CN113554387A - Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium - Google Patents
Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113554387A CN113554387A CN202110719257.2A CN202110719257A CN113554387A CN 113554387 A CN113554387 A CN 113554387A CN 202110719257 A CN202110719257 A CN 202110719257A CN 113554387 A CN113554387 A CN 113554387A
- Authority
- CN
- China
- Prior art keywords
- data
- preference
- logistics order
- order
- logistics
- 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
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
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- 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
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0633—Lists, e.g. purchase orders, compilation or processing
- G06Q30/0635—Processing of requisition or of purchase orders
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Human Resources & Organizations (AREA)
- Mathematical Physics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses an e-commerce logistics order distribution method, device, equipment and storage medium based on driver preference, and the method comprises the following steps: generating data input to the motivation prediction HMM model from feedback data of the carrier vehicle and the logistics order data; the motivation prediction HMM neural network model outputs the motivation type of the carrier vehicle and the corresponding occurrence probability value; vehicle data of a carrier vehicle with an initiative type of active order grabbing and data of a logistics order are input into a corresponding carrier preference judgment neural network model; the carrier preference judges the output matching type and the corresponding confidence coefficient of the neural network model; and dispatching the logistics order to the carrier vehicle with the matching type being the matching preference and the confidence coefficient being more than or equal to the preset confidence coefficient threshold value. The method, the device, the equipment and the storage medium for distributing the E-commerce logistics orders based on the driver preference are beneficial to providing the method, the device, the equipment and the storage medium for efficiently realizing human-oriented E-commerce logistics order distribution based on the driver preference by predicting the intention and preference of the driver for accepting the logistics orders.
Description
Technical Field
The application relates to the field of e-commerce logistics, in particular to an e-commerce logistics order allocation method, device, equipment and storage medium based on driver preference.
Background
The existing small and medium-sized convenience stores usually carry out off-line commodity purchase at suppliers through own channels, and due to the characteristics of the small and medium-sized convenience stores, the small and medium-sized convenience stores cannot purchase goods in large batches, so that effective bargaining can not be carried out with the suppliers, and meanwhile, due to the requirement of wholesale of the suppliers on the purchase quantity, the small and medium-sized convenience stores need to guarantee a certain scale for each purchase, so that the inventory problem is caused.
From the perspective of suppliers, the scattered purchasing mode of small and medium-sized convenience stores leads to the increase of the warehousing cost of the suppliers, so that the supply price is kept high.
In the related art, as disclosed in chinese patent publication No. CN112801759A, a buyer, a seller and a carrier are integrated together through an e-commerce platform, so that orders of multiple buyers can be 'listed' in a single logistics order, thereby implementing centralized purchasing and centralized distribution, and reducing the purchasing cost of small and medium-sized convenience stores.
In another related art, in order to realize the optimal distribution of technical logistics, as disclosed in chinese patent publication No. CN112016876B, there is provided a method for automatically configuring goods by a computer program, which can configure the path and assembly scheme of a vehicle according to the location of a supplier and a merchant and the order situation.
In other related art, schemes for configuring logistics routes based on pick and delivery addresses and vehicle locations are provided.
However, none of these schemes are configured in real time, and they tend to order the items over a period of time, such as one day. And the temporarily generated orders cannot be distributed, so that real-time response is realized.
In other related art, to respond to real-time orders, a preemptive order is sent to vehicles in the vicinity of the target area to respond to temporary, random orders in a preemptive manner.
However, the method can not know the actual intention of the driver to take the order, and in many cases, the driver takes the order, but cancels the order, or does not finish the order actively because the logistics order is not matched with the usual delivery area or habit of the driver, thereby reducing the logistics distribution efficiency.
Disclosure of Invention
In order to solve the defects of the prior art, the application provides an e-commerce logistics order distribution method based on driver preference, which comprises the following steps: acquiring data of a newly added logistics order; sending data of the logistics order to a carrier vehicle within a preset range; receiving feedback data of whether the carrier vehicle accepts the logistics order; generating characteristic data according to the feedback data of the carrier vehicle receiving the logistics order and the data of the logistics order, and inputting the characteristic data into a motivation prediction HMM model corresponding to the carrier vehicle; the motivational prediction HMM model outputs a motivational type for the carrier vehicle and a corresponding occurrence probability value, the motivational type comprising: actively and randomly preempting the order; inputting vehicle data of the carrier vehicle with the motivation type of active order grabbing and the occurrence probability value larger than a preset occurrence probability threshold value and related data of the logistics order into a corresponding carrier preference judgment neural network model; the carrier preference judgment neural network model outputs a matching type and a corresponding confidence coefficient, wherein the matching type comprises: matching preferences, bias preferences, and unknown preferences; and dispatching the logistics order to the carrier vehicle of which the matching type is matching preference and the confidence coefficient is more than or equal to a preset confidence coefficient threshold value.
Further, the e-commerce logistics order distribution method based on driver preference further comprises the following steps: collecting historical motivation characteristic data of a user of each carrier vehicle, and training the motivation type prediction HMM model corresponding to each carrier vehicle; wherein the historical motivational characteristic data comprises: the system comprises a goods taking address, goods taking time, goods delivery address, goods delivery time, one-way time consumption, order dispatching time, order grabbing time, reaction time consumption, order receiving time and response time consumption; wherein the one-way elapsed time is equal to an absolute value of a time difference between the delivery time and the pickup time; the reaction time is equal to the absolute value of the time difference between the order dispatching time and the order grabbing time; the response elapsed time is equal to the absolute value of the time difference between the pickup time and the order taking time.
Further, the observable sequence of the motivation type prediction HMM model is historical motivation feature data, and the hidden state of the motivation type prediction HMM model is the motivation type.
Further, the value range of the occurrence probability threshold of the active form snatching type is greater than or equal to 60% to 100%.
Further, the e-commerce logistics order distribution method based on driver preference further comprises the following steps: collecting historical carrying characteristic data of a user of each carrying vehicle as input data to train the carrying preference judgment neural network model corresponding to each carrying vehicle; setting corresponding preference types for historical carrying characteristic data of a user of each carrying vehicle to serve as output data for training the carrying preference judgment neural network model corresponding to each carrying vehicle; wherein the historical shipper characteristic data comprises: the goods taking address, the goods taking time, the goods delivering address, the goods delivering time, the absolute length of the route, the actual length of the route and the occupied volume of the goods.
Further, the e-commerce logistics order distribution method based on driver preference further comprises the following steps: and determining the preference type according to at least the number of times that the pick-up address or/and the delivery address appears in the historical carrying characteristic data and all address numbers of the historical carrying characteristic data.
Further, the preference type is determined according to the ratio of the average value obtained by taking the absolute length of the route, the actual length of the route or \ and the occupied volume of the goods as the median to the average value of all data of the historical carrying characteristic data.
As another aspect of the present application, there is also provided an e-commerce logistics order distribution apparatus based on driver preference, including: the acquisition module is used for acquiring data of the newly added logistics order; the transmitting module is used for transmitting the data of the logistics order to the carrying vehicles within the preset range; a receiving module for receiving feedback data of whether the carrier vehicle accepts the logistics order; the motivation module is used for generating characteristic data according to feedback data of a carrier vehicle receiving the logistics order and data of the logistics order, inputting the characteristic data into a motivation prediction HMM model corresponding to the carrier vehicle, and enabling the motivation prediction HMM model to output a motivation type and a corresponding occurrence probability value of the carrier vehicle; a preference module, configured to input vehicle data of a carrier vehicle and related data of the logistics order, where the motivation type is active order grabbing and the occurrence probability value is greater than a preset occurrence probability threshold, into a corresponding carrier preference judgment neural network model and enable the carrier preference judgment neural network model to output a matching type and a corresponding confidence level, where the matching type includes: matching preferences, bias preferences, and unknown preferences; and the dispatching module is used for dispatching the logistics order to the carrier vehicle with the matching type being the matching preference and the confidence coefficient being greater than or equal to a preset confidence coefficient threshold value.
As another aspect of the present application, the present application further provides an e-commerce logistics order allocation apparatus based on driver preference, a memory for storing a computer program; a processor for implementing the method for assigning e-commerce logistics orders based on driver preferences as described above when executing the computer program.
As another aspect of the present application, there is also provided a computer client storage medium having a computer program stored therein, the computer program when executed by a processor implementing the steps of the e-commerce logistics order allocation method based on driver preference as previously described.
The application has the advantages that: the method, the device, the equipment and the storage medium for distributing the E-commerce logistics orders based on the driver preference are provided, and people-oriented goods distribution method, device and equipment based on the driver preference are efficiently realized by predicting the intention and preference of a driver for accepting the logistics orders.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, serve to provide a further understanding of the application and to enable other features, objects, and advantages of the application to be more apparent. The drawings and their description illustrate the embodiments of the invention and do not limit it. In the drawings:
FIG. 1 is a schematic block diagram illustrating the main steps of a method for assigning e-commerce logistics orders based on driver preferences according to one embodiment of the present application;
FIG. 2 is a schematic diagram illustrating steps of constructing an incentive type prediction HMM model in an e-commerce logistics order allocation method based on driver preferences according to an embodiment of the application;
fig. 3 is a schematic diagram illustrating steps of constructing a carrier preference decision neural network model in an e-commerce logistics order allocation method based on driver preferences according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a driver-side order grabbing interface in an e-commerce logistics order allocation method based on driver preferences according to an embodiment of the present application;
FIG. 5 is a schematic diagram of a driver-side route navigation interface in an e-commerce logistics order allocation method based on driver preferences according to one embodiment of the subject application;
FIG. 6 is a block diagram of an electronic commerce logistics order distribution apparatus based on driver preferences according to one embodiment of the subject application;
FIG. 7 is a block diagram of an electronic commerce logistics order distribution facility based on driver preferences according to one embodiment of the subject application;
FIG. 8 is a schematic step diagram of an active allocation method of an e-commerce logistics order allocation method based on driver preferences according to one embodiment of the present application;
FIG. 9 is a schematic diagram illustrating the conversion of a logistics order into a characteristic field in the active allocation method shown in FIG. 8;
FIG. 10 is a schematic diagram of the inputs and outputs of the active dispatching neural network of the active dispatching method of FIG. 8;
fig. 11 is a schematic diagram illustrating an "order sharing" distribution principle of an e-commerce logistics order distribution method according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It should be understood that the data so used may be interchanged under appropriate circumstances such that embodiments of the application described herein may be used. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In this application, the terms "upper", "lower", "left", "right", "front", "rear", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings. These terms are used primarily to better describe the present application and its embodiments, and are not used to limit the indicated devices, elements or components to a particular orientation or to be constructed and operated in a particular orientation.
Moreover, some of the above terms may be used to indicate other meanings besides the orientation or positional relationship, for example, the term "on" may also be used to indicate some kind of attachment or connection relationship in some cases. The specific meaning of these terms in this application will be understood by those of ordinary skill in the art as appropriate.
Furthermore, the terms "mounted," "disposed," "provided," "connected," and "sleeved" are to be construed broadly. For example, it may be a fixed connection, a removable connection, or a unitary construction; can be a mechanical connection, or an electrical connection; may be directly connected, or indirectly connected through intervening media, or may be in internal communication between two devices, elements or components. The specific meaning of the above terms in the present application can be understood by those of ordinary skill in the art as appropriate.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Referring to fig. 1, the e-commerce logistics order allocation method based on driver preference according to the present application includes the following main steps:
and S101, acquiring data of the newly added logistics order.
And S102, sending the data of the logistics order to the carrier vehicles within the preset range.
And S103, receiving feedback data of whether the carrier vehicle accepts the logistics order or not.
And S104, generating characteristic data according to the feedback data of the carrier vehicle receiving the logistics order and the data of the logistics order, and inputting the characteristic data into the motivation prediction HMM model corresponding to the carrier vehicle.
S105, outputting an motivation type of the carrier vehicle and a corresponding occurrence probability value by a motivation prediction HMM model, wherein the motivation type comprises: active order grabbing and random order grabbing.
And S106, inputting the vehicle data of the carrying vehicle with the motive type of active order grabbing and the occurrence probability value larger than a preset occurrence probability threshold value and the related data of the logistics order into the corresponding carrying preference judgment neural network model.
S107, the carrier preference judgment neural network model outputs a matching type and a corresponding confidence coefficient, wherein the matching type comprises the following steps: matching preferences, bias preferences, and unknown preferences.
And S108, dispatching the logistics order to the carrier vehicle with the matching type being the matching preference and the confidence coefficient being greater than or equal to the preset confidence coefficient threshold value.
As a specific solution, the new logistics order referred to herein is an order that is generated immediately and is not allocated. The preset range can be set in various ways, the simplest way is to divide all available carrier vehicles in the system into the preset range, in addition, the preset range can be divided according to the preset business circles and the carrier vehicles belonging to the business circles, as a preferred scheme, the preset range adopts a dynamic setting way, when the logistics order is generated, all the carrier vehicles upload position data, the system takes a certain geographical range (taking the goods taking address as the center of a circle and taking the preset distance as the radius) taking the goods taking address of the logistics order as the preset range, and sends the order grabbing data of the logistics order to the carrier vehicles in the circle center area, and a specific interface at the end of a driver can be shown in fig. 4.
The user of the carrier vehicle uses the terminal device to perform corresponding operations through driver-side software, so as to send a feedback data reflecting whether the user (the carrier vehicle) accepts the logistics order or not to the system (the system server). Preferably, the feedback data includes not only data on whether the user receives the logistics order, but also: order grabbing time and picture checking duration. The order grabbing time is the time when the user clicks to receive the logistics order, and if the user refuses to grab the order, the order grabbing time is null; the image checking duration is the time for the user to check the logistics order delivery address and the pickup address navigation map.
Specifically, the motivation prediction HMM model is essentially an HMM model (hidden markov model), and the basic idea of the present application is to use the psychological motivation of the driver end user to pick up an order as the hidden state of the HMM model, and use the action of the driver end user to pick up an order and the post-pick up action as the observable sequence.
And (3) forming an motivation prediction HMM model for each driver end user, and training and constructing the model by adopting the past data of the driver end user.
Specifically, as shown in fig. 2, the method for constructing the motivational type prediction HMM model includes:
s201, original logistics order data of the carrier vehicle are obtained.
S202, historical motivation characteristic data are calculated or generated based on the original logistics data.
And S203, classifying the motive types of the historical motive characteristic data according to the actual logistics order completion condition.
And S204, taking the historical motive feature data as an observation sequence, taking the corresponding motive type as a hidden state, and training a motive type prediction HMM model corresponding to the driver end.
S205, storing the motivation type prediction neural network model and parameters thereof.
Specifically, the historical motivational characteristic data includes: pick-up address, pick-up time, delivery address, delivery time, one-way time consumption, order delivery time, order grabbing time, reaction time consumption, order receiving time and response time consumption.
Wherein the one-way elapsed time is equal to the absolute value of the time difference between the delivery time and the pickup time; the reaction time consumption is equal to the absolute value of the time difference between the order dispatching time and the order grabbing time; the response elapsed time is equal to the absolute value of the time difference between the pickup time and the order taking time.
Through the construction of the historical motivation characteristic data, the sequence characteristics of the user during order grabbing and delivery can be accurately reflected. The goods taking address, the goods taking time, the goods delivery address and the goods delivery time reflect the basic attribute characteristics of the group of characteristic data.
A single pass time is the amount of time a user takes from pick to delivery when completing a single pass. The single pass time-consuming reflects the time-attribute characteristics of the set of characteristic data.
The order dispatching time is the time when the system sends the logistics order to the driver end, the order grabbing time is the time when the driver end user clicks to receive the logistics order, and the order receiving time is the time when the system formally sends the logistics order to the user. The order dispatching time, the order grabbing time, the reaction time consumption, the order receiving time and the response time consumption reflect the quantitative attribute characteristics of the characteristic data, namely whether the willingness of order grabbing and order completing is strong or not.
As a further preferred solution, the historical motivational characteristic data further includes a chart duration in the feedback data. Preferably, zero values are set for the data of the incomplete links for rejected or cancelled logistics orders.
As a specific scheme, the specific method for classifying the motivation types of the historical motivation characteristic data according to the actual logistics order completion condition is that the motivation type of the completely completed logistics order is set as an active form-grabbing, and the motivation type of the uncompleted logistics order is set as a passive form-grabbing. More preferably, the motivational type classification is based on the ratio of the response time to the one-way time.
Specifically, the one-way elapsed time is Th, and the response elapsed time is Tx, and their ratio q, i.e., q = Th/Tx. And then, comparing the Q value with a preset ratio threshold Q, if Q is greater than the ratio threshold Q, determining that the type of the motivation of the group of data is active order grabbing, and otherwise, determining that the motivation is passive order grabbing. The unfinished logistics orders are uniformly classified as passive order grabbing. As a preferred scheme, the value range of the ratio threshold value Q is 4 to 22; more specifically 5 to 10, the value of the ratio threshold Q is 7.6 as a more precise range.
After an HMM model for predicting the motivation type is trained, generating characteristic data items to be judged according to the current feedback data and existing data in the data of the logistics order.
For example, if a driver end receives a robbed order, the characteristic data of the order as shown in fig. 10 should be as follows: 123 # along the longitude and latitude area of Hangzhou city, sky, 567 # along the longitudinal and horizontal areas of Xiaoshan area of Hangzhou city, sky, 11 minutes at 11 points for 23 seconds, 24 minutes for 55 seconds, 1 minute for 44 seconds, sky and sky. All addresses may be put into digital data format, streets replaced with special numbers, and time may be set into digital format as well.
As can be seen from the above, many null data appear in the feature data, and even if the null data is assigned with a value, for example, the value is assigned as 0, the null data will not affect the model, because the a posteriori data is always null data in the prediction process, but the effect of the hidden state on the observation sequence cannot be reflected without using the feature data after the HMM model is predicted by training the motivation type.
Preferably, the training is still performed by using the data of the picking time, the delivery time, the one-way time consumption, the order receiving time and the response time consumption, and during the prediction, only the predictable characteristic data is input for prediction processing.
Preferably, the feature data input to the motivational type prediction HMM model includes: the system comprises a goods picking address, a goods delivering address, one-way time consumption, order dispatching time, order grabbing time, reaction time consumption and a chart checking time.
After being processed, the motive type prediction HMM model outputs the probability that the observation sequence corresponds to the hidden state, namely the active order grabbing or the passive order grabbing, for example, the probability is 78%. Then, the system judges whether the probability exceeds the occurrence probability threshold of the active order grabbing type, and if the threshold is 70%, the order grabbing action can be considered as the active order grabbing.
It should be noted that, based on historical data, parameters in the motivation type prediction HMM model, such as conversion probability values between hidden states, and the like, can be calculated by using an existing algorithm, and meanwhile, a complete HMM model tool is configured in the prior art, and the creativity of the technical solution of the application is not embodied in the HMM model itself, but lies in a specific application scenario of solving logistics single motivation detection by using the HMM model itself.
After determining which driver-side users have real motivation to complete the logistics task according to the above method, it is necessary to analyze whether the order meets the driver's preference.
In a general sense, when it is determined that the driver has the real intention of getting an order, the logistics order can be distributed to the driver end users arranged in the front according to the sorting mode of distance and the like.
But many times the specific scenarios are: from the aspect of time, a plurality of newly added or possibly added logistics orders are available, the orders are more suitable for a driver end user than the current orders, but the driver end user cannot predict the future and can be allocated with an order which does not accord with the habit and the preference of the driver end user, and at the moment, if the order which accords with the preference comes in, the driver end user still can rob the order (different from the situation that the order grabbing such as drip-drop bus is in an idle state, and for the logistics 'order-piecing' scene, the system expects that a vehicle can simultaneously bear a plurality of logistics orders), so that the waste of system resources is caused; from a group distribution perspective, if the driver is not considered to prefer to distribute in a familiar area, it may result in a low logistics distribution efficiency, and the driver always performs order taking in an unfamiliar area (as a result of distance determination), because the driver is already in an unfamiliar area in position, and because the need to continue order taking reduces the return cost, resulting in a real willingness to perform order taking, but rather reduces the enthusiasm of the carrier users over time.
Therefore, the electronic commerce logistics order distribution method based on the preference of the driver realizes the identification of the preference of the user by constructing a carrying preference judgment neural network model.
Specifically, the carrier preference determination neural network model is used for outputting a corresponding matching type according to input logistics order data.
As shown in fig. 3, the method for constructing the carrier preference determination neural network model includes:
s301, original logistics order data of the carrier vehicle are obtained.
S302, historical carrying characteristic data is calculated or generated based on the original data.
And S303, determining a corresponding preference type according to the historical carrying characteristic data.
S304, taking the historical carrier characteristic data as input data, taking the corresponding preference type as output data, and training the preference judgment neural network model corresponding to the driver end until the model converges.
S305, storing the preference judgment neural network model and the parameters thereof.
As a specific scheme, collecting historical carrying characteristic data of a user of each carrying vehicle as input data to train a carrying preference judgment neural network model corresponding to each carrying vehicle; setting corresponding preference types for historical carrying characteristic data of a user of each carrying vehicle to serve as output data for training a carrying preference judgment neural network model corresponding to each carrying vehicle; wherein the historical shipper characteristic data comprises: the goods taking address, the goods taking time, the goods delivering address, the goods delivering time, the absolute length of the route, the actual length of the route and the occupied volume of the goods.
The core technical point of the scheme is how to determine whether the historical shipping characteristic data is the user preference and the corresponding to the matching type, and as a preferred scheme, the preference type is determined according to the number of times that the pickup address or/and the shipping address appear in the historical shipping characteristic data and all the addresses of the historical shipping characteristic data.
The frequency of the pick-up address and the delivery address in the historical data can reflect the range characteristics of the activity area of the user at the terminal of the driver.
As a more specific technical scheme, the number of times of occurrence of a pickup address of a current logistics order is Z1, and the total number of pickup addresses in all historical data is Z2, then a pickup address ratio Z can be obtained, wherein Z = Z1/Z2; similarly, if the number of times of occurrence of the delivery address of the current logistics order is Y1 and the total number of the delivery addresses in all the historical data is Y2, a delivery address ratio Y can be obtained, wherein Y = Y1/Y2; the number of times that the picking address and the delivery address of the current logistics order are repeated simultaneously is X1, and the total number of orders in the historical data is X2, then a picking and delivery address ratio value X can be obtained, wherein X = X1/X2.
Then, according to the formula C =0.2z +0.3y +0.5x, the address preference coefficient C is calculated, and then it is determined whether the address preference coefficient C is greater than or equal to a preset address preference coefficient threshold value, which may be a variable empirical value.
As a preferable scheme, the judgment is carried out by the address preference coefficient C, and the preference type is determined according to the ratio of the average value obtained by taking the absolute length of the route, the actual length of the route or \ and the occupied volume of the goods as the median to the average value of all data of the historical carrying characteristic data.
Specifically, the generation of the capacity preference coefficient G also needs to be calculated.
For convenience of introduction, the present application is defined as follows, assuming that the historical data is 1, 2, 3, 5, 6, 7, 8, the current data is 4, and the truncated data set is 1, 2, 3, 5, 6, 7 with 4 as the median, in which case the median is 4, and the overall average is 1, 2, 3, 5, 6, 7, 8.
Similarly, the absolute length of the route (i.e. the straight-line distance between the pick-up address and the delivery address) of the current logistics order is used as a median, the absolute lengths of a plurality of routes in the history data are intercepted, then the average value of the absolute lengths of the routes (excluding the logistics order data) is calculated and defined as a median absolute route average value H1, then the average value of the absolute lengths of all routes (excluding the logistics order data) in the history data is calculated and defined as a whole absolute route average value H2, and the ratio of the median absolute route average value H1 to the whole absolute route average value H2 is defined as an absolute route ratio H.
Similarly, the actual length of the route of the current logistics order (i.e. the distance between the pick-up address and the delivery address, the actual length of the route of the current logistics order is based on the navigation road recommended by the system, and the actual distance of the route of the logistics order in the history data is based on the distance of the actual route of the logistics order in 1, 2, 3, 5, 6, 7, 8) is taken as a median, several actual lengths of the routes in the history data are intercepted, then the average value of the actual lengths of the routes (excluding the data of the logistics order) is calculated and defined as a median actual route average value J1, then the average value of the actual lengths of all routes (excluding the data of the logistics order) in the history data is calculated and defined as an overall actual route average value J2, and the ratio of the median actual route average value J1 to the overall actual route average value J2 is defined as an actual route ratio J.
Similarly, the cargo occupancy volume of the current logistics order (i.e. the volume occupied by the cargo in the order) is used as a median, the cargo occupancy volumes in the historical data are intercepted, then the average values (not including this time) of the cargo occupancy mentioned above are obtained, and defined as a median occupancy volume average value P1, then the average value of all the cargo occupancy volumes (not including this time) in the historical data is calculated and positioned as a whole occupancy volume average value P2, and the median occupancy volume average value P1 and the whole occupancy volume average value P2 are defined as an actual route ratio P.
The transport capacity preference coefficient G =0.5h +0.75j +0.4p, and after the transport capacity preference coefficient G is calculated, whether the transport capacity preference coefficient G is greater than or equal to a preset transport capacity preference coefficient threshold value is judged, and the transport capacity preference coefficient threshold value may be a variable empirical value.
If the address preference coefficient C and the transport capacity preference coefficient G both meet the judgment condition (are greater than or equal to the threshold value), the matching type is set as the matching preference; if one of the address preference coefficient C and the transport capacity preference coefficient G satisfies a judgment condition (is greater than or equal to a threshold value), the matching type is set to unknown preference; if neither the address preference coefficient C nor the capacity preference coefficient G satisfies the judgment condition (is smaller than the threshold), the matching type is set as the deviation preference.
The prepared historical carrier characteristic data is used as input data, correspondingly set (automatically set) preference types are used as output data, and then the carrier preference judgment neural network model corresponding to each driver end is trained respectively.
After the judgment of the motivation is completed, the carrying vehicles meeting the active order grabbing condition are selected, carrying characteristic data is generated according to data of the current logistics order, namely a goods picking address, goods picking time, a goods delivery address, goods delivery time, absolute length of a route, actual length of the route and occupied volume of goods are input into a carrying preference judgment neural network model, the carrying preference judgment neural network model outputs a matching type and a corresponding confidence coefficient, if the confidence coefficient is greater than a threshold value, the type condition is set strictly in the previous learning process, the threshold value range of the confidence coefficient can be properly relaxed, as an optimal scheme, the value range is 78% -100%, as a specific scheme is 85%, a candidate ranking list is entered, the confidence coefficient is used as a ranking basis, and the ranking is more advanced when the confidence coefficient is larger. The system dispatches the logistics order to the first ranked carrier vehicle.
And the carrier preference judgment neural network model is a convolutional neural network model.
As shown in fig. 6, the present application also provides an e-commerce logistics order distribution apparatus 400 based on driver preference, which includes: an acquisition module 401, a sending module 402, a receiving module 403, an motivation module 404, a preference module 405, and a billing module 406. The obtaining module 401 is configured to obtain data of a newly added logistics order; the sending module 402 is configured to send data of the logistics order to the carrier vehicles within a preset range; the receiving module 403 is configured to receive feedback data of whether the logistics vehicle receives the logistics order; the motivation module 404 is configured to input vehicle data and feedback data of the carrier vehicle receiving the logistics order into corresponding motivation prediction HMM neural network models respectively and cause the motivation prediction HMM neural network models to output motivation types and corresponding occurrence probability values of the carrier vehicle; the preference module 405 is configured to input vehicle data of the carrier vehicle and related data of the logistics order, where the motivation type is active order grabbing and the occurrence probability value is greater than a preset occurrence probability threshold, into the corresponding carrier preference judgment neural network model and enable the carrier preference judgment neural network model to output a matching type and a corresponding confidence level, where the matching type includes: matching preferences, bias preferences, and unknown preferences; the order module 406 is configured to order the logistics order to a carrier vehicle with a matching type as a matching preference and a confidence level greater than or equal to a preset confidence level threshold.
Alternatively, the driver preference-based e-commerce logistics order distribution device 400 can further include a main control module (not shown in the figure), which can implement the rest of the driver preference-based e-commerce logistics order distribution method of the present application.
As another aspect of the present application, as shown in fig. 7, the present application further provides an e-commerce logistics order distribution apparatus 500 based on driver preference, which includes a memory 501 and a processor 502. Wherein the memory 501 is adapted to store a computer program and the processor 502 is adapted to carry out the steps of the method as provided above when executing the computer program.
The present application also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method as provided above.
It should be noted that for the description of the relevant parts of the device, the equipment and the storage medium for allocating the e-commerce logistics order based on the driver preference provided by the embodiment of the present invention, reference is made to the detailed description of the corresponding parts in the method for allocating the e-commerce logistics order based on the driver preference provided by the embodiment of the present invention, and details are not repeated herein. In addition, parts of the technical solutions provided in the embodiments of the present invention that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail, so as to avoid redundant description.
As an extension of the technical solution of the present application, the electronic commerce logistics order allocation method based on driver preference of the present application may further include a method of actively allocating an order, and specifically, the method of actively allocating an order includes:
acquiring character data of a newly added logistics order;
dividing the character data into a plurality of characteristic fields according to a classification rule of a preset characteristic parameter;
inputting the characteristic field into a trained artificial neural network model;
outputting, by the artificial neural network model, a recommended vehicle for carrying the logistics order and a confidence level corresponding to the recommended vehicle;
judging whether the confidence of the recommended vehicle is greater than a preset confidence threshold;
and when the confidence coefficient of the recommended vehicle is greater than the confidence coefficient threshold value, the newly added logistics orders are dispatched to the recommended vehicle.
As a specific scheme, in order to implement the method of the present application, an artificial neural network model needs to be constructed in advance, and as a preferred scheme, the artificial neural network model may be an RNN neural network model, a CNN neural network model, or an LSTM neural network model.
The training set data and the test set of the artificial neural network model comprise two parts, wherein one part is a preset data set, namely a logistics order which is distributed by a logistics distribution system, such as historical logistics distribution data or logistics distribution data distributed on the current day, and the other part is a dynamic data set, namely the logistics distribution system generates the logistics distribution data through order grabbing or dynamic manual assignment.
Preferably, as shown in fig. 11, the driver side of the vehicle in the logistics system of the present application can perform "order sharing" delivery according to the addresses of the seller and the buyer, so as to generate the whole logistics order for the driver side. When the artificial neural network model training is performed, in order to match the user habits of the driver and analyze the delivery habits of the driver in the overall logistics order, the overall logistics order data is decomposed into individual logistics orders, that is, orders with only one delivery address corresponding to one receiving address, for example, the logistics orders taken by the driver in fig. 11 are split into individual logistics orders from a supplier a to a store 1, from the supplier a to a store 4, from the supplier B to a store 2, and the like, and the individual logistics orders are used as a training set and a test set. The above data of the logistics order can also be split based on this to obtain a "separate" logistics order.
The logistics orders are preferably subjected to data cleaning and a standard loop, and the logistics order data are labeled into a standard format of the logistics order shown in fig. 9, wherein the expected occupied space volume can be calculated through unique SKU codes and numbers and prestored volume data as a preferable scheme. As an optimal scheme, the predicted occupied space volume is not adopted as a characteristic field during the training of the artificial neural network model, so that the influence of the predicted occupied space volume on the training and analysis results is reduced.
As a training set and a testing set of the artificial neural network model, the logistics order can be converted into a feature field with a specific format, and particularly, the feature field shown in fig. 9 and fig. 10 can be referred to. Likewise, the characteristic field may not contain a field of the expected occupied space volume. The field of the expected occupied space volume can be used for avoiding the problem that the logistics order selects the vehicle type in a wrong way.
When the artificial neural network model is trained to converge through the data, some data sets which are not used as training sets and test sets are collected to be used as verification sets to adjust the verification sets of the hyper-parameters of the artificial neural network model.
As a basic idea, the driver of the vehicle responsible for the logistics order, often for personal or vehicular reasons, tends to or is better suited to accept a logistics order in a certain area or with certain geographical characteristics; in the case where there is history data, the logistics order itself may be taken as text field data to be classified, and the recommended vehicle may be taken as the type of classification. Since the vehicle data in the system is known, this corresponds to a problem of classification based on text.
The logistics method utilizes the characteristic of logistics 'order matching' and is different from the common passenger transportation order matching and order grabbing modes, the logistics characteristic of delivery is that the operation range of vehicles is relatively fixed, and the passenger transportation order matching and order grabbing, such as dripping and hitting, have larger vehicle operation range and more random positions.
Compared with the application of the conventional artificial neural network in logistics order matching, the method and the system have the advantages that the text classification mode is creatively adopted to carry out the rapid allocation of the temporarily added logistics orders. In the conventional logistics distribution of artificial intelligence, a neural network is often used for route planning or area division, for example, a K-means clustering mode and the like are adopted, but the method does not pay attention to data such as distance and the like, and only excavates delivery addresses, pickup addresses and the adaptation degree of vehicles from historical orders for distribution.
As a more specific technical solution, referring to fig. 8, the logistics distribution method applicable to e-commerce orders according to the present application further includes the following steps: when the confidence coefficient of the recommended vehicle is smaller than or equal to the confidence coefficient threshold value, inputting the characteristic field into the artificial neural network model again; judging whether the times that the confidence coefficient of the recommended vehicle is less than or equal to the confidence coefficient threshold exceeds a preset time threshold or not; when the times that the confidence coefficient of the recommended vehicle is less than or equal to the confidence coefficient threshold value are greater than the time threshold value, the newly added logistics orders are processed by a common logistics distribution program; and when the times that the confidence coefficient of the recommended vehicle is less than or equal to the confidence coefficient threshold value are less than or equal to the time threshold value, inputting the newly added logistics order into the artificial neural network model.
The way the artificial neural network model outputs the recommended vehicle is to output the driver account number or user name data of the recommended vehicle. Since the account ID is a unique identifier in the system, the artificial neural network model preferably outputs ID data of the driver account, i.e., a list of strings.
Specifically, referring to fig. 8, when the confidence of the recommended vehicle is less than or equal to the confidence threshold, a counting procedure is executed, whether the currently counted value (i.e., the number of times less than or equal to the confidence threshold) exceeds the number threshold is determined, if the currently counted value exceeds the number threshold, the newly added logistics order is processed by a general logistics distribution procedure, and if the currently counted value does not exceed the number threshold, the newly constructed artificial neural network model is still used for processing again.
As a further preferable scheme, the number threshold may be a dynamic value, and as a specific scheme, the number threshold takes a value of N, where N is a positive integer. The value of N can adopt the following method:
firstly, acquiring a task specific gravity value M, wherein the formula of the task specific gravity value M is as follows: m = T2/L-T1/L + K.
Wherein, T1 is newly added logistics orders which are not allocated, T2 is logistics orders which are already allocated (including planned allocation and temporary classification), L is the number of vehicles of available vehicles in the system, K is a correction coefficient, and the correction coefficient K = V1/V2, wherein V1 is the transport volume used by the available vehicles in the system, and V2 is the transport volume not used by the available vehicles in the system.
After the value of M is calculated, the value of M is rounded to obtain the number threshold N. As a preferable scheme, in order to avoid the efficiency reduction caused by the over-circulation of the system, an upper limit value is set for the number threshold N, and the value of the upper limit value ranges from 5 to 9.
As a further preferred scheme, the general logistics distribution program is to obtain the positions of the vehicles, obtain a list of vehicles within a preset range by taking the pickup location as the center of a circle, call the logistics orders which are being distributed by the vehicles, judge the next destination of the current vehicle, use the distance between the next destination and the delivery address in the newly added order as the basis for sorting, and then use the vehicle at the first position in the sorting as the recommended vehicle. Of course, other general logistics distribution procedures may be employed.
In order to dynamically adjust the matching capability of the artificial neural network model, as shown in fig. 8, the logistics distribution method applicable to the e-commerce order further includes the following steps: acquiring a newly added logistics order and processing the output recommended vehicle by a general logistics distribution program; judging whether the data output by the general logistics distribution program processing exists in the recommended vehicle data output by the artificial neural network model; if the data output by the general logistics distribution program processing exists in the recommended vehicle data output by the artificial neural network model, adjusting a preset confidence threshold value to be a confidence corresponding to the recommended vehicle data; and if the data output by the general logistics distribution program processing does not exist in the recommended vehicle data output by the artificial neural network model, taking the newly added logistics order and the recommended vehicle output by the general logistics distribution program as training data of the artificial neural network.
By adopting the scheme, the artificial neural network model can dynamically adapt to the actual distribution situation of the logistics task, so that the artificial neural network model can more efficiently adapt to the requirement of rapid distribution of the temporarily newly-added order.
As a further scheme, as shown in fig. 8, the logistics distribution method suitable for e-commerce orders further includes the following steps: acquiring feedback data of the recommended vehicle after the newly added logistics order is dispatched to the recommended vehicle; judging whether the recommended vehicle receives a logistics order or not; and when the recommended vehicle receives the logistics order, storing the data of the recommended vehicle and the logistics order, generating corresponding logistics task data and respectively sending the logistics task data to the goods taking fee and the goods receiver.
As shown in fig. 4, the driver-side device (typically a mobile phone) receives the interface containing the logistics order data, and the driver of the recommended vehicle can receive or reject the system's order through the interface, or can view information such as location and route through a map provided by the system. Once the driver end user receives the logistics task, the system stores data of recommended vehicles and the logistics order, generates corresponding logistics task data and respectively sends the logistics task data to the goods taking fee and the goods receiver; the logistics task data and the driver account ID data are stored and can be used as training set data of an artificial neural network model to further train the artificial neural network.
Alternatively, when the driver-side user rejects the logistics order dispatched by the system, the driver-side user recommended by the next system, that is, the next recommended vehicle, may send the logistics order, and then determine whether to pick up the order, and when a continuous non-order-picking condition occurs, for example, three continuous non-order-picking conditions occur, the order dispatching method based on the order grabbing introduced above is adopted.
The generation and the sequencing of the recommended vehicles can adopt two schemes, wherein one scheme is the sequencing according to a general logistics distribution program, which has been introduced in the above, the sequencing is carried out according to the distance of the delivery address, the efficiency seems to be the highest, but the individual delivery will of the driver user does not depend on the distance, so that as another preferable scheme, at least two or more recommended vehicles and corresponding confidence degrees can be output in the step of outputting the confidence degrees of the recommended vehicles and the corresponding recommended vehicles for carrying logistics orders by the artificial neural network model; the recommended vehicles with the confidence degrees larger than the confidence degree threshold value are sorted according to the confidence degrees, and then the recommendation of the vehicles can be carried out according to the order. Preferably, the ranking can be performed according to the confidence level whether the confidence level exceeds a preset value or not.
As a further preferred scheme, three artificial neural network models can be constructed, wherein the first group of artificial neural network models adopts logistics order data extracted by decomposing an overall logistics order in a planned logistics order (i.e. centrally distributing logistics distribution orders for the next day according to the daily ordering condition) into a logistics order corresponding to a single address as a training set; the second group of artificial neural network models adopt temporarily newly added logistics order data (namely logistics orders which need to be sent in the day of generation) as a training set; the third group of artificial neural network models adopt logistics order data of general logistics order grabbing as a training set; the generation and the ranking of the recommended vehicles can also adopt a scheme that the three groups of artificial neural networks are respectively output and then ranking is carried out according to the confidence degree.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. An e-commerce logistics order distribution method based on driver preference is characterized in that:
the e-commerce logistics order distribution method based on driver preference comprises the following steps:
acquiring data of a newly added logistics order;
sending data of the logistics order to a carrier vehicle within a preset range;
receiving feedback data of whether the carrier vehicle accepts the logistics order;
generating characteristic data according to the feedback data of the carrier vehicle receiving the logistics order and the data of the logistics order, and inputting the characteristic data into a motivation prediction HMM model corresponding to the carrier vehicle;
the motivational prediction HMM model outputs a motivational type for the carrier vehicle and a corresponding occurrence probability value, the motivational type comprising: actively and randomly preempting the order;
inputting vehicle data of the carrier vehicle with the motivation type of active order grabbing and the occurrence probability value larger than a preset occurrence probability threshold value and related data of the logistics order into a corresponding carrier preference judgment neural network model;
the carrier preference judgment neural network model outputs a matching type and a corresponding confidence coefficient, wherein the matching type comprises: matching preferences, bias preferences, and unknown preferences;
and dispatching the logistics order to the carrier vehicle of which the matching type is matching preference and the confidence coefficient is more than or equal to a preset confidence coefficient threshold value.
2. The driver preference-based e-commerce logistics order distribution method as claimed in claim 1, wherein:
the e-commerce logistics order distribution method based on driver preference further comprises the following steps:
collecting historical motivation characteristic data of a user of each carrier vehicle, and training the motivation type prediction HMM model corresponding to each carrier vehicle;
wherein the historical motivational characteristic data comprises: the system comprises a goods taking address, goods taking time, goods delivery address, goods delivery time, one-way time consumption, order dispatching time, order grabbing time, reaction time consumption, order receiving time and response time consumption; wherein the one-way elapsed time is equal to an absolute value of a time difference between the delivery time and the pickup time; the reaction time is equal to the absolute value of the time difference between the order dispatching time and the order grabbing time; the response elapsed time is equal to the absolute value of the time difference between the pickup time and the order taking time.
3. The driver preference-based e-commerce logistics order distribution method of claim 2, wherein:
the observable sequence of the motivation type prediction HMM model is historical motivation characteristic data, and the hidden state of the motivation type prediction HMM model is the motivation type.
4. The driver preference-based e-commerce logistics order distribution method of claim 3, wherein:
the value range of the occurrence probability threshold of the active order grabbing type is more than or equal to 60% to 100%.
5. The driver preference-based e-commerce logistics order distribution method as claimed in claim 1, wherein:
the e-commerce logistics order distribution method based on driver preference further comprises the following steps:
collecting historical carrying characteristic data of a user of each carrying vehicle as input data to train the carrying preference judgment neural network model corresponding to each carrying vehicle;
setting corresponding preference types for historical carrying characteristic data of a user of each carrying vehicle to serve as output data for training the carrying preference judgment neural network model corresponding to each carrying vehicle;
wherein the historical shipper characteristic data comprises: the goods taking address, the goods taking time, the goods delivering address, the goods delivering time, the absolute length of the route, the actual length of the route and the occupied volume of the goods.
6. The driver preference-based e-commerce logistics order distribution method of claim 5, wherein:
the e-commerce logistics order distribution method based on driver preference further comprises the following steps:
and determining the preference type according to at least the number of times that the pick-up address or/and the delivery address appears in the historical carrying characteristic data and all address numbers of the historical carrying characteristic data.
7. The driver preference-based e-commerce logistics order distribution method of claim 6, wherein:
and determining the preference type according to the ratio of the average value obtained by taking the absolute length of the route, the actual length of the route or \ and the occupied volume of the goods as the median to the average value of all data of the historical carrying characteristic data.
8. An e-commerce logistics order distribution device based on driver preference, characterized in that:
the e-commerce logistics order distribution device based on driver preference comprises:
the acquisition module is used for acquiring data of the newly added logistics order;
the transmitting module is used for transmitting the data of the logistics order to the carrying vehicles within the preset range;
a receiving module for receiving feedback data of whether the carrier vehicle accepts the logistics order;
the motivation module is used for generating characteristic data according to feedback data of a carrier vehicle receiving the logistics order and data of the logistics order, inputting the characteristic data into a motivation prediction HMM model corresponding to the carrier vehicle, and enabling the motivation prediction HMM model to output a motivation type and a corresponding occurrence probability value of the carrier vehicle;
a preference module, configured to input vehicle data of a carrier vehicle and related data of the logistics order, where the motivation type is active order grabbing and the occurrence probability value is greater than a preset occurrence probability threshold, into a corresponding carrier preference judgment neural network model and enable the carrier preference judgment neural network model to output a matching type and a corresponding confidence level, where the matching type includes: matching preferences, bias preferences, and unknown preferences;
and the dispatching module is used for dispatching the logistics order to the carrier vehicle with the matching type being the matching preference and the confidence coefficient being greater than or equal to a preset confidence coefficient threshold value.
9. An e-commerce logistics order distribution device based on driver preferences, characterized in that:
the e-commerce logistics order distribution device based on driver preference comprises:
a memory for storing a computer program;
a processor for implementing the driver preference based e-commerce logistics order allocation method of any one of claims 1 to 7 when executing the computer program.
10. A computer client storage medium, characterized in that: the computer-readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps of the driver preference based e-commerce logistics order allocation method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110719257.2A CN113554387A (en) | 2021-06-28 | 2021-06-28 | Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110719257.2A CN113554387A (en) | 2021-06-28 | 2021-06-28 | Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113554387A true CN113554387A (en) | 2021-10-26 |
Family
ID=78131010
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110719257.2A Pending CN113554387A (en) | 2021-06-28 | 2021-06-28 | Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113554387A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114493405A (en) * | 2021-12-21 | 2022-05-13 | 重庆特斯联智慧科技股份有限公司 | Dispatching method and dispatching system based on movement track of logistics robot |
CN116029704A (en) * | 2023-03-27 | 2023-04-28 | 浙江虎哥数字科技有限公司 | Method, system and medium for recycling recyclable articles |
CN116433138A (en) * | 2023-06-13 | 2023-07-14 | 长沙争渡网络科技有限公司 | Logistics platform information pushing method and system based on genetic algorithm |
CN116681373A (en) * | 2023-08-02 | 2023-09-01 | 四川集鲜数智供应链科技有限公司 | Logistics supply chain management method |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107133697A (en) * | 2017-05-03 | 2017-09-05 | 百度在线网络技术(北京)有限公司 | Estimate method, device, equipment and the storage medium of driver's order wish |
CN110119928A (en) * | 2019-05-07 | 2019-08-13 | 宏图物流股份有限公司 | A kind of vehicle match recommended method based on driver's feature |
CN110869953A (en) * | 2018-02-06 | 2020-03-06 | 北京嘀嘀无限科技发展有限公司 | System and method for recommending transportation travel service |
CN110874700A (en) * | 2018-09-03 | 2020-03-10 | 菜鸟智能物流控股有限公司 | Logistics order matching method and device and electronic 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 |
US20200380629A1 (en) * | 2019-06-03 | 2020-12-03 | International Business Machines Corporation | Intelligent on-demand management of ride sharing in a transportation system |
CN112488806A (en) * | 2020-12-18 | 2021-03-12 | 深圳依时货拉拉科技有限公司 | Method and device for predicting order willingness, computer equipment and computer-readable storage medium |
CN112488794A (en) * | 2020-12-11 | 2021-03-12 | 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 |
CN112580945A (en) * | 2020-12-08 | 2021-03-30 | 吉首大学 | Dispatching method based on multiple correlation coefficients and vehicle dispatching optimization system |
CN112825182A (en) * | 2019-11-21 | 2021-05-21 | 北京京东尚科信息技术有限公司 | Method and device for determining recommended commodities |
-
2021
- 2021-06-28 CN CN202110719257.2A patent/CN113554387A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107133697A (en) * | 2017-05-03 | 2017-09-05 | 百度在线网络技术(北京)有限公司 | Estimate method, device, equipment and the storage medium of driver's order wish |
CN110869953A (en) * | 2018-02-06 | 2020-03-06 | 北京嘀嘀无限科技发展有限公司 | System and method for recommending transportation travel service |
CN110874700A (en) * | 2018-09-03 | 2020-03-10 | 菜鸟智能物流控股有限公司 | Logistics order matching method and device and electronic 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 |
CN110119928A (en) * | 2019-05-07 | 2019-08-13 | 宏图物流股份有限公司 | A kind of vehicle match recommended method based on driver's feature |
US20200380629A1 (en) * | 2019-06-03 | 2020-12-03 | International Business Machines Corporation | Intelligent on-demand management of ride sharing in a transportation system |
CN112825182A (en) * | 2019-11-21 | 2021-05-21 | 北京京东尚科信息技术有限公司 | Method and device for determining recommended commodities |
CN112580945A (en) * | 2020-12-08 | 2021-03-30 | 吉首大学 | Dispatching method based on multiple correlation coefficients and vehicle dispatching optimization system |
CN112488794A (en) * | 2020-12-11 | 2021-03-12 | 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 |
CN112488806A (en) * | 2020-12-18 | 2021-03-12 | 深圳依时货拉拉科技有限公司 | Method and device for predicting order willingness, computer equipment and computer-readable storage medium |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114493405A (en) * | 2021-12-21 | 2022-05-13 | 重庆特斯联智慧科技股份有限公司 | Dispatching method and dispatching system based on movement track of logistics robot |
CN114493405B (en) * | 2021-12-21 | 2024-04-09 | 重庆特斯联智慧科技股份有限公司 | Commodity circulation robot motion trail-based dispatch method and system |
CN116029704A (en) * | 2023-03-27 | 2023-04-28 | 浙江虎哥数字科技有限公司 | Method, system and medium for recycling recyclable articles |
CN116433138A (en) * | 2023-06-13 | 2023-07-14 | 长沙争渡网络科技有限公司 | Logistics platform information pushing method and system based on genetic algorithm |
CN116433138B (en) * | 2023-06-13 | 2023-09-22 | 长沙争渡网络科技有限公司 | Logistics platform information pushing method and system based on genetic algorithm |
CN116681373A (en) * | 2023-08-02 | 2023-09-01 | 四川集鲜数智供应链科技有限公司 | Logistics supply chain management method |
CN116681373B (en) * | 2023-08-02 | 2023-10-20 | 四川集鲜数智供应链科技有限公司 | Logistics supply chain management method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113554387A (en) | Driver preference-based e-commerce logistics order allocation method, device, equipment and storage medium | |
CN106600036B (en) | Based on Android multiple spot express delivery allocator | |
Berbeglia et al. | Dynamic pickup and delivery problems | |
Cheung et al. | Dynamic routing model and solution methods for fleet management with mobile technologies | |
CN112183852B (en) | Logistics distribution route screening and freight accounting method, system, terminal and medium | |
US11288603B2 (en) | System for real-time optimal matching of ride sharing requests | |
Zurheide et al. | A revenue management slot allocation model for liner shipping networks | |
US20200160251A1 (en) | Adaptive Dispatching Engine For Advanced Taxi Management | |
US20210201214A1 (en) | System and method for recommending bidding bundle options in bidding-based ridesharing | |
Raviv et al. | Crowd-shipping of small parcels in a physical internet | |
CN109685411A (en) | A kind of share-car matching process of logistics shipping platform | |
Zhang et al. | Dynamic vehicle routing with random requests: A literature review | |
Correia et al. | Smart supply chain management: The 5W1H open and collaborative framework | |
Wang et al. | On optimizing shared-ride mobility services with walking legs | |
US10902379B2 (en) | System for customized unrequested item resolution | |
Li et al. | Emerging optimization problems for distribution in same-day delivery | |
Billhardt et al. | Coordinating open fleets. A taxi assignment example | |
CN112418550A (en) | Vehicle and goods matching method under various driver working states in logistics industry | |
CN113554386A (en) | Logistics distribution method, device, equipment and storage medium suitable for e-commerce orders | |
Hou et al. | Matching models for crowd-shipping considering shipper’s acceptance uncertainty | |
Megantara et al. | Mathematical Modeling on Integrated Vehicle Assignment and Rebalancing in Ride-hailing System with Uncertainty Using Fuzzy Linear Programming | |
Çakır et al. | Intuitionistic fuzzy multi-objective milk-run modelling under time window constraints | |
Gómez-Marín et al. | Fostering collaboration and coordination in urban delivery: a multi-agent microsimulation model | |
Verma et al. | A predictive taxi dispatching system for improved user satisfaction and taxi utilization | |
Sethuraman et al. | Amazon locker capacity management |
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 |