figure FDA0002110040740000013
the time required for the taxi to search for the taxi in the jth hot spot,pjmthe probability of carrying a passenger for the jth hotspot, rjmThe average passenger revenue for the jth hotspot.
2. The intelligent passenger searching method for the taxi according to claim 1, wherein the first step comprises the following steps:
the method comprises the following steps: extracting all passenger boarding points from the historical data, and counting the number of passenger loads occurring in different time periods, as shown in formula (2):
Figure FDA0002110040740000014
wherein h isiDenotes the ith hotspot zone, djIndicating the jth day in the taxi history, L indicating that the day is divided into L time segments,
Figure FDA0002110040740000015
representing the number of passengers present during the kth time period;
the first step is: collecting hotspotsiAll passenger load data in the historical data, as shown in equation (3):
M(hi)＝{P(hi,d1),P(hi,d2),...,P(hi,dn)} (3)
assuming that the load of the mth time slot of the ith hot spot is predicted and the load of the first k time slots is used as the prediction feature, the machine learning feature and the label as shown in formula (4) and formula (5) can be constructed,
Figure FDA0002110040740000016
Figure FDA0002110040740000021
step one is three: according toTraining machine learning classification model by using characteristics and labels provided by formula (4) and formula (5), and finding out passenger capacity close to predicted result from formula (5)
Figure FDA0002110040740000022
And selecting the most reasonable passenger searching index of the historical data analysis hotspot of the mth time period on the jth day.
3. The method as claimed in claim 1, wherein in step two, specifically, for each taxi
Figure FDA0002110040740000023
pjk，rjkQuantization is performed, and the quantization result is shown in formula (6):
Figure FDA0002110040740000024
wherein
Figure FDA0002110040740000025
And
Figure FDA0002110040740000026
respectively are the passenger searching time, the passenger searching probability and the quantitative result of the passenger carrying income,
Figure FDA0002110040740000027
Figure FDA0002110040740000028
and
Figure FDA0002110040740000029
are all between 0 and k-1,
use of
Figure FDA00021100407400000210
Quantized results instead of seek time
Figure FDA00021100407400000211
Then, formula (7):
Figure FDA00021100407400000212
changing selection of targets to find
Figure FDA00021100407400000213
And
Figure FDA00021100407400000214
all with a large hot spot.
4. The intelligent passenger searching method for the taxi according to claim 1, wherein the third step comprises the following steps:
step three, firstly: setting weight for each hotspot, wherein the weight of the hotspot is the sum of passenger searching time, passenger searching probability and passenger searching income of the passenger carrying hotspot:
Figure FDA00021100407400000215
step three: sorting the weights of the passenger-carrying hotspots from big to small, and recording the sorting result as:
S(hi,m)＝(weight1,weight2,...,weightl) (9)
and screening out the edge hot spots, and recommending the screened edge hot spots to a car rental driver.
CN201910567942.0A 2019-06-27 2019-06-27 Intelligent passenger searching method for taxi Active CN110347937B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910567942.0A CN110347937B (en) 2019-06-27 2019-06-27 Intelligent passenger searching method for taxi

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910567942.0A CN110347937B (en) 2019-06-27 2019-06-27 Intelligent passenger searching method for taxi

Publications (2)

Publication Number Publication Date
CN110347937A CN110347937A (en) 2019-10-18
CN110347937B true CN110347937B (en) 2022-02-11

Family

ID=68176772

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910567942.0A Active CN110347937B (en) 2019-06-27 2019-06-27 Intelligent passenger searching method for taxi

Country Status (1)

Country Link
CN (1) CN110347937B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112052405B (en) * 2020-08-24 2021-12-24 杭州电子科技大学 A method for recommending customer-seeking areas based on driver experience
CN112070529A (en) * 2020-08-24 2020-12-11 贵州民族大学 Parallel prediction method, system, terminal and computer storage medium for passenger hotspots

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6456207B1 (en) * 2001-02-20 2002-09-24 John Yen Intelligent taxi total service system
CN103544834A (en) * 2013-11-14 2014-01-29 孙林 Taxi customer seeking strategy selection method based on GPS track
CN103854472A (en) * 2012-12-05 2014-06-11 深圳先进技术研究院 Taxi cloud-intelligent scheduling method and system
CN107590175A (en) * 2017-07-31 2018-01-16 同济大学 A kind of taxi pickup point based on deep learning recommends method
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

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001022777A (en) * 1999-07-09 2001-01-26 Hitachi Ltd Data analysis method
JP2006040007A (en) * 2004-07-28 2006-02-09 Nobutoshi Umeda Taxi allocating system and allocating method
US7577665B2 (en) * 2005-09-14 2009-08-18 Jumptap, Inc. User characteristic influenced search results
US7603360B2 (en) * 2005-09-14 2009-10-13 Jumptap, Inc. Location influenced search results
CN102572697A (en) * 2012-02-26 2012-07-11 沈哲 Taxi calling system and taxi calling method based on handheld mobile terminals

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6456207B1 (en) * 2001-02-20 2002-09-24 John Yen Intelligent taxi total service system
CN103854472A (en) * 2012-12-05 2014-06-11 深圳先进技术研究院 Taxi cloud-intelligent scheduling method and system
CN103544834A (en) * 2013-11-14 2014-01-29 孙林 Taxi customer seeking strategy selection method based on GPS track
CN107590175A (en) * 2017-07-31 2018-01-16 同济大学 A kind of taxi pickup point based on deep learning recommends method
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

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
基于出租车GPS数据的高效益寻客推荐策略研究;刘丽;《中国优秀硕士学位论文全文数据库工程科技||辑》;20170915;C034-63 *
基于北京市载客热点区的出租车出行需求研究;王贝贝;《中国优秀硕士学位论文全文数据库工程科技||辑》;20180615;C034-467 *

Also Published As

Publication number Publication date
CN110347937A (en) 2019-10-18

Similar Documents

Publication Publication Date Title
US12094344B2 (en) System for providing big data-based artificial intelligence automatic allocation matching service using taxi demand prediction
CN106875066B (en) Vehicle travel behavior prediction method, device, server and storage medium
Ma et al. T-share: A large-scale dynamic taxi ridesharing service
US7679531B2 (en) System and method of optimizing a fixed-route transit network
CN102324128B (en) Method for predicting OD (Origin-Destination) passenger flow among bus stations on basis of IC (Integrated Circuit)-card record and device
CN107423837A (en) The Intelligent planning method and system of tourism route
CN117829536B (en) Riding prediction and traffic system based on user preference
CN110836675B (en) Decision tree-based automatic driving search decision method
CN111932341B (en) A method and system for determining a carpooling order
CN107392389A (en) Taxi dispatching processing method based on ARIMA models
CN106372674A (en) Driver classification method and device applied to online taxi booking service platform
CN116663811A (en) Scheduling matching method and device for reciprocating dynamic carpooling of inter-city passenger transport
CN112601187B (en) Bus frequent passenger prediction method and system based on mobile phone signaling
CN115186905B (en) Multitask optimization method and system for inter-city network taxi-booking path planning
CN110347937B (en) Intelligent passenger searching method for taxi
CN119313060A (en) A dynamic carpooling scheduling method for driverless electric taxis
Wang et al. Taxi-cruising recommendation via real-time information and historical trajectory data
CN115127569A (en) Personalized intermodal navigation method and system
Abdelmoumène et al. Dynamic matching optimization in ridesharing system based on reinforcement learning
CN113850295B (en) Site selection prediction method based on Weighted-LeaderRank and GMM clustering
CN119227926A (en) Railway transfer travel route optimization method, device, equipment and storage medium
CN113379159B (en) Taxi driver passenger searching route recommendation method based on gray model and Markov decision process
CN117592788B (en) A method and device for identifying public transportation operation risks
Rodrigues et al. Hybrid context-aware multimodal routing
Wan et al. Online recommendation system for autonomous and human-driven ride-hailing taxi services

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
GR01 Patent grant
GR01 Patent grant
CB03 Change of inventor or designer information

Inventor after: Wang Tong

Inventor after: Shen Zhaoxian

Inventor after: Zhang Lejun

Inventor before: Wang Tong

Inventor before: Shen Zhaoxian

Inventor before: Zhang Lejun

Inventor before: Li Shengbo

CB03 Change of inventor or designer information