CN104599168A - Method and device for allocating taxi-calling orders - Google Patents
Method and device for allocating taxi-calling orders Download PDFInfo
- Publication number
- CN104599168A CN104599168A CN201510053500.6A CN201510053500A CN104599168A CN 104599168 A CN104599168 A CN 104599168A CN 201510053500 A CN201510053500 A CN 201510053500A CN 104599168 A CN104599168 A CN 104599168A
- Authority
- CN
- China
- Prior art keywords
- order
- orders
- competition
- group
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and device for allocating taxi-calling orders. The method includes: defining an order group including multiple orders to be allocated and a user group including multiple users whose orders to be received. Based on that the allocation way of the orders is decided by the order group and the user group and the decided order allocation way, orders of the order group are sent to the users of the user group. The method and the device for allocating the taxi orders can achieve the optimal order allocation way in terms of a taxi-calling software platform.
Description
Technical field
Embodiments of the invention relate generally at the method and apparatus used in software platform of calling a taxi, and relate more particularly to distribution method and the device of chauffeur order.
Background technology
Along with popularizing of software of calling a taxi, the custom of calling a taxi of people is by deep change.For certain software of calling a taxi, passenger opens software and sends chauffeur demand, after the server of this software receives chauffeur request, and the multiple drivers around this passenger of Auto-matching, and utilize large data to calculate some drivers of mating the most with this order, this order is pushed to relevant driver.Driver's end receives order and is represented, and determines that initiating competition for orders behavior decision accepts or do not accept by driver.
Calling a taxi software as the platform connecting passenger and taxi driver, how to mate driver and order, namely how to select " which driver which order is broadcast to, and which order which driver listens ", is that the core of platform is experienced.
First, for passenger, certainly wish that its order has higher probability to be replied by driver; For driver, it is desirable to oneself to be ready that the order accepted has greater probability to get; For software platform of calling a taxi, wish that the order of each passenger can strike a bargain, most of driver can get order simultaneously.
The software of calling a taxi of the transaction platform of current most contact driver and order, mostly adopts the following two kinds mode to distribute in Order splitting.
One, centered by order, is distributed by the mode finding suitable driver for order.Under this allocation scheme, the triggering mode of order distribution system is that single order triggers, and a certain passenger initiates the request of calling a taxi, and after order distribution system receives order request, for suitable driver found by this order in single driver waiting, pushes; Whether driver receives this order and pushes, and determines whether order according to the wish of oneself, and can successful order, also will see that other drivers rob one-state to this order.
Its two, centered by driver, distributed by the mode finding suitable order for driver.This allocation scheme comparatively above allocation scheme system realize on relatively simple, directly trigger with single driver, such as, a certain driver is after previous order has pushed, client is just continuous to be initiated to push order request to service end, after service end receives the request of driver's client, this driver is transmitted to order distribution system, after order distribution system takes this driver, the order that do not strike a bargain is searched in order database, calculate the correlativity of this driver and each order, that selects N number of (such as N=2) to mate most is pushed to driver, determine whether respond by driver.If driver does not respond and after order pushed, driver's client is initiated to push order request to server again, continues next round order and pushes.
But, be no matter order distribution system or the many-to-one order distribution system of order-driver of order-driver's one-to-many, all cannot guarantee accurately to find optimum matching way, bring loss therefore to the whole structure of platform.
As can be seen from the above analysis, be no matter order distribution system or the many-to-one order distribution system of order-driver of order-driver's one-to-many, all cannot guarantee accurately to find optimum matching way, bring certain loss therefore to the whole structure of platform.And the basic reason producing this problem is: one, for the allocation scheme of one-to-many, there is a side (order or driver) all the time centered by individuality, other individualities cannot be taken into account; Its two, this distribution mechanism does not meet the compatible incentives principle in Mechanism Design, namely individual target and integral platform target not quite identical.
Summary of the invention
In view of the above-mentioned problems in the prior art, one of object of embodiments of the invention is: provide a kind of in the assignment procedure using order group and driver group as the order of distribution basis-the order distribution system model of driver's multi-to-multi, thus solve above-mentioned technical matters existing in prior art.
According to a first aspect of the invention, provide a kind of distribution method of chauffeur order, comprising: determine to comprise the order group of multiple order to be allocated and comprise the customer group of multiple single user waiting; Order splitting mode is determined based on described order group and described customer group; And according to determined Order splitting mode, push the order in described order group to the user in described customer group.
According to some embodiments of the present invention, wherein determine that the order group comprising multiple order to be allocated comprises with the customer group comprising multiple single user waiting: determine described order group and described customer group based on a geographic area.
According to some embodiments of the present invention, wherein determine that described order group and described customer group comprise based on a geographic area: the current all orders in described geographic area are defined as described order group, and the current all single user waiting in described geographic area are defined as described customer group.
According to some embodiments of the present invention, wherein said geographic area comprises city.
According to some embodiments of the present invention, wherein determine that Order splitting mode comprises based on described order group and described customer group: based on described order group and described customer group, and according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine described Order splitting mode.
According to some embodiments of the present invention, wherein according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine described Order splitting mode comprises: calculate described order probability of transaction, competition for orders success ratio and listen the weighted sum of single competition for orders rate; And determine the Order splitting mode that makes described weighted sum maximum, as determined Order splitting mode.
According to some embodiments of the present invention, wherein based on the competition for orders probability of the arbitrary user in described customer group to the arbitrary order in described order group, calculate respectively described order probability of transaction, described competition for orders success ratio and described in listen single competition for orders rate.
According to some embodiments of the present invention, wherein calculate described competition for orders probability based on the status flag relevant with the passenger placed an order with single user waiting.
According to some embodiments of the present invention, wherein said status flag comprises at least one item in the following: single user waiting is raised the price with the distance between passenger, order expected revenue, passenger, whether the direction, destination of passenger is consistent with the expection travel direction of single user waiting.
According to some embodiments of the present invention, wherein use at least one in following algorithm to determine the Order splitting mode making described weighted sum maximum: the method for exhaustion, genetic algorithm, ant group algorithm, tabu search algorithm, simulated annealing and the hill-climbing algorithm based on greed.
According to some embodiments of the present invention, wherein determine that the Order splitting mode making described weighted sum maximum comprises: generate initial Order splitting mode based on pre-defined rule; And use and based on the hill-climbing algorithm of greed, described initial Order splitting mode is optimized, thus determine the Order splitting mode that makes described weighted sum maximum.
According to some embodiments of the present invention, wherein use order buffer area and user cache district to store described order group and described customer group respectively, and read described order buffer area and described user cache district termly to determine Order splitting mode.
According to some embodiments of the present invention, but wherein when order is created or pushed over time, become not by competition for orders, this order is added in described order buffer area; And after order is pushed to user, delete this order from described order buffer area.
According to some embodiments of the present invention, wherein when user is in until order state, this user is added in described user cache district; And after user's competition for orders success, delete this user from described user cache district.
According to a second aspect of the invention, provide a kind of distributor of chauffeur order, comprising: the first determining unit, be configured to the order group determining to comprise multiple order to be allocated and the customer group comprising multiple single user waiting; Second determining unit, is configured to determine Order splitting mode based on described order group and described customer group; And push unit, be configured to, according to determined Order splitting mode, push the order in described order group to the user in described customer group.
By adopting distribution method and the device of the chauffeur order of embodiments of the invention, achieve using order group and driver group as the order of distribution basis-order distribution system the model of driver's multi-to-multi, and determine overall expectation target and individual expectation target, set up the contact between global object and individual goal, thus the order matching way of the optimum for the whole structure of platform can be obtained.
Accompanying drawing explanation
By reference to accompanying drawing reading detailed description hereafter, above-mentioned and other objects of embodiments of the invention, feature and advantage will be easy to understand.In the accompanying drawings, show some embodiments of the present invention by way of example, and not by way of limitation, wherein:
Fig. 1 schematically shows the distribution method of a kind of chauffeur order according to an embodiment of the invention;
Fig. 2 schematically shows a kind of embodiment in engineering of the distribution method of chauffeur order; And
Fig. 3 schematically shows the distributor of a kind of chauffeur order according to an embodiment of the invention.
Embodiment
Principle and the spirit of embodiments of the invention are described below with reference to the accompanying drawings.Should be appreciated that describing these embodiments is only understand better to enable those skilled in the art to and implement the present invention, and not limit the scope of the invention by any way.
The distribution method 100 of a kind of chauffeur order according to an embodiment of the invention is schematically shown with reference to figure 1, Fig. 1.Method 100 enters step 101 after start-up.In a step 101, determine comprise the order group of multiple order to be allocated and comprise the customer group of multiple single user waiting.
Those skilled in the art will appreciate that but order to be allocated can be comprised passenger and created by software platform of calling a taxi not yet be pushed to the order of calling a taxi of driver.Those skilled in the art be further appreciated that order to be allocated also can comprise passenger by software platform of calling a taxi create and to driver push but over time, become not by the order of calling a taxi of competition for orders.Those skilled in the art can understand further, as long as to need the order pushed to driver can belong to order to be allocated here sometime, the present invention is unrestricted in this.
Those skilled in the art will appreciate that single user waiting can comprise the driver user of taxi taking platform.Those skilled in the art is further appreciated that single user waiting can also comprise the various operating personnel of other vehicles various if software of calling a taxi according to an embodiment of the invention uses in the transport scene of other type.As long as it will be further appreciated by those skilled in the art that and can belong to single user waiting here the user of the software platform of calling a taxi being in state of can accepting an order sometime, the present invention is unrestricted in this.
It should be noted that in a step 101, multiple order to be allocated also comprises two orders to be allocated, and multiple single user waiting also comprises two single user waiting.In other words, although the distribution method 100 of chauffeur order is generally used in the scene of a large amount of order to be allocated and a large amount of single user waiting according to an embodiment of the invention, method 100 also can be applied in the scene only having two orders and two drivers.
According to some embodiments of the present invention, in a step 101, order group and customer group can be determined based on a geographic area.Specifically, the current all orders in certain geographic area can be defined as order group, and the current all single user waiting in this geographic area are defined as customer group.Such as, this geographic area can comprise the geographic area that province, city, counties and cities etc. are divided by administrative division, also can be the geographic area divided by geographic longitude and latitude.Those skilled in the art will appreciate that the present invention not by the restriction of the dividing mode of concrete geographic area.
Then, method 100 advances to step 102.In a step 102, Order splitting mode is determined based on order group and customer group.
According to some embodiments of the present invention, may further include in a step 102: based on order group and customer group, and according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine Order splitting mode.Order probability of transaction E
orderSuccRaterefer to the ratio of the order eventually through software platform conclusion of the business of calling a taxi in order group and all orders in order group.Competition for orders success ratio E
strivesSuccRaterefer to the ratio robbing the quantity of single operation that success that user in customer group implements is robbed the quantity of single operation and user and implemented.Listen single competition for orders rate E
striveRaterefer to that user in customer group implements to rob the ratio being pushed to his quantity on order in the quantity of single operation and order group to the order being pushed to him.In this way, method 100 can realize software platform optimum allocation mode generally speaking of calling a taxi.But, those skilled in the art will appreciate that according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine Order splitting mode is only determine a kind of embodiment of Order splitting mode based on order group and customer group.Those skilled in the art can when concrete different technological accumulation and inheritance, application scenarios and designing requirement, adopt other index or other mode to determine Order splitting mode based on order group and customer group.
According to some embodiments of the present invention, according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine that Order splitting mode can comprise: calculate order probability of transaction, competition for orders success ratio and listen the weighted sum of single competition for orders rate; And determine the Order splitting mode that makes this weighted sum maximum, as determined Order splitting mode.Specifically, in the implementation process of method 100, those skilled in the art can arrange weight between these three indexs according to the technology scene of reality or environment, and this can be formulated as E=α
1* E
orderSuccRate+ α
2* E
striveSuccRate+ α
3* E
striveRate, E represents the weighted sum of these three indexs, α
1, α
2, α
3represent the weighted value of these three indexs respectively.According to one embodiment of present invention, method 100 selects the Order splitting mode making this weighted sum E maximum as the Order splitting mode finally determined.It should be noted that, those skilled in the art can adopt other mode finally to determine Order splitting mode to consider these three indexs, the present invention is not limited to linear weighted function mode as described above, and which is only a kind of example of embodiments of the invention.
According to some embodiments of the present invention, based on the competition for orders probability of the arbitrary user in customer group to the arbitrary order in order group, order probability of transaction, competition for orders success ratio can be calculated respectively and listen single competition for orders rate.In a kind of example of one embodiment of the present of invention, suppose to there is N in order group
o, in customer group, there is N in individual order
dindividual user, then the competition for orders probability of the user j in customer group to the order i in order group can be expressed as P
ij.So order probability of transaction can be expressed as further
Competition for orders success ratio can be expressed as
Single competition for orders rate is listened to be expressed as
Wherein
it will be appreciated by those skilled in the art that, can also derive order probability of transaction, competition for orders success ratio according to other mode from competition for orders probability and listen single competition for orders rate, embodiments of the invention be not limited to derive order probability of transaction, competition for orders success ratio from competition for orders probability and listen the concrete mode of single competition for orders rate.
According to some embodiments of the present invention, competition for orders probability P can be calculated based on the status flag relevant with the passenger placed an order with single user waiting
ij.Such status flag includes but not limited to: single user waiting is raised the price with the distance between passenger, order expected revenue, passenger, whether the direction, destination of passenger is consistent with the expection travel direction of user and will have influence on single user waiting other factors for the competition for orders wish of order, etc.In this case, competition for orders probability P
ijcan be expressed as
wherein X
ijrepresent the proper vector that status flag forms, W represents X
ijin weight corresponding to each status flag, it can be arranged according to concrete technology scene and demand.
Then, method 100 advances to step 103.In step 103, according to determined Order splitting mode, the order in order group is pushed to the user in customer group.
According to some embodiments of the present invention, determined Order splitting mode can be push an order to a user, also can be push multiple order to a user; An order can only be pushed to a user a time, also can be pushed to multiple user a time; An order can be at least pushed to each user, also order can be do not pushed to certain user when not being applicable to order.Those skilled in the art will appreciate that number and the rule of concrete propelling movement order can be selected according to concrete technology scene and demand, the present invention is unrestricted in this.
In addition, those skilled in the art will appreciate that the present invention also not by the restriction of concrete propelling movement mode.Concrete propelling movement mode can be by voice broadcast, text importing and any propelling movement mode making single user waiting can learn order to be allocated, etc.
Therefore, as mentioned above, according to one embodiment of present invention, the distribution method 100 of the order of embodiments of the invention can be analyzed by setting up following model to carry out and solve.First, business is supposed a driver user can only accept an order simultaneously, same order can be pushed to multiple driver.When multiple driver initiates competition for orders request to same order, only has driver's energy competition for orders success, all the other drivers then competition for orders failure.Under this business scenario, to Order Allocation Model can mathematically modeling be as follows:
In superincumbent model, E
orderSuccRate(probability of transaction), E
striveSuccRate(competition for orders success ratio), E
striveRate(listening single competition for orders rate) is core business index, and E is the weighted sum of above some core index, and the optimization aim of model is that weighted target is maximum.Solution to model is matrix { d
ij, i ∈ [0, N
o], j ∈ [0, N
d].In other words, be exactly to N
oindividual order to be allocated, N
dindividual single user waiting, according to the competition for orders rate P of constraint above and correspondence
ij, obtain the allocation scheme D={d making integration objective E maximum
ij.
According to some embodiments of the present invention, at least one in following algorithm can be used determine the Order splitting mode making above-mentioned weighted sum E maximum: the method for exhaustion, genetic algorithm, ant group algorithm, tabu search algorithm, simulated annealing and the hill-climbing algorithm based on greed.It will be apparent to one skilled in the art that the intelligent algorithm that can also adopt other and non intelligent algorithm solve above-mentioned model, embodiments of the invention are not restricted to concrete derivation algorithm.
According to some embodiments of the present invention, due in the software platform of reality, above-mentioned model needs to carry out calculating online and solves, and combination property and effect can generate initial Order splitting mode based on pre-defined rule; And use and based on the hill-climbing algorithm of greed, initial Order splitting mode is optimized, thus determine the Order splitting mode that makes weighted sum maximum.
Fig. 2 schematically shows a kind of embodiment 200 in engineering of the distribution method 100 of chauffeur order.In Project Realization, in order to ensure the model of the multi-to-multi realized between order to be allocated and single user waiting, have employed order buffer area and user cache district.Two buffer areas are responsible for storing the work (frame 201) such as deletion of current order to be allocated and single user waiting, the entering of new data, stale data respectively.Specifically, but when order is created or pushed over time, become not by competition for orders, this order is added in order buffer area; And after order is pushed to user, delete this order from order buffer area.In addition, when user is in until order state, this user is added in user cache district; And after user's competition for orders success, delete this user (frame 201) from user cache district.Order splitting service at interval of certain hour (such as 500 milliseconds) from two buffer areas fetch current user to be allocated (frame 202), order to be allocated set, to current all single user waiting and order to be allocated to calculate the competition for orders probability (frame 203) of arbitrary user to arbitrary order, then according to method above, model is solved, obtain current most suitable Order splitting mode matrix D (frame 204).
Fig. 3 schematically shows the distributor 300 of a kind of chauffeur order according to an embodiment of the invention.Device 300 can comprise the first determining unit 301, second determining unit 302 and push unit 303.Device 300 can comprise order buffer area 304 and user cache district 305 further alternatively.
According to some embodiments of the present invention, the first determining unit 301 can be configured to the order group determining to comprise multiple order to be allocated and the customer group comprising multiple single user waiting.Determining unit 302 can be configured to determine Order splitting mode based on order group and customer group.Push unit 303 can be configured to, according to determined Order splitting mode, push the order in order group to the user in customer group.
According to some embodiments of the present invention, the first determining unit 301 can be configured to further: determine described order group and described customer group based on a geographic area.According to some embodiments of the present invention, the first determining unit 301 can be configured to further: the current all orders in geographic area are defined as order group, and the current all single user waiting in geographic area are defined as customer group.According to some embodiments of the present invention, geographic area can comprise city.
According to some embodiments of the present invention, the second determining unit 302 can be configured to further: based on order group and customer group, and according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine Order splitting mode.Second determining unit 302 can be configured to further: calculate order probability of transaction, competition for orders success ratio and listen the weighted sum of single competition for orders rate; And determine the Order splitting mode that makes this weighted sum maximum, as determined Order splitting mode.Second determining unit 302 can be configured to further: based on the competition for orders probability of the arbitrary user in customer group to the arbitrary order in order group, calculates order probability of transaction, competition for orders success ratio respectively and listens single competition for orders rate.
According to some embodiments of the present invention, the second determining unit 302 can be configured to further: calculate competition for orders probability based on the status flag relevant with the passenger placed an order with single user waiting.Status flag comprises at least one item in the following: whether single user waiting raise the price with the distance between passenger, order expected revenue, passenger, the treat direction, destination of order passenger is consistent with the expection travel direction of user.
According to some embodiments of the present invention, second determining unit 302 can be configured to further, uses at least one in following algorithm to determine the Order splitting mode making weighted sum maximum: the method for exhaustion, genetic algorithm, ant group algorithm, tabu search algorithm, simulated annealing and the hill-climbing algorithm based on greed.According to some embodiments of the present invention, the second determining unit 302 can be configured to further: generate initial Order splitting mode based on pre-defined rule; And use and based on the hill-climbing algorithm of greed, initial Order splitting mode is optimized, thus determine the Order splitting mode that makes weighted sum maximum.
According to some embodiments of the present invention, device 300 may further include order buffer area 304 and user cache district 305 stores order group and customer group respectively, and reading order buffer area 304 and user cache district 305 determine Order splitting mode termly.In figure 3 order buffer area 304 and user cache district 305 represented by dashed line.Specifically, but when order is created or pushed over time, become not by competition for orders, this order is added in order buffer area 304; And after order is pushed to user, delete this order from order buffer area 304.When user is in until order state, this user is added in user cache district 305; And after user's competition for orders success, delete this user from user cache district 305.
Below the distribution method of chauffeur order according to an embodiment of the invention and device is discussed relative to the advantage of existing scheme and technique effect from concrete exemplary aspect respectively.Before how discussion carries out Order splitting, first carry out following hypothesis, and consider two concrete examples next under these assumptions.
Suppose 1: driver determines whether initiate competition for orders request, and the distance positive correlation between driver and order, driver's distance order is nearer, and driver is more ready competition for orders, and apart from far away, driver is more unwilling competition for orders.Suppose 2: if driver's distance order is very far away, then this driver participates in competition for orders hardly, will sail very long distance by sky, and cause income not high because welcome the emperor.Suppose 3: after passenger places an order, have certain patience, if the long period does not also have driver to reply, passenger can cancel an order.
Example 1: order 1 has 300 meters apart from driver 1, distance driver 2 has 400 meters; Order 2 has 2000 meters apart from driver 1, and concrete driver 2 has 500 meters.
In this illustration, be not difficult to find (driver 1, order 1), (driver 2, order 2) are best couplings.To example above, adopt (x, y) to represent and order x is distributed to driver y, order distribution system has 4 kinds of allocation scheme.
The first, (1,1), (2,2), this allocation scheme is obviously best, if driver 1,2 all initiates competition for orders, due to not competition, two successes per capita; For passenger, two passengers all have driver to reply.
The second, (1,1), (1,2), in this fashion, probably all there is competition for orders behavior, at this moment must have driver's competition for orders failure in two drivers are the pushed order nearest apart from oneself.
If driver 2 competition for orders failure, now, driver 1 achieves one's goal and completes order 1, and after system knows driver 1 failure, redistributes order again to driver 2, now only have 1 order residue, select to push order 2, after driver 2 receives order 2, and competition for orders success.Although result is good, obviously efficiency is a bit on the low side; And in another situation, because order 2 is spaced a period of time just to broadcast, if passenger eliminates order because the stand-by period is longer just at this moment, then cannot obtain the optimum solution as mode 1.
If driver 1 failure, now driver 2 completes order 1.In like manner, system pushes order 2 to driver 1.If driver 1 determines competition for orders, combination so is now (1,2), (2,1), compares (1,1), (2,2), and two drivers all will pay and larger welcome the emperor distance; And another kind of situation, if driver 1 feels too far away and be reluctant competition for orders, then order 1 does not strike a bargain, and driver 1 there is no suitable order simultaneously.
Similar, to other two kinds of situations (1,2), (2,1) and (1,2), (2,2), all can not obtain Optimum Matching efficiently, maybe cannot obtain Optimum Matching.
Consider another example 2 again: order 1 has 300 meters apart from driver 1, distance driver 2 has 400 meters; Order 2 has 800 meters apart from driver 1, and distance driver 2 has 9000 meters.
Under this example, because order 2 is all very far away apart from two drivers, therefore the competition for orders wish of two drivers to this order is almost nil, and two drivers also not necessarily to order 1, competition for orders can occur, and just the competition for orders wish of two drivers to order 1 is larger.At this moment optimum solution is (1,1), and suboptimal solution is (1,2).
In Order splitting angle, still there are 4 kinds of distribution condition above.First, consider that order 2 is all far apart from two drivers, its hope struck a bargain is little; Meanwhile, although order 1 is all closer apart from each driver, there is certain following possibility after all, namely certain driver wherein or two drivers all do not respond.Therefore, for this situation, allocation strategy should strike a bargain by guarantee order 1 large as far as possible, and therefore reasonable allocation scheme is that order 1 is broadcast to driver 1 and driver 2 simultaneously.
The two classes more typical situation of two example representation above, relatively easily draws best allocation scheme.Further, best allocation scheme is judged to this situation is just more difficult below.
Example 3: order 1 has 300 meters apart from driver 1, distance driver 2 has 400 meters; Order 2 has 1100 meters apart from driver 1, and distance driver 2 has 1000 meters.
To this example, order 1 is comparatively near apart from two drivers, and order 2 is in one and says near, says far not far distance, therefore this example is an example being between example 1 and example 2, so (1,1), (1,2) and (1,1), (2,3) which good is just difficult to judge actually.
Be two orders and two drivers above, if N number of order M driver's situation can be more complicated; Simultaneously, under superincumbent hypothesis, only considered apart from the impact of this factor on driver's competition for orders wish, in fact several factors is also had to affect the factor of driver's competition for orders wish, the complexity of the final valuation of such as order, the order of order destination, the jam situation of road, whether consistent with driver's projected trip direction etc.
On the basis of above-mentioned example, the solution of prior art is analyzed below.
Call a taxi in transaction platform at present at great majority contact order user (driver) and order (passenger), in Order splitting, mostly adopt the following two kinds mode to distribute.
The first, centered by order, distributed by the mode finding suitable driver for order.Under this allocation scheme, the triggering mode of order distribution system is that single order triggers, and a certain passenger initiates the request of calling a taxi, and after order distribution system receives order request, for suitable driver found by this order in single driver waiting, pushes; Whether driver is subject to this order and pushes, and determines whether order according to the wish of oneself, and can successful order, also will see that other drivers rob one-state to this order.
In the realizing of this system, need two considerations: need a database, the current driver needing order can be obtained, be called for short driver's buffer area.When a certain driver needs to push order, driver enters this buffer area; When a certain order needs to distribute, distribution system selects some drivers from driver's buffer area, and after this order and each driver are calculated correlativity, finally selected some drivers push; A certain driver removes, needs to enter during order until its next time again after obtaining and pushing order from buffer area.In addition, need a control system, at any time for the order that do not strike a bargain is reallocated, be called for short order dispatch system.
When order just creates, distribution system distributes for this order carries out first time, driver's channel is excessively taken for ensureing that this order is unlikely, other orders are pushed and has an impact, generally, system can do a restriction, namely primary distribution mostly be most this order select N number of (such as N=10) driver push, if these 10 drivers all do not respond this order, so when reallocating to this order, is the task of order dispatch system.
Generally, order dispatch system is (as 5 seconds) run-down database at set intervals, to the order that do not strike a bargain, send to order distribution system successively, after system takes order, repeat work above, continue to find suitable driver to this order and push.
Based on the order distribution system of logical design above, it is characterized by centered by order, each process order, finds suitable coupling, distributes in numerous driver.Therefore, this system becomes the order distribution system of order-driver's one-to-many.
By the example 1 above this systematic analysis, for the purpose of simple, suppose N=1 here, i.e. every sub-distribution, 1 driver only selected by each order.
Suppose that order 1 is first processed, to order 1, select from driver 1, driver 2, obvious driver 1 is nearer, and therefore order 1 selects driver 1 to push, and driver 1 is removed from driver's buffer area simultaneously; To order 2, current only have driver 2 optional, therefore selects driver 2 to push.This allocation scheme, reaches the optimum allocation mode in ideal.
Suppose that order 2 is first processed.Analysis above similar, in this case, driver 2 will be caused to push order 1, and driver 1 pushes order 2, and the allocation scheme of distance optimum differs greatly.
As can be seen from the above analysis, the order distribution system mode of this order-driver's one-to-many, can obtain optimum solution be a random occurrence completely, because the time loss of each process Order splitting of the computing machine of service end is in millisecond rank, optimum solution can be obtained and depend on that two orders arrive the time sequencing of server completely.
The second, centered by driver, distributed by the mode finding suitable order for driver.This allocation scheme comparatively above allocation scheme relatively simple on system realizes, directly trigger with single driver.Such as, a certain driver is after previous order has pushed, client service of just constantly giving is initiated to push order request, this driver is transmitted to order distribution system, after order distribution system takes this driver after receiving the request of driver's end by service end, the order that do not strike a bargain is searched in order database, calculate the correlativity of this driver and each order, that selects N number of (such as N=2) to mate most is pushed to driver, determines whether respond by driver.If driver does not respond and after order pushed, driver's client is initiated to push order request to server again, continues next round order and pushes.
Based on the order distribution system of logical design above, it is characterized by centered by driver, each process driver, finds suitable coupling, distributes in numerous order.Therefore, this system becomes the many-to-one order distribution system of order-driver.
Can prove theoretically, order distribution system and the many-to-one order distribution system of order-driver of order-driver's one-to-many are of equal value.And the many-to-one order distribution system of order-driver is relatively easy in Project Realization, current industry great majority adopt the many-to-one order distribution system of order-driver.Analyze example above according to this allocation scheme, be not difficult to find, this allocation scheme can not ensure to find optimum allocation scheme.
To sum up, be no matter order distribution system or the many-to-one order distribution system of order-driver of order-driver's one-to-many, all cannot guarantee accurately to find optimum matching way, bring certain loss therefore to the whole structure of platform.
As can be seen from the above analysis, be no matter order distribution system or the many-to-one order distribution system of order-driver of order-driver's one-to-many, all cannot guarantee accurately to find optimum matching way, bring certain loss therefore to the whole structure of platform.And the basic reason producing this problem is: the first, the allocation scheme of one-to-many, has a side (driver or order) all the time centered by individuality, cannot take into account other individualities; The second, distribution mechanism of the prior art does not meet the compatible incentives principle in Mechanism Design, namely individual target and integral platform target not quite identical.
Such as, following example is analyzed, only analyze the many-to-one order distribution system of order-driver, namely analyze centered by single driver, the existing problem discussed above can finding out.
Example 4: order 1 has 500 meters apart from driver 1, distance driver 2 has 400 meters; Order 2 has 2000 meters apart from driver 1, and distance driver 2 has 500 meters.When distributing order to driver 2, In the view of driver 2, system only knows that order 1 is optimum to him, and cannot know and in fact should say 1 point, order and driver 1, and driver 2 should have to take the second best, and selects order 2.
Secondly, the order that be contemplated to be push of platform to each driver responds as far as possible; And the overall goals of platform to be the probability of transaction of all orders large as far as possible, all driver's competition for orders conflicts are little as far as possible, individual goal and global object Incomplete matching.
For how to solve first problem, key is how to be considered together with all drivers by all orders, so just can avoid the information loss between individuality; And to Second Problem, need contacting between individual expectation target and overall expectation target, and then ensure that individual goal is consistent with global object.
Embodiments of the invention are on solve on two problems, adopt following scheme: first, by current all passengers (order) integrally, be designated as order group, by current all single user waiting (driver) integrally, customer group is designated as, in the assignment procedure, using order group and customer group as allocation unit, therefore this model is the order distribution system of order-driver's multi-to-multi.The second, determine overall expectation target and individual expectation target, set up the contact between global object and individual goal.For individual goal, wish the order that each driver's active response pushes it, therefore we suppose that all drivers obey identical conditions probability distribution P (STRIVE=1|X) to the competition for orders wish of order, wherein, STRIVE=1 represents driver's competition for orders, STRIVE=0 represents driver's not competition for orders, and X represents the status flag information of current driver and driver, as distance, anticipated price, pays cost etc.Meanwhile, determine that platform global object is all order probability of transaction E, obvious probability of transaction is the function that all driver's competition for orders are expected, E=F (P).
For competition for orders probability P (STRIVE=1|X), the method for machine learning can be adopted to train and estimate, training sample is that history is broadcast and singly robbed forms data { Y
ij| X
ij, this record represents and once pushes record, and namely certain order is to the once propelling movement of certain driver, wherein X
ijrepresent some proper vectors in this propelling movement moment, as department take advantage of distance, order expected revenue, passenger raises the price, to expect that with driver whether travel direction consistent etc., Y
ijrepresent after pushing, whether driver competition for orders occurs, 1 is competition for orders, and 2 is non-competition for orders.Forecast model can adopt the widely used LR model (close copy also has linear regression, svm, gbdt etc.) of industry member.
Probability is estimated the equation be expressed as below by LR model:
Wherein, X
ijfor some proper vectors in this propelling movement moment, as department take advantage of distance, order expected revenue, passenger raises the price, to expect that with driver whether travel direction consistent etc.; W is X
ijin weight corresponding to each feature.
Adopt the method for off-line, broadcast by history and singly rob unirecord to model training, then on line just can be right to each (order, driver) in real time, carry out competition for orders rate and estimate, be called STR (strive through rate) and estimate.Obtain arbitrary P to (order i, driver j)
ijafter, just can start to carry out modeling to Order splitting mode according in mode as described in Figure 1.
In order to absolutely prove that the multi-to-multi model that embodiments of the invention provide obviously is better than one-to-many model, first with this model, example is above analyzed below, secondly in the advantage analyzing this model theoretically.
Or for example 4 above, first calculate competition for orders probability P
ij, here in order to the cause of simplicity only employs apart from this feature.Machine training under line, the weight of distance feature is 0.001, and therefore the formula of competition for orders probability can be designated as:
according to this formulae discovery, following result can be obtained.
| Order 1 | Order 2 |
| Driver 1 | 0.38 | 0.12 |
| Driver 2 | 0.4 | 0.38 |
Then, solving model obtains the solution of Order splitting, and in this example, for simplicity, only can consider probability of transaction core index, then the probability of transaction that each allocation scheme is corresponding is as follows:
Obviously, model, after traveling through 4 kinds of feasible solutions, will find (order 1, driver 1), (order 2, driver 2) be the solution making probability of transaction maximum, therefore select the solution of this Order splitting.In like manner, according to the method, example 2, example 3 are solved, all can obtain the solution of the most rational Order splitting.
As can be seen here, the distribution method of chauffeur order and device are relative to the method for prior art according to an embodiment of the invention, can obtain the Order splitting scheme of the optimum generally speaking relative to all orders and all drivers.
Discuss the distribution method of chauffeur order according to an embodiment of the invention and the advantage of device and technique effect from the angle of model analysis below.
Conceptually incite somebody to action, Order Allocation Model is a kind of Mechanism Design.Mechanism Design comprises following three parts: { B (strategic space), P (allocation rule), M (payment rule) }.Here, B is the competition for orders wish of driver, and system cannot learn the true competition for orders wish of driver, can only suppose that the competition for orders wish of driver obeys unified conditional probability distribution, and then adopt the method for machine learning to estimate; P comprises two parts: the propelling movement distribution before competition for orders and the distribution after competition for orders, simply distributes according to the mode of first trying to be the first here after competition for orders.M is the payment behavior after driver obtains order, and under current operation scene, driver only need complete passenger and picks.
The core of Order Allocation Model to be exactly B the be competition for orders wish of driver estimate with competition for orders before propelling movement distribute two models.Faced by Order splitting is driver and order colony, and the target of Order splitting is probability of transaction, competition for orders success ratio, listen the weighting of the targets such as single competition for orders rate maximum.
According to Mechanism Design Theory, a good mechanism should be compatible incentives.Compatible incentives is exactly that participant's rationality realizes the maximized strategy of individual interests, consistent with the strategy desired by Mechanism Design person, thus participant is taken action according to the strategy desired by Mechanism Design person voluntarily.
From the angle analysis of compatible incentives, the Combinatorial Optimization Model of multi-to-multi meets compatible incentives principle, and under this Policy model, as long as each driver is according to the competition for orders wish estimated, so the probability of transaction of platform is maximized; And for each driver, for the trend of individual interests, competition for orders response can be carried out according to corresponding wish.
Analyze probability of transaction computing formula again:
For a certain specific indent x, its probability of transaction is expected as follows:
To a certain driver y, consider that order x pushes driver than y, order
Implication is except driver y, all the other drivers are to the contribution of this order probability of transaction, and namely order x is before broadcasting to driver x, according to the propelling movement situation of current order x, the probability of transaction of order x is estimated (predict success rate), so formula above becomes:
To P
xydifferentiate obtains:
Can find out, derived function with
it is negative correlation.In business, if the current probability of transaction contribution of propelling movement situation to this order of order is larger, newly adding an order so again will be less to the increment that this order probability of transaction is contributed.For whole platform, the propelling movement channel of order is limited resources, is existed by this point, can adjust the propelling movement situation of order automatically.If more abundant (PSR is higher) that order pushes, so system can select the order pushing insufficient (PSR is lower) to single driver to be broadcast automatically.
In like manner, right
differentiate:
Can find out equally, derived function is and P
xynegative correlation.In business, if the competition for orders rate of a driver to a certain order is higher, when this order is broadcast after to this driver, this order is broadcast to the probability of transaction contribution of other drivers to this order lower again.By this point, if an order is pushed to the stronger driver of a competition for orders wish, so according to this model, this order can automatically the least possiblely be broadcast to other drivers by system again, thus ensures that this driver can get this order.
It should be noted that embodiments of the invention can be realized by the combination of hardware, software or software and hardware.Hardware components can utilize special logic to realize; Software section can store in memory, and by suitable instruction execution system, such as microprocessor or special designs hardware perform.It will be appreciated by those skilled in the art that above-mentioned equipment and method can use computer executable instructions and/or be included in processor control routine to realize, such as, in the data carrier of programmable memory or such as optics or electrical signal carrier, provide such code.
In addition, although describe the operation of the inventive method in the accompanying drawings with particular order, this is not that requirement or hint must perform these operations according to this particular order, or must perform the result that all shown operation could realize expectation.On the contrary, the step described in process flow diagram can change execution sequence.Additionally or alternatively, some step can be omitted, multiple step is combined as a step and perform, and/or a step is decomposed into multiple step and perform.Shall also be noted that and can to specialize in one apparatus according to the Characteristic and function of two or more devices of the present invention.Otherwise, the Characteristic and function of an above-described device can Further Division for be specialized by multiple device.
Although describe the present invention with reference to some specific embodiments, should be appreciated that and the invention is not restricted to disclosed specific embodiment.The present invention is intended to contain the interior included various amendment of spirit and scope and the equivalent arrangements of claims.
Claims (28)
1. a distribution method for chauffeur order, comprising:
Determine comprise the order group of multiple order to be allocated and comprise the customer group of multiple single user waiting;
Order splitting mode is determined based on described order group and described customer group; And
According to determined Order splitting mode, push the order in described order group to the user in described customer group.
2. method according to claim 1, wherein determine that the order group comprising multiple order to be allocated comprises with the customer group comprising multiple single user waiting:
Described order group and described customer group is determined based on a geographic area.
3. method according to claim 2, wherein determine that described order group and described customer group comprise based on a geographic area:
Current all orders in described geographic area are defined as described order group, and the current all single user waiting in described geographic area are defined as described customer group.
4. method according to claim 3, wherein said geographic area comprises city.
5. method according to claim 1, wherein determine that Order splitting mode comprises based on described order group and described customer group:
Based on described order group and described customer group, and according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine described Order splitting mode.
6. method according to claim 5, wherein according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine described Order splitting mode comprises:
Calculate described order probability of transaction, competition for orders success ratio and listen the weighted sum of single competition for orders rate; And
Determine the Order splitting mode making described weighted sum maximum, as determined Order splitting mode.
7. method according to claim 5, wherein based on the competition for orders probability of the arbitrary user in described customer group to the arbitrary order in described order group, calculate respectively described order probability of transaction, described competition for orders success ratio and described in listen single competition for orders rate.
8. method according to claim 7, wherein calculates described competition for orders probability based on the status flag relevant with the passenger placed an order with single user waiting.
9. method according to claim 8, wherein said status flag comprises at least one item in the following: single user waiting is raised the price with the distance between passenger, order expected revenue, passenger, whether the direction, destination of passenger is consistent with the expection travel direction of single user waiting.
10. method according to claim 6, wherein uses at least one in following algorithm to determine the Order splitting mode making described weighted sum maximum:
The method of exhaustion, genetic algorithm, ant group algorithm, tabu search algorithm, simulated annealing and the hill-climbing algorithm based on greed.
11. methods according to claim 6, wherein determine that the Order splitting mode making described weighted sum maximum comprises:
Initial Order splitting mode is generated based on pre-defined rule; And
Use and based on the hill-climbing algorithm of greed, described initial Order splitting mode is optimized, thus determine the Order splitting mode that makes described weighted sum maximum.
12. methods according to claim 1, wherein use order buffer area and user cache district to store described order group and described customer group respectively, and read described order buffer area and described user cache district termly to determine Order splitting mode.
13. methods according to claim 12, but wherein when order is created or pushed over time, become not by competition for orders, described order is added in described order buffer area; And after order is pushed to user, delete described order from described order buffer area.
14. methods according to claim 12, wherein when user is in until order state, add to described user in described user cache district; And after user's competition for orders success, delete described user from described user cache district.
The distributor of 15. 1 kinds of chauffeur orders, comprising:
First determining unit, is configured to the order group determining to comprise multiple order to be allocated and the customer group comprising multiple single user waiting;
Second determining unit, is configured to determine Order splitting mode based on described order group and described customer group; And
Push unit, is configured to, according to determined Order splitting mode, push the order in described order group to the user in described customer group.
16. devices according to claim 15, wherein said first determining unit is configured to further:
Described order group and described customer group is determined based on a geographic area.
17. devices according to claim 16, wherein said first determining unit is configured to further:
Current all orders in described geographic area are defined as described order group, and the current all single user waiting in described geographic area are defined as described customer group.
18. devices according to claim 17, wherein said geographic area comprises city.
19. devices according to claim 15, wherein said second determining unit is configured to further:
Based on described order group and described customer group, and according to order probability of transaction, competition for orders success ratio and listen at least one item in single competition for orders rate to determine described Order splitting mode.
20. devices according to claim 19, wherein said second determining unit is configured to further:
Calculate described order probability of transaction, competition for orders success ratio and listen the weighted sum of single competition for orders rate; And
Determine the Order splitting mode making described weighted sum maximum, as determined Order splitting mode.
21. devices according to claim 19, wherein said second determining unit is configured to further:
Based on the competition for orders probability of the arbitrary user in described customer group to the arbitrary order in described order group, calculate respectively described order probability of transaction, described competition for orders success ratio and described in listen single competition for orders rate.
22. devices according to claim 21, wherein said second determining unit is configured to further:
Described competition for orders probability is calculated based on the status flag relevant with the passenger placed an order with single user waiting.
23. devices according to claim 22, wherein said status flag comprises at least one item in the following: single user waiting is raised the price with the distance between passenger, order expected revenue, passenger, whether the direction, destination of passenger is consistent with the expection travel direction of single user waiting.
24. devices according to claim 20, wherein said second determining unit is configured to further: use at least one in following algorithm to determine the Order splitting mode making described weighted sum maximum:
The method of exhaustion, genetic algorithm, ant group algorithm, tabu search algorithm, simulated annealing and the hill-climbing algorithm based on greed.
25. devices according to claim 20, wherein said second determining unit is configured to further:
Initial Order splitting mode is generated based on pre-defined rule; And
Use and based on the hill-climbing algorithm of greed, described initial Order splitting mode is optimized, thus determine the Order splitting mode that makes described weighted sum maximum.
26. devices according to claim 15, wherein order buffer area and user cache district store described order group and described customer group respectively, and read described order buffer area and described user cache district termly to determine Order splitting mode.
27. devices according to claim 26, but wherein when order is created or pushed over time, become not by competition for orders, described order is added in described order buffer area; And after order is pushed to user, delete described order from described order buffer area.
28. devices according to claim 26, wherein when user is in until order state, add to described user in described user cache district; And after user's competition for orders success, delete described user from described user cache district.
Priority Applications (9)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510053500.6A CN104599168A (en) | 2015-02-02 | 2015-02-02 | Method and device for allocating taxi-calling orders |
| GB1712642.6A GB2550523A (en) | 2015-02-02 | 2016-01-29 | Methods and systems for order processing |
| PCT/CN2016/072837 WO2016124118A1 (en) | 2015-02-02 | 2016-01-29 | Order processing method and system |
| SG11201706269QA SG11201706269QA (en) | 2015-02-02 | 2016-01-29 | Methods and systems for order processing |
| PH12017501388A PH12017501388A1 (en) | 2015-02-02 | 2016-01-29 | Methods and systems for order processing |
| US15/547,528 US10657581B2 (en) | 2015-02-02 | 2016-01-29 | Methods and systems for order processing |
| MYPI2017001131A MY181464A (en) | 2015-02-02 | 2016-01-29 | Methods and systems for order processing |
| HK18106251.1A HK1246941A1 (en) | 2015-02-02 | 2018-05-15 | Methods and systems for order processing |
| US16/869,447 US11315170B2 (en) | 2015-02-02 | 2020-05-07 | Methods and systems for order processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510053500.6A CN104599168A (en) | 2015-02-02 | 2015-02-02 | Method and device for allocating taxi-calling orders |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104599168A true CN104599168A (en) | 2015-05-06 |
Family
ID=53124920
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510053500.6A Pending CN104599168A (en) | 2015-02-02 | 2015-02-02 | Method and device for allocating taxi-calling orders |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104599168A (en) |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809527A (en) * | 2015-05-11 | 2015-07-29 | 华侨大学 | Automatic selection method for orders of taking taxies by mobilephones |
| CN105139228A (en) * | 2015-08-20 | 2015-12-09 | 北京嘀嘀无限科技发展有限公司 | Order allocation method and device |
| CN105303817A (en) * | 2015-09-16 | 2016-02-03 | 北京嘀嘀无限科技发展有限公司 | Travel manner planning method and device |
| CN105373840A (en) * | 2015-10-14 | 2016-03-02 | 深圳市天行家科技有限公司 | Designated-driving order predicting method and designated-driving transport capacity scheduling method |
| CN105701634A (en) * | 2016-03-15 | 2016-06-22 | 江阴中科今朝科技有限公司 | Logistics monitoring and dispatching system based on mobile terminal |
| CN105719111A (en) * | 2015-05-22 | 2016-06-29 | 北京小度信息科技有限公司 | Method and device for dynamically adjusting transport capacity |
| WO2016124118A1 (en) * | 2015-02-02 | 2016-08-11 | 北京嘀嘀无限科技发展有限公司 | Order processing method and system |
| WO2017000488A1 (en) * | 2015-06-30 | 2017-01-05 | 百度在线网络技术（北京）有限公司 | Order pushing method and apparatus, and storage medium |
| CN107146123A (en) * | 2016-03-01 | 2017-09-08 | 滴滴（中国）科技有限公司 | A kind of order allocation method and system |
| CN107392512A (en) * | 2016-11-25 | 2017-11-24 | 北京小度信息科技有限公司 | Grouped task method and apparatus |
| WO2018076951A1 (en) * | 2016-10-31 | 2018-05-03 | Beijing Didi Infinity Technology And Development Co., Ltd. | Device and method for order distribution |
| WO2018095065A1 (en) * | 2016-11-23 | 2018-05-31 | 北京小度信息科技有限公司 | Method and apparatus for allocating data object, and electronic device |
| CN108182524A (en) * | 2017-12-26 | 2018-06-19 | 北京三快在线科技有限公司 | A kind of order allocation method and device, electronic equipment |
| CN108537619A (en) * | 2018-03-05 | 2018-09-14 | 新智数字科技有限公司 | A kind of method for allocating tasks, device and equipment based on maximum-flow algorithm |
| WO2018218946A1 (en) * | 2017-06-01 | 2018-12-06 | 北京小度信息科技有限公司 | Order processing method and device |
| CN108960747A (en) * | 2018-08-21 | 2018-12-07 | 安吉汽车物流股份有限公司 | Optimization method and device, storage medium, the terminal of Logistic Scheduling |
| WO2019015661A1 (en) * | 2017-07-20 | 2019-01-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for service request allocation |
| CN109359843A (en) * | 2018-09-30 | 2019-02-19 | 山西游骑兵电子商务有限公司 | Data processing method and order equipment |
| CN109377317A (en) * | 2018-10-26 | 2019-02-22 | 天津五八到家科技有限公司 | Data processing method and device |
| CN109636213A (en) * | 2018-12-19 | 2019-04-16 | 拉扎斯网络科技（上海）有限公司 | Order splitting, evaluation method and device, electronic equipment and storage medium |
| CN110009131A (en) * | 2019-02-28 | 2019-07-12 | 河海大学 | A kind of net about vehicle worksheet processing method considering multifactor impact |
| CN110020818A (en) * | 2018-01-09 | 2019-07-16 | 车伯乐（北京）信息科技有限公司 | A kind of order processing method, apparatus, system and server |
| CN110322159A (en) * | 2019-07-10 | 2019-10-11 | 金瓜子科技发展（北京）有限公司 | A kind of data processing method and device |
| CN110447050A (en) * | 2017-12-04 | 2019-11-12 | 北京嘀嘀无限科技发展有限公司 | System and method for distributing order in online on-demand service |
| TWI684107B (en) * | 2018-12-18 | 2020-02-01 | 國立中山大學 | Method for data imputation and classification and system for data imputation and classification |
| CN110798800A (en) * | 2018-08-02 | 2020-02-14 | 北京嘀嘀无限科技发展有限公司 | Information pushing method, device, equipment and computer readable storage medium |
| CN110874777A (en) * | 2018-08-30 | 2020-03-10 | 北京嘀嘀无限科技发展有限公司 | Order processing method and device |
| CN111523968A (en) * | 2015-11-26 | 2020-08-11 | 滴滴（中国）科技有限公司 | Order splicing method and equipment |
| CN111724091A (en) * | 2019-03-21 | 2020-09-29 | 天津五八到家科技有限公司 | Order distribution method, server side equipment and storage medium |
| CN111882109A (en) * | 2020-06-22 | 2020-11-03 | 北京嘀嘀无限科技发展有限公司 | Order allocation method and system |
| CN112396375A (en) * | 2020-11-18 | 2021-02-23 | 江苏满运物流信息有限公司 | Delivery method and device, storage medium and electronic device |
| CN112418616A (en) * | 2020-11-05 | 2021-02-26 | 深圳依时货拉拉科技有限公司 | Order broadcasting method and device, computer equipment and computer-readable storage medium |
| CN112465602A (en) * | 2020-12-11 | 2021-03-09 | 深圳依时货拉拉科技有限公司 | Order pushing method and device, computer equipment and computer readable storage medium |
| CN104867065B (en) * | 2015-06-05 | 2021-07-02 | 北京嘀嘀无限科技发展有限公司 | Method and equipment for processing order |
| CN113159458A (en) * | 2021-05-20 | 2021-07-23 | 南京领行科技股份有限公司 | Distribution method and device of network appointment vehicles, electronic equipment and storage medium |
| CN114331645A (en) * | 2022-03-14 | 2022-04-12 | 广州宸祺出行科技有限公司 | Method and system for improving transport capacity utilization rate of network appointment car |
| CN114723125A (en) * | 2022-04-01 | 2022-07-08 | 华侨大学 | Intercity vehicle order allocation method combining deep learning and multi-task optimization |
| CN114827100A (en) * | 2022-04-26 | 2022-07-29 | 郑州锐目通信设备有限公司 | Taxi calling method and system |
| US11892312B2 (en) | 2015-01-27 | 2024-02-06 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for providing information for an on-demand service |
| CN114723125B (en) * | 2022-04-01 | 2024-06-28 | 华侨大学 | Inter-city vehicle order allocation method combining deep learning and multitask optimization |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008088387A2 (en) * | 2007-01-11 | 2008-07-24 | Dynamic Intelligence Inc. | Strategic planning and revenue management system |
| CN103150698A (en) * | 2012-10-31 | 2013-06-12 | 中华电信股份有限公司 | Automatic taxi dispatching system |
| CN104156868A (en) * | 2014-08-22 | 2014-11-19 | 北京嘀嘀无限科技发展有限公司 | Taxi integrating system for promoting order transaction based on order value judgment |
| CN104183118A (en) * | 2014-08-19 | 2014-12-03 | 北京嘀嘀无限科技发展有限公司 | Order sending system for obtaining passenger optimal picking-up driver based on auction model |
-
2015
- 2015-02-02 CN CN201510053500.6A patent/CN104599168A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008088387A2 (en) * | 2007-01-11 | 2008-07-24 | Dynamic Intelligence Inc. | Strategic planning and revenue management system |
| CN103150698A (en) * | 2012-10-31 | 2013-06-12 | 中华电信股份有限公司 | Automatic taxi dispatching system |
| CN104183118A (en) * | 2014-08-19 | 2014-12-03 | 北京嘀嘀无限科技发展有限公司 | Order sending system for obtaining passenger optimal picking-up driver based on auction model |
| CN104156868A (en) * | 2014-08-22 | 2014-11-19 | 北京嘀嘀无限科技发展有限公司 | Taxi integrating system for promoting order transaction based on order value judgment |
Cited By (60)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11892312B2 (en) | 2015-01-27 | 2024-02-06 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for providing information for an on-demand service |
| US10657581B2 (en) | 2015-02-02 | 2020-05-19 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for order processing |
| GB2550523A (en) * | 2015-02-02 | 2017-11-22 | Beijing Didi Infinity Tech And Dev Co Ltd | Methods and systems for order processing |
| US11315170B2 (en) | 2015-02-02 | 2022-04-26 | Beijing Didi Infinity Technology And Development Co., Ltd. | Methods and systems for order processing |
| WO2016124118A1 (en) * | 2015-02-02 | 2016-08-11 | 北京嘀嘀无限科技发展有限公司 | Order processing method and system |
| CN104809527B (en) * | 2015-05-11 | 2018-05-25 | 华侨大学 | The order automatic selecting method that a kind of mobile phone is called a taxi |
| CN104809527A (en) * | 2015-05-11 | 2015-07-29 | 华侨大学 | Automatic selection method for orders of taking taxies by mobilephones |
| CN105719111A (en) * | 2015-05-22 | 2016-06-29 | 北京小度信息科技有限公司 | Method and device for dynamically adjusting transport capacity |
| CN104867065B (en) * | 2015-06-05 | 2021-07-02 | 北京嘀嘀无限科技发展有限公司 | Method and equipment for processing order |
| WO2017000488A1 (en) * | 2015-06-30 | 2017-01-05 | 百度在线网络技术（北京）有限公司 | Order pushing method and apparatus, and storage medium |
| CN105139228A (en) * | 2015-08-20 | 2015-12-09 | 北京嘀嘀无限科技发展有限公司 | Order allocation method and device |
| CN112508616A (en) * | 2015-08-20 | 2021-03-16 | 北京嘀嘀无限科技发展有限公司 | Order allocation method and device |
| CN105303817A (en) * | 2015-09-16 | 2016-02-03 | 北京嘀嘀无限科技发展有限公司 | Travel manner planning method and device |
| CN105373840A (en) * | 2015-10-14 | 2016-03-02 | 深圳市天行家科技有限公司 | Designated-driving order predicting method and designated-driving transport capacity scheduling method |
| CN105373840B (en) * | 2015-10-14 | 2018-12-11 | 深圳市天行家科技有限公司 | In generation, drives order forecast method and generation drives Transport capacity dispatching method |
| CN111523968A (en) * | 2015-11-26 | 2020-08-11 | 滴滴（中国）科技有限公司 | Order splicing method and equipment |
| CN111523968B (en) * | 2015-11-26 | 2023-11-21 | 滴滴（中国）科技有限公司 | Method and equipment for spelling bill |
| CN107146123A (en) * | 2016-03-01 | 2017-09-08 | 滴滴（中国）科技有限公司 | A kind of order allocation method and system |
| CN105701634A (en) * | 2016-03-15 | 2016-06-22 | 江阴中科今朝科技有限公司 | Logistics monitoring and dispatching system based on mobile terminal |
| CN108022139A (en) * | 2016-10-31 | 2018-05-11 | 北京嘀嘀无限科技发展有限公司 | Distribute the method and device of order |
| CN108022139B (en) * | 2016-10-31 | 2021-06-04 | 北京嘀嘀无限科技发展有限公司 | Order distribution method and device |
| WO2018076951A1 (en) * | 2016-10-31 | 2018-05-03 | Beijing Didi Infinity Technology And Development Co., Ltd. | Device and method for order distribution |
| WO2018095065A1 (en) * | 2016-11-23 | 2018-05-31 | 北京小度信息科技有限公司 | Method and apparatus for allocating data object, and electronic device |
| CN107392512A (en) * | 2016-11-25 | 2017-11-24 | 北京小度信息科技有限公司 | Grouped task method and apparatus |
| CN109118334A (en) * | 2017-06-01 | 2019-01-01 | 北京小度信息科技有限公司 | order processing method and device |
| WO2018218946A1 (en) * | 2017-06-01 | 2018-12-06 | 北京小度信息科技有限公司 | Order processing method and device |
| WO2019015661A1 (en) * | 2017-07-20 | 2019-01-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for service request allocation |
| TWI690867B (en) * | 2017-07-20 | 2020-04-11 | 大陸商北京嘀嘀無限科技發展有限公司 | Systems and methods for service request allocation |
| CN110447050A (en) * | 2017-12-04 | 2019-11-12 | 北京嘀嘀无限科技发展有限公司 | System and method for distributing order in online on-demand service |
| CN108182524B (en) * | 2017-12-26 | 2021-07-06 | 北京三快在线科技有限公司 | Order allocation method and device and electronic equipment |
| CN108182524A (en) * | 2017-12-26 | 2018-06-19 | 北京三快在线科技有限公司 | A kind of order allocation method and device, electronic equipment |
| CN110020818A (en) * | 2018-01-09 | 2019-07-16 | 车伯乐（北京）信息科技有限公司 | A kind of order processing method, apparatus, system and server |
| CN108537619A (en) * | 2018-03-05 | 2018-09-14 | 新智数字科技有限公司 | A kind of method for allocating tasks, device and equipment based on maximum-flow algorithm |
| CN110798800A (en) * | 2018-08-02 | 2020-02-14 | 北京嘀嘀无限科技发展有限公司 | Information pushing method, device, equipment and computer readable storage medium |
| CN108960747A (en) * | 2018-08-21 | 2018-12-07 | 安吉汽车物流股份有限公司 | Optimization method and device, storage medium, the terminal of Logistic Scheduling |
| CN108960747B (en) * | 2018-08-21 | 2021-10-01 | 安吉汽车物流股份有限公司 | Logistics scheduling optimization method and device, storage medium and terminal |
| CN110874777A (en) * | 2018-08-30 | 2020-03-10 | 北京嘀嘀无限科技发展有限公司 | Order processing method and device |
| CN109359843A (en) * | 2018-09-30 | 2019-02-19 | 山西游骑兵电子商务有限公司 | Data processing method and order equipment |
| CN109377317B (en) * | 2018-10-26 | 2021-03-30 | 天津五八到家科技有限公司 | Data processing method and device |
| CN109377317A (en) * | 2018-10-26 | 2019-02-22 | 天津五八到家科技有限公司 | Data processing method and device |
| TWI684107B (en) * | 2018-12-18 | 2020-02-01 | 國立中山大學 | Method for data imputation and classification and system for data imputation and classification |
| CN109636213A (en) * | 2018-12-19 | 2019-04-16 | 拉扎斯网络科技（上海）有限公司 | Order splitting, evaluation method and device, electronic equipment and storage medium |
| CN110009131B (en) * | 2019-02-28 | 2020-12-25 | 河海大学 | Network appointment vehicle dispatching method considering multi-factor influence |
| CN110009131A (en) * | 2019-02-28 | 2019-07-12 | 河海大学 | A kind of net about vehicle worksheet processing method considering multifactor impact |
| CN111724091A (en) * | 2019-03-21 | 2020-09-29 | 天津五八到家科技有限公司 | Order distribution method, server side equipment and storage medium |
| CN110322159A (en) * | 2019-07-10 | 2019-10-11 | 金瓜子科技发展（北京）有限公司 | A kind of data processing method and device |
| CN111882109A (en) * | 2020-06-22 | 2020-11-03 | 北京嘀嘀无限科技发展有限公司 | Order allocation method and system |
| CN112418616A (en) * | 2020-11-05 | 2021-02-26 | 深圳依时货拉拉科技有限公司 | Order broadcasting method and device, computer equipment and computer-readable storage medium |
| CN112418616B (en) * | 2020-11-05 | 2024-01-23 | 深圳依时货拉拉科技有限公司 | Order broadcasting method, order broadcasting device, computer equipment and computer readable storage medium |
| CN112396375A (en) * | 2020-11-18 | 2021-02-23 | 江苏满运物流信息有限公司 | Delivery method and device, storage medium and electronic device |
| CN112396375B (en) * | 2020-11-18 | 2022-06-21 | 江苏满运物流信息有限公司 | Delivery method and device, storage medium and electronic device |
| CN112465602B (en) * | 2020-12-11 | 2023-12-15 | 深圳依时货拉拉科技有限公司 | Order pushing method, order pushing device, computer equipment and computer readable storage medium |
| CN112465602A (en) * | 2020-12-11 | 2021-03-09 | 深圳依时货拉拉科技有限公司 | Order pushing method and device, computer equipment and computer readable storage medium |
| CN113159458B (en) * | 2021-05-20 | 2022-04-08 | 南京领行科技股份有限公司 | Distribution method and device of network appointment vehicles, electronic equipment and storage medium |
| CN113159458A (en) * | 2021-05-20 | 2021-07-23 | 南京领行科技股份有限公司 | Distribution method and device of network appointment vehicles, electronic equipment and storage medium |
| CN114331645A (en) * | 2022-03-14 | 2022-04-12 | 广州宸祺出行科技有限公司 | Method and system for improving transport capacity utilization rate of network appointment car |
| CN114723125A (en) * | 2022-04-01 | 2022-07-08 | 华侨大学 | Intercity vehicle order allocation method combining deep learning and multi-task optimization |
| CN114723125B (en) * | 2022-04-01 | 2024-06-28 | 华侨大学 | Inter-city vehicle order allocation method combining deep learning and multitask optimization |
| CN114827100A (en) * | 2022-04-26 | 2022-07-29 | 郑州锐目通信设备有限公司 | Taxi calling method and system |
| CN114827100B (en) * | 2022-04-26 | 2023-10-13 | 郑州锐目通信设备有限公司 | Taxi calling method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104599168A (en) | Method and device for allocating taxi-calling orders | |
| Zhou et al. | Multi-agent reinforcement learning for order-dispatching via order-vehicle distribution matching | |
| US11631333B2 (en) | Multi-agent reinforcement learning for order-dispatching via order-vehicle distribution matching | |
| Clemente et al. | The vehicle relocation problem in car sharing systems: Modeling and simulation in a petri net framework | |
| Berg et al. | Comparison of static ambulance location models | |
| Lowalekar et al. | Online repositioning in bike sharing systems | |
| Shavandi et al. | A fuzzy queuing location model with a genetic algorithm for congested systems | |
| CN111033535A (en) | System and method for bus order scheduling | |
| Teimoori et al. | A secure cloudlet-based charging station recommendation for electric vehicles empowered by federated learning | |
| CN109598368B (en) | Path decision evaluation and order distribution method and device, electronic equipment and medium | |
| Shi et al. | ParkCrowd: Reliable crowdsensing for aggregation and dissemination of parking space information | |
| CN112488400B (en) | Traffic travel behavior regulation and control method based on block chain technology and travel plan sharing | |
| Tang et al. | Value function is all you need: A unified learning framework for ride hailing platforms | |
| CN109146280B (en) | Method, device and system for pushing information | |
| Yu et al. | A dynamic holding strategy in public transit systems with real-time information | |
| CN110110903B (en) | Neural evolution-based distribution vehicle path planning method | |
| Cepolina et al. | A methodology for planning a new urban car sharing system with fully automated personal vehicles | |
| CN106600057A (en) | Express distribution task scheduling algorithm and apparatus | |
| CN110020215A (en) | Look for the method for pushing and device, electronic equipment of single recommendation information | |
| Fernández et al. | Bike3S: A tool for bike sharing systems simulation | |
| Cangialosi et al. | Designing a multimodal generalised ride sharing system | |
| Han et al. | Three‐stage electric vehicle scheduling considering stakeholders economic inconsistency and battery degradation | |
| Saharan et al. | DyPARK: A dynamic pricing and allocation scheme for smart on-street parking system | |
| Hu et al. | An artificial-neural-network-based model for real-time dispatching of electric autonomous taxis | |
| Tran et al. | Adaptive passenger-finding recommendation system for taxi drivers with load balancing problem |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150506 |