CN114819414B - Block demand prediction method, system and computer storage medium - Google Patents
Block demand prediction method, system and computer storage medium Download PDFInfo
- Publication number
- CN114819414B CN114819414B CN202210722389.5A CN202210722389A CN114819414B CN 114819414 B CN114819414 B CN 114819414B CN 202210722389 A CN202210722389 A CN 202210722389A CN 114819414 B CN114819414 B CN 114819414B
- Authority
- CN
- China
- Prior art keywords
- block
- time
- vehicle
- real
- blocks
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 239000013598 vector Substances 0.000 claims abstract description 58
- 230000003993 interaction Effects 0.000 claims abstract description 24
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 238000004590 computer program Methods 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 3
- 230000000875 corresponding effect Effects 0.000 description 18
- 230000009286 beneficial effect Effects 0.000 description 7
- 230000002123 temporal effect Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000002452 interceptive effect Effects 0.000 description 4
- 230000002596 correlated effect Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013277 forecasting method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
Images
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
- 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
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
- G06N3/08—Learning methods
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of vehicle scheduling, in particular to a block demand prediction method, a system and a computer storage medium, comprising the following steps: acquiring spatial relation characteristics, real-time characteristics and historical order characteristics of a current block and respectively preprocessing the spatial relation characteristics, the real-time characteristics and the historical order characteristics; generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features; and generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relationship characteristics. The invention can accurately estimate the requirements of each block by combining a plurality of characteristics.
Description
Technical Field
The present invention relates to the field of vehicle scheduling technologies, and in particular, to a block demand prediction method, a block demand prediction system, and a computer storage medium.
Background
In a short-distance travel scene, a shared electric bicycle is a 'what you see is what you get' travel mode, whether a vehicle exists in one place or not is particularly important for user experience and user demand meeting, but the vehicle delivery can not be carried out in a 'saturated mode', and how to redistribute allocated limited vehicle resources in space and time is achieved, so that more vehicle values are brought, and meeting the short-distance travel demand of a user is an important problem that needs to be solved urgently.
The current estimation of the demand of each block is estimated based on historical order data and other data of each block, that is, the average of the past order quantity of the period is used for estimating the demand of the future period, but the demand estimated by the method is not accurate.
Disclosure of Invention
In order to solve the problem of inaccurate prediction of vehicle demands for blocks, the invention provides a block demand prediction method, a system and a computer storage medium.
In order to solve the technical problems, the invention provides the following technical scheme: a block demand prediction method for guiding vehicle scheduling comprises the following steps:
acquiring spatial relationship characteristics, real-time characteristics and historical order characteristics of a current block and respectively preprocessing the characteristics;
generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relation characteristics;
the method for acquiring and preprocessing the spatial relationship features comprises the following steps:
acquiring the block correlation between the current block and the rest blocks;
and screening the rest blocks of which the correlation with the blocks is in a preset range, extracting the characteristics of the screened rest blocks as spatial relationship characteristics, and generating corresponding vector representations.
Preferably, the spatial dimension features comprise one or more of block number, POI data, heat value, road network data, AOI data, resident population density, number of nearby stations, and people flow rate; the time dimension characteristics comprise one or more of time period serial numbers, holiday information and weather information.
Preferably, the spatial feature vector and the temporal feature vector are interacted by adopting a Dense Layer or an FM Layer to form the spatial-temporal interaction feature.
Preferably, the steps of obtaining the real-time characteristics and the historical order characteristics and correspondingly preprocessing the real-time characteristics and the historical order characteristics respectively comprise:
acquiring historical order quantity of the block, selecting the order quantity in the current preset time as real-time characteristics, extracting the real-time characteristics by adopting a time series model or an LSTM (least partial real-time metric) to obtain real-time characteristic vectors, or directly adopting at least partial real-time characteristics as the real-time characteristic vectors for representation;
and obtaining the total order amount and the average order value in a preset period based on the historical order amount as historical order characteristics, and directly representing the historical order characteristics as vectors.
Preferably, the type of the tile relevance comprises one or more of a geographical relevance between tiles, a vehicle traffic relevance and a POI relevance.
Preferably, after generating the block vehicle inflow amount, the vehicle outflow amount and the block demand prediction result, the method further comprises the following steps:
acquiring the current number of vehicles in the block, and judging whether the block contains a station or not;
if the station exists, calculating the number of required gaps of the block, wherein the number of required gaps of the block is equal to the number obtained by subtracting the inflow of vehicles from the required amount of the block and then subtracting the current number of vehicles in the block;
and if no station exists, calculating the surplus quantity of the blocks, wherein the surplus quantity of the blocks is equal to the sum of the current vehicle number of the blocks and the vehicle inflow amount and then the vehicle outflow amount.
Preferably, the vehicles move in the blocks with the block demand gap number larger than 0, and the vehicles move out the blocks with the block surplus number larger than 0.
In order to solve the above technical problems, the present invention provides another technical solution as follows: a block demand prediction system for implementing the block demand prediction method described above, comprising the following modules:
an acquisition module: acquiring spatial relation characteristics, real-time characteristics and historical order characteristics of a current block and respectively and correspondingly preprocessing the spatial relation characteristics, the real-time characteristics and the historical order characteristics;
an interaction module: generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
a prediction module: and generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relation characteristics.
In order to solve the above technical problems, the present invention provides another technical solution as follows: a computer storage medium, on which a computer program is stored, when the computer program runs on a computer, the block demand prediction method as described above is implemented.
Compared with the prior art, the block demand prediction method, the block demand prediction system and the computer storage medium have the following beneficial effects:
1. according to the block demand prediction method provided by the invention, the spatial characteristics and the temporal characteristics are interacted to form new interactive characteristics, and the prediction result is obtained through the neural network by combining with other characteristics, so that certain correlation exists between the temporal characteristics and the spatial characteristics, and a large amount of practical data shows that the accuracy of the prediction result can be improved after the characteristics are interactively correlated.
2. According to the block demand prediction method provided by the invention, various spatial characteristics of the blocks are utilized in prediction, and understandably, the characteristics of the blocks are greatly different, so that the prediction result obtained by considering the characteristics of the blocks, which influence the vehicle demand, is closer to the actual situation.
3. In the block demand prediction method provided by the invention, the spatial relationship characteristics are also expressed as the characteristics of the blocks having correlation with the prediction blocks, and the correlation between the related blocks, namely the correlation between the blocks can influence the demand of the actual situation, so that the correlation between the blocks is considered to improve the prediction accuracy.
4. According to the block demand forecasting method provided by the invention, after the block vehicle inflow amount, the vehicle outflow amount and the block demand forecasting result are generated, the demand gap number of the obtained block can be accurately calculated through the actual current number of the blocks.
5. According to the block demand prediction method provided by the invention, it can be understood that the station block calculation demand gap is used for paying attention to the blocks with gap values larger than 0 and guiding to move in the vehicle, and the non-station block calculation surplus is used for paying attention to the blocks with surplus values larger than 0 and guiding to move out the vehicle.
6. The embodiment of the present invention further provides a block demand prediction system, which has the same beneficial effects as the block demand prediction method described above, and is not described herein again.
7. The embodiment of the present invention further provides a computer storage medium, which has the same beneficial effects as the block demand prediction method described above, and is not described herein again.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
Fig. 1 is a flowchart illustrating a block demand prediction method according to a first embodiment of the present invention.
Fig. 2 is a flowchart illustrating a first step of S1 of a block demand prediction method according to a first embodiment of the present invention.
Fig. 3 is a flowchart illustrating a step S1 of a block demand prediction method according to a first embodiment of the present invention.
Fig. 4 is a flowchart illustrating steps after step S3 of a block demand prediction method according to a first embodiment of the present invention.
Fig. 5 is a block diagram of a block demand prediction system according to a second embodiment of the present invention.
Description of the figures:
1. a block demand prediction system;
10. an acquisition module; 20. an interaction module; 30. and a prediction module.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, a first embodiment of the present invention provides a block demand prediction method for guiding vehicle scheduling, including the following steps:
s1: acquiring spatial relation characteristics, real-time characteristics and historical order characteristics of a current block and respectively preprocessing the spatial relation characteristics, the real-time characteristics and the historical order characteristics;
s2: generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
s3: and generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relation characteristics.
It should be noted that the block is a plurality of small areas divided according to a mining algorithm based on a large area of the open shared electric bicycle operation service in the city, the small areas may be irregular in shape, and the shape, the size and the like of each small area may be the same or different; meanwhile, each block can also be used for reflecting the requirements of different areas in the city on the shared electric bicycle.
The space also represents the block, that is, the spatial relationship feature also represents the correlation between blocks, and the spatial dimension feature also represents the feature of each block.
According to the invention, after the block vehicle inflow, the vehicle outflow and the block demand of the station are predicted, the shunting demand of the block can be obtained through calculation, and vehicle operation and maintenance personnel can carry out shunting planning according to the demand of each block, so that the vehicle scheduling can be conveniently guided.
It should be noted that, in the prediction, the time is divided into a plurality of time periods, and what is predicted is the result of each time period, where the vehicle inflow amount represents the number of vehicle inflow blocks in a certain period, the vehicle outflow amount represents the number of vehicle outflow blocks in a certain period, and the block demand amount represents the number of vehicles that need to be used in a certain period block.
Specifically, the spatial dimension features include one or more of block number, POI data, heat value, road network data, AOI data, resident population density, number of nearby stations, and people flow rate.
The block numbers are only used to distinguish different blocks, and the rest of POI data, heat value, road network data, AOI data, resident population density, number of nearby stations, people flow rate, etc. are obtained according to the actual situation of each block, and can be obtained by means of a platform or a specific APP, and these data are all used to reflect the characteristics of the corresponding block, and moreover, the considered characteristics all affect the requirements of this area, such as: in the area a, the POI data of the area a includes a plurality of tourist attractions, and the corresponding area a may attract more tourists to go, so that the passenger flow is larger, and the demand of the area on the electric bicycle sharing is larger; the same reasoning applies to the rest of the data, and the accuracy of the prediction result is improved by considering the information.
Further, the time dimension characteristics comprise one or more of a time period serial number, holiday information and weather information.
It should be noted that the time period number represents a number for each time period, and the example is as follows: 24 hours in 1 day are divided into 48 time periods according to 0.5 hour as a granularity, namely, for example, 12.00-12.30 are one time period, a sequence number is set for distinguishing different time periods corresponding to each time period, and the time period corresponding to the sequence number can be known only according to the sequence number in the following process.
It can be understood that by interacting the temporal feature and the spatial feature to form a new interactive feature, in practice, by observing a large amount of sample data, it can be found that the correlation between some features and the label is improved after the features are correlated, and the prediction accuracy can be improved, for example: predicting the demand of the area A in the future time interval B, and only giving out spatial dimension characteristics that the area A contains tourist attractions, only considering the factors of the tourist attractions in the prediction process, and assuming that the predicted result is 50; and if the time dimension characteristic is given again that the weather in the time interval B is rainstorm, the interactive characteristic obtained after interaction is that the area A contains tourist attractions and the weather in the time interval B is rainstorm, the trip of people is influenced by the rainstorm is also considered in the prediction, and the prediction result of the area may be 10, so that the accuracy of the prediction result is improved by interacting the time characteristic and the space characteristic.
It should be noted that the preprocessing of the corresponding features in step S1 is to obtain vector representations of the corresponding features, the corresponding vector representations can be generated by the corresponding embedding layers, and the corresponding vectors are also generated by the embedding layers in step S2.
In general, in the prediction, a plurality of blocks are simultaneously predicted at the same time, or prediction results of all blocks of a city with shared electric bicycle service opened all day after the prediction is performed in advance are obtained, and if the prediction result of one block is required, the query can be performed according to the prediction result.
Specifically, in the present embodiment, a density Layer or an FM Layer is used for interacting the spatial feature vector and the temporal feature vector to form the spatial-temporal interaction feature.
The Dense Layer is also a Dense Layer (full connection Layer) and is used for combining the previous features, the FM Layer can also be called a factorization machine and can perform inner product on feature vectors to construct a feature cross pair, and a second-order expression of the feature cross pair is as follows when two features are crossed:
(ii) a Wherein n represents the feature number of the sample, xi is the value of the ith feature, xi is the value of the jth feature, and W0, wi and Wij are model parameters; from the formulation, the first two parts represent linear combinations, and the last part represents cross terms, i.e., combinations of features.
Further, in step S3, the feature interaction may be performed first, and then the prediction result is generated, so as to improve the accuracy of the prediction result.
Further, referring to fig. 2, the steps of obtaining the real-time characteristics and the historical order characteristics and performing the pre-processing correspondingly respectively include:
s11: acquiring historical order quantity of a block, selecting the order quantity in the current preset time as a real-time feature, extracting the real-time feature by adopting a time series model or LSTM to obtain a real-time feature vector, or directly expressing by adopting at least part of the real-time feature as the real-time feature vector;
s12: and obtaining the total order amount and the average order value in a preset period based on the historical order amount as historical order characteristics, and directly inputting the historical order characteristics as vectors into the neural network.
The order quantity represents the riding history data of the user, namely the user forms an order from unlocking the vehicle to locking the vehicle after riding, and the order can reflect the riding route, riding time and the starting point and the terminal point of riding of the user; the order may be used to determine which block the vehicle has ridden from and to.
In step S11, at least part of the real-time features are directly used as the real-time feature vector for representing, which is equivalent to that the real-time features do not need to be subjected to other data processing, and can be directly used as vectors.
It can be understood that, in order to predict the demand of the block at a future time, the order quantity that is relatively close to the predicted time period is generally used as the real-time characteristic in the present invention, so that the order quantity of the current block at the latest time period, such as the previous 3 hours, the previous 5 hours, and the like, is generally taken as the basis in step S11, without specific limitation, and according to the actual situation, since the order quantity is obtained according to the actual situation, the order quantity of the block at each time period, that is, the actual vehicle outflow quantity, the vehicle inflow quantity, can be obtained based on the order, and a part or all of the order quantities can be directly input as a vector, for example, the total vehicle outflow quantity, the total vehicle inflow quantity, and the average inflow quantity and/or the average outflow quantity of each time period of the previous five time periods can be directly input as a vector.
It is understood that the historical order characteristics are also obtained by using the historical order quantity, and generally the characteristics are considered to be comprehensive, so the data is relatively large, and the examples are as follows: adopting a historical period, such as the average value of orders in the previous month, the total order quantity and the like; alternatively, assuming that the demand of a certain period is predicted, the order quantity of the historical order in the previous month in the period may be used as the historical order feature, and the order quantity data may be directly input as the feature vector.
The real-time characteristic and the historical order characteristic are real data of each block in the past time period, and the actual demand condition of each block can be reflected more intuitively, so that the accuracy of the prediction result is improved.
Further, referring to fig. 3, the obtaining and preprocessing the spatial relationship features includes the following steps:
s13: acquiring the block correlation between the current block and the rest blocks;
s14: and screening the rest blocks of which the correlation with the blocks is in a preset range, extracting the characteristics of the screened blocks as spatial relationship characteristics, and generating corresponding vector representations.
Specifically, the type of relevance includes one or more of geographic relevance between the tiles, vehicle traffic relevance, and POI relevance.
It can be understood that the correlation between the blocks can be determined according to the spatial dimension characteristics of the blocks, for example, if two blocks have similar POIs, such as schools, banks, and the like, it indicates that the POIs of the two blocks have correlation, and if two blocks have bus stops, and a certain bus passes through the two bus stops at the same time, the POI correlation between the two blocks is relatively large.
The geographic relevance indicates whether the blocks are adjacent in geographic position, and the adjacent blocks indicate that the blocks are geographically relevant; the vehicle circulation relevance can be obtained according to historical orders, namely, the vehicle rides out of a certain block and rides into another block to finish riding, and the two blocks form the vehicle circulation relevance.
It can be understood that, generally, features of several blocks with high correlation with the block to be predicted are screened out as spatial relationship features, for example, historical order quantity and spatial dimension features of the blocks can be input as the spatial relationship features, and since the spatial relationship features are also features of other blocks, a vector generation mode can be the same as that described above, that is, historical order quantity can also be directly input into the neural network model as a vector, and for the spatial dimension features, an embedding layer can be used to generate corresponding vectors.
Therefore, the correlation between the blocks can influence the actual requirements of the blocks, and the like, so that the accuracy of the estimation result is improved.
Referring to fig. 4, after generating the block vehicle inflow, vehicle outflow, and block demand prediction results, the method further includes the following steps:
s31: acquiring the current number of vehicles in the block, and judging whether the block contains a station or not;
s32: if there is a station, calculating the number of gaps required by the block, wherein the number of the gaps required by the block is equal to the number obtained by subtracting the vehicle inflow from the block demand and then subtracting the current number of the vehicles in the block;
s33: and if no station exists, calculating the surplus quantity of the blocks, wherein the surplus quantity of the blocks is equal to the sum of the current vehicle number of the blocks and the vehicle inflow amount and then the vehicle outflow amount.
It should be noted that the stations are mined by an algorithm in advance, not all the blocks correspondingly include the stations, and when shunting, the blocks including the stations are generally scheduled according to the estimated demand.
It can be understood that a station is used as a basic dispatching unit, the dispatching is generally carried out on the station, and a block without the station is equivalent to a block without the dispatching point, so that vehicles are not dispatched into the block without the station, but surplus vehicles in the block can be dispatched to meet other stations with stations and demands, wherein surplus can be understood that the number of the vehicles in the block currently meets the demands of the current area, and the exceeding number is a surplus number beyond the demands, and the surplus vehicles are dispatched to improve the use efficiency of the vehicles.
The current vehicle number of the block can be obtained through GPS positioning, it can be understood that only the vehicle inflow, the vehicle outflow, and the block demand of each time period block are estimated in advance, in actual application, the number of gaps or the surplus number of each block can be calculated according to the current vehicle number of each block in real time and the estimated vehicle inflow, vehicle outflow, and block demand of each block, for example, if we want to meet the demand of a block a of a time period 18.00-18.30, we can judge whether the vehicle of the block can meet the demand of the time period 18.00-18.30 according to the current vehicle number of the block a of the time period 17.50 and the estimated vehicle inflow, vehicle outflow, or block demand of the block a of the time period 18.00-18.30, and if not, there can be enough time for vehicle scheduling; when the number of vehicle gaps or the surplus number is calculated, adjustment can be made according to actual conditions, and the current number of vehicles in a block with proper time is selected for calculation.
Specifically, the vehicles are moved in the blocks with the number of the gaps required by the blocks larger than 0, and the vehicles are moved out of the blocks with the surplus number of the blocks larger than 0.
It should be noted that vehicles moved from the blocks with the surplus number greater than 0 can be used for scheduling to the blocks with the block demand gap number greater than 0 to meet the demand, so as to improve the vehicle utilization rate, and the moved number can be the calculated surplus number or less than the calculated surplus number, and generally does not exceed the surplus number, so that the situation that the vehicles in the area cannot meet the demand of the user later is avoided.
Referring to fig. 5, a block demand prediction system 1 for implementing the block demand prediction method according to a second embodiment of the present invention includes the following modules:
the acquisition module 10: acquiring spatial relation characteristics, real-time characteristics and historical order characteristics of a current block and respectively and correspondingly preprocessing the spatial relation characteristics, the real-time characteristics and the historical order characteristics;
the interaction module 20: generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
the prediction module 30: and generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relation characteristics.
The above modules correspondingly implement the steps of the block demand prediction method, and it is within the scope of the present invention to integrate or separate the modules.
The third embodiment of the present invention further provides a computer storage medium, which stores a computer program, and when the computer program runs on a computer, the block demand prediction method described above is implemented.
The computer storage medium has the same beneficial effects as the block demand prediction method, and can realize the corresponding steps of the block demand prediction method, which are not described herein again.
In the embodiments provided herein, it should be understood that "B corresponding to a" means that B is associated with a, from which B can be determined. It should also be understood, however, that determining B from a does not mean determining B from a alone, but may also be determined from a and/or other information.
It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. Those skilled in the art should also appreciate that the embodiments described in this specification are exemplary and alternative embodiments, and that the acts and modules illustrated are not required in order to practice the invention.
In various embodiments of the present invention, it should be understood that the sequence numbers of the above-mentioned processes do not imply an inevitable order of execution, and the execution order of the processes should be determined by their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
The flowchart and block diagrams in the figures of the present application illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
Compared with the prior art, the block demand prediction method, the system and the computer storage medium provided by the invention have the following beneficial effects:
1. according to the block demand prediction method provided by the invention, the spatial characteristics and the temporal characteristics are interacted to form new interactive characteristics, and the prediction result is obtained through the neural network in combination with other characteristics, so that certain correlation exists between the temporal characteristics and the spatial characteristics, and a large amount of practical data shows that the accuracy of the prediction result can be improved after the characteristics are interactively correlated.
2. According to the block demand prediction method provided by the invention, various spatial characteristics of the blocks are utilized in prediction, and understandably, the characteristics of the blocks are greatly different, so that the prediction result obtained by considering the characteristics of the blocks, which influence the vehicle demand, is closer to the actual situation.
3. In the block demand prediction method provided by the invention, the spatial relationship characteristics are also expressed as the characteristics of the blocks having correlation with the prediction blocks, and the correlation between the related blocks, namely the correlation between the blocks can influence the demand of the actual situation, so that the correlation between the blocks is considered to improve the prediction accuracy.
4. According to the block demand prediction method provided by the invention, after the block vehicle inflow amount, the vehicle outflow amount and the block demand prediction result are generated, the acquired demand gap number of the block can be accurately calculated through the actual current number of the blocks.
5. According to the block demand forecasting method provided by the invention, it can be understood that the station block calculation demand gap is used for paying attention to the blocks with the gap value larger than 0 and guiding the moving in of the vehicle, and the non-station block calculation surplus is used for paying attention to the blocks with the surplus value larger than 0 and guiding the moving out of the vehicle.
6. The embodiment of the present invention further provides a block demand prediction system, which has the same beneficial effects as the block demand prediction method described above, and is not described herein again.
7. The embodiment of the present invention further provides a computer storage medium, which has the same beneficial effects as the block demand prediction method described above, and is not described herein again.
The block demand prediction method, system and computer storage medium disclosed in the embodiments of the present invention are described in detail above, and the principles and embodiments of the present invention are explained herein by applying specific embodiments, and the descriptions of the above embodiments are only used to help understanding the method and its core ideas of the present invention; meanwhile, for the persons skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present description should not be construed as a limitation to the present invention, and any modification, equivalent replacement, and improvement made within the principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. A block demand prediction method for guiding vehicle dispatching is characterized in that: the method comprises the following steps:
acquiring spatial relationship characteristics, real-time characteristics and historical order characteristics of a current block and respectively preprocessing the characteristics;
generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics and the preprocessed real-time characteristics, historical order characteristics and spatial relation characteristics;
the method for acquiring and preprocessing the spatial relationship features comprises the following steps:
acquiring the block correlation between the current block and the rest blocks;
and screening the rest blocks of which the correlation with the blocks is in a preset range, extracting the characteristics of the screened rest blocks as spatial relationship characteristics, and generating corresponding vector representations.
2. The block demand prediction method of claim 1, wherein: the spatial dimension characteristics comprise one or more of block serial numbers, POI data, heat value, road network data, AOI data, resident population density, number of nearby stations and people flow grade; the time dimension characteristics comprise one or more of time period serial numbers, holiday information and weather information.
3. The block demand prediction method of claim 1, wherein: and adopting a Dense Layer or an FM Layer to carry out interaction on the space characteristic vector and the time characteristic vector to form space-time interaction characteristics.
4. The block demand prediction method of claim 1, wherein: the method for acquiring the real-time characteristics and the historical order characteristics and correspondingly preprocessing the real-time characteristics and the historical order characteristics respectively comprises the following steps:
acquiring historical order quantity of a block, selecting the order quantity in the current preset time as a real-time feature, extracting the real-time feature by adopting a time series model or LSTM to obtain a real-time feature vector, or directly expressing by adopting at least part of the real-time feature as the real-time feature vector;
and obtaining the total order amount and the average order value in a preset period based on the historical order amount as historical order characteristics, and directly representing the historical order characteristics as vectors.
5. The block demand prediction method of claim 1, wherein: the type of the block relevance comprises one or more of geographical relevance, vehicle circulation relevance and POI relevance among blocks.
6. The block demand prediction method of claim 1, wherein: the method also comprises the following steps after the block vehicle inflow amount, the vehicle outflow amount and the block demand amount prediction result are generated:
acquiring the current number of vehicles in the block, and judging whether the block contains a station or not;
if the station exists, calculating the number of required gaps of the block, wherein the number of required gaps of the block is equal to the number obtained by subtracting the inflow of vehicles from the required amount of the block and then subtracting the current number of vehicles in the block;
and if no station exists, calculating the surplus quantity of the block, wherein the surplus quantity of the block is equal to the sum of the current vehicle number of the block and the vehicle inflow amount, and then subtracting the vehicle outflow amount.
7. The block demand prediction method of claim 6, wherein: and carrying out vehicle moving in on the blocks with the block demand gap number larger than 0, and carrying out vehicle moving out on the blocks with the block surplus number larger than 0.
8. A block demand prediction system for implementing the method of any one of claims 1 to 7, wherein: the system comprises the following modules:
an acquisition module: acquiring spatial relation characteristics, real-time characteristics and historical order characteristics of a current block and respectively and correspondingly preprocessing the spatial relation characteristics, the real-time characteristics and the historical order characteristics;
an interaction module: generating corresponding space feature vectors and time feature vectors based on the space dimension features and the time dimension features of the current block, and interacting the space feature vectors and the time feature vectors to form space-time interaction features;
a prediction module: and generating block vehicle inflow, vehicle outflow and block demand prediction results through a neural network by combining the space-time interaction characteristics, the preprocessed real-time characteristics, the preprocessed historical order characteristics and the spatial relation characteristics.
9. A computer storage medium having a computer program stored thereon, characterized in that: the block demand prediction method according to any one of claims 1 to 7, when the computer program is run on a computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210722389.5A CN114819414B (en) | 2022-06-24 | 2022-06-24 | Block demand prediction method, system and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210722389.5A CN114819414B (en) | 2022-06-24 | 2022-06-24 | Block demand prediction method, system and computer storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114819414A CN114819414A (en) | 2022-07-29 |
| CN114819414B true CN114819414B (en) | 2022-10-28 |
Family
ID=82521673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210722389.5A Active CN114819414B (en) | 2022-06-24 | 2022-06-24 | Block demand prediction method, system and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114819414B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117852730A (en) * | 2024-03-08 | 2024-04-09 | 北京阿帕科蓝科技有限公司 | Vehicle circulation prediction method, device, computer equipment and storage medium |
| CN117876027A (en) * | 2024-03-13 | 2024-04-12 | 北京阿帕科蓝科技有限公司 | Data processing method, device, computer equipment and storage medium |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102110365B (en) * | 2009-12-28 | 2013-11-06 | 日电（中国）有限公司 | Road condition prediction method and road condition prediction system based on space-time relationship |
| CN108985475B (en) * | 2018-06-13 | 2021-07-23 | 厦门大学 | Network taxi appointment and taxi calling demand prediction method based on deep neural network |
| CN111275228B (en) * | 2018-12-04 | 2022-08-02 | 北京骑胜科技有限公司 | Prediction method, prediction device, electronic device and storage medium |
| CN111476588B (en) * | 2019-01-24 | 2023-10-24 | 北京嘀嘀无限科技发展有限公司 | Order demand prediction method and device, electronic equipment and readable storage medium |
| CN109978040A (en) * | 2019-03-19 | 2019-07-05 | 西南交通大学 | A kind of traffic transport power distribution forecasting method |
| CN112561187B (en) * | 2020-12-22 | 2022-06-03 | 武汉理工大学 | Network taxi booking target order prediction method based on CNN-LSTM |
-
2022
- 2022-06-24 CN CN202210722389.5A patent/CN114819414B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN114819414A (en) | 2022-07-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114819414B (en) | Block demand prediction method, system and computer storage medium | |
| Miao et al. | Taxi dispatch with real-time sensing data in metropolitan areas: A receding horizon control approach | |
| Gurumurthy et al. | Analyzing the dynamic ride-sharing potential for shared autonomous vehicle fleets using cellphone data from Orlando, Florida | |
| CN109035770B (en) | Real-time analysis and prediction method for bus passenger capacity in big data environment | |
| Monteiro et al. | On-street parking prediction using real-time data | |
| Haghani et al. | Real-time emergency response fleet deployment: concepts, systems, simulation & case studies | |
| CN109872535A (en) | A kind of current prediction technique of wisdom traffic, device and server | |
| CN109612488B (en) | Big data micro-service-based mixed travel mode path planning system and method | |
| Qi et al. | How long a passenger waits for a vacant taxi--large-scale taxi trace mining for smart cities | |
| CN110852499A (en) | Method and device for recovering fault vehicle | |
| Sayarshad et al. | Optimizing dynamic switching between fixed and flexible transit services with an idle-vehicle relocation strategy and reductions in emissions | |
| CN110210652A (en) | A kind of rail traffic website passenger flow control method and device | |
| CN113326641B (en) | Path planning method and device, computer equipment and storage medium | |
| Chen et al. | ASC: Actuation system for city-wide crowdsensing with ride-sharing vehicular platform | |
| Ebrahimi et al. | Using a GIS-based spatial approach to determine the optimal locations of bikeshare stations: The case of Washington DC | |
| Sayarshad et al. | A scalable non-myopic atomic game for a smart parking mechanism | |
| Wang et al. | Large-scale intelligent taxicab scheduling: A distributed and future-aware approach | |
| Jaramillo-Alvarez et al. | Route optimization of urban public transportation | |
| He et al. | Station importance evaluation in dynamic bike-sharing rebalancing optimization using an entropy-based TOPSIS approach | |
| Liu et al. | Exploring the impact of spatiotemporal granularity on the demand prediction of dynamic ride-hailing | |
| Jiang et al. | A unified probabilistic framework for spatiotemporal passenger crowdedness inference within urban rail transit network | |
| Cui et al. | Usage demand forecast and quantity recommendation for urban shared bicycles | |
| CN117610734A (en) | Deep learning-based user behavior prediction method, system and electronic equipment | |
| Faghih et al. | Predicting short-term Uber demand using spatio-temporal modeling: A New York City Case Study | |
| Ku et al. | Real-time taxi demand prediction using recurrent neural network |
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 |