Order allocation method and device, storage medium and electronic equipment
The order allocation method addresses logistical inefficiencies by using non-Euclidean clustering and weighted bipartite graph matching to enhance task division and courier allocation, improving accuracy and adaptability in logistics operations.
Patent Information
- Application Number
- CN202310665739.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-06
- Publication Date
- 2024-12-06
AI Technical Summary
Existing logistics systems face challenges in accurately and timely scheduling delivery tasks due to manual reliance on manager experience, lack of consideration for multi-dimensional order information, and inability to adapt to diverse business scenarios, leading to unreasonable task division and courier allocation.
An order allocation method utilizing non-Euclidean distance-based clustering and weighted bipartite graph matching to identify business scenarios, calculate multi-dimensional order distances, and match tasks with delivery personnel based on comprehensive order and courier information.
Improves the accuracy and rationality of order clustering, enhances the diversity of task division, and incentivizes delivery personnel by considering historical service quality, resulting in more efficient and personalized logistics operations.
Smart Images

Figure CN119090159A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of logistics technology, and in particular to an order distribution method, an order distribution device, a storage medium and an electronic device. Background Art
[0002] As the "last mile" of the logistics system, the overall capacity of the distribution link directly affects customer loyalty to the company.
[0003] At present, the division of delivery areas and the matching of couriers with areas in actual scenarios mainly rely on manual designation or adjustment by managers such as station managers of the sales department, which is completely dependent on manual experience; or only considers single information such as the spatial location and quantity information of the order, making the task division and the binding between tasks and couriers unreasonable. With the continuous growth of delivery demand, it is no longer possible to meet the scheduling requirements of accuracy and timeliness.
[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention
[0005] The purpose of the present disclosure is to provide an order distribution method, an order distribution device, a storage medium and an electronic device, aiming to solve the problems of order task division and unreasonable matching between order tasks and delivery personnel during order delivery.
[0006] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.
[0007] According to one aspect of an embodiment of the present disclosure, there is provided an order allocation method, comprising: identifying a target business scenario corresponding to a site based on order information of the site, and extracting clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario; calculating a non-Euclidean distance between each order based on the order information and the clustering algorithm weight information, so as to obtain a clustering result by performing a clustering analysis based on the non-Euclidean distance; obtaining an order task unit composed of orders by performing division according to the task division method based on the clustering result; constructing a weighted bipartite graph based on the order task unit, the delivery personnel of the site and the bipartite graph weight information, and calculating a matching result between the order task unit and the delivery personnel based on the weighted bipartite graph.
[0008] According to some embodiments of the present disclosure, based on the aforementioned scheme, the clustering algorithm weight information includes distance weight values corresponding to each dimension of order features, and the non-Euclidean distance between each order is calculated according to the order information and the clustering algorithm weight information, including: obtaining feature data corresponding to each dimension of order features of each order according to the order information; wherein the order information includes one or more of the order's collection address, delivery address, time wave, product type, task type, channel source, and object characteristics; calculating order distance data corresponding to each dimension of order features between each order based on the feature data; and calculating the non-Euclidean distance according to the order distance data and the distance weight values corresponding to each dimension of order features.
[0009] According to some embodiments of the present disclosure, based on the aforementioned scheme, the clustering results are obtained by performing cluster analysis according to the non-Euclidean distance, including: obtaining a distance matrix according to the non-Euclidean distance between two orders; and tuning the hyperparameters in the clustering model based on the distance matrix to obtain the clustering results corresponding to the optimal hyperparameter values.
[0010] According to some embodiments of the present disclosure, based on the aforementioned scheme, the hyperparameters in the clustering model are tuned based on the distance matrix to obtain the optimal values of the hyperparameters and the clustering results corresponding to the optimal values of the hyperparameters, including: configuring an initial optimal silhouette coefficient, an initial list index value and a hyperparameter candidate list; wherein the hyperparameter candidate list includes a neighborhood radius and a minimum number of neighborhood samples; selecting a target hyperparameter from the hyperparameter candidate list according to the initial list index value, and inputting the target hyperparameter and the distance matrix into the clustering model to obtain the cluster center coordinates output by the clustering model; calculating the current silhouette coefficient based on the cluster center coordinates, and when the current silhouette coefficient is greater than the optimal silhouette coefficient, updating the initial optimal silhouette coefficient to the current silhouette coefficient, and recording the clustering results; updating the initial list index value, repeating the above clustering and judgment process until the hyperparameter candidate list is completely traversed to output the clustering results.
[0011] According to some embodiments of the present disclosure, based on the aforementioned scheme, a weighted bipartite graph is constructed according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, including: creating a left node according to the order task unit, and creating a right node according to the delivery personnel; generating an adjacency matrix based on the matching degree between the order task unit and the delivery personnel to create an edge between the left node and the right node to obtain an initial bipartite graph; traversing all elements in the adjacency matrix in turn, and when it is checked that there is no possible matching between the order task unit and the delivery personnel, pruning the initial bipartite graph to obtain a target bipartite graph; calculating the edge weight values corresponding to each edge in the target bipartite graph according to the bipartite graph weight information and the pre-configured weight values of the delivery personnel to obtain the weighted bipartite graph.
[0012] According to some embodiments of the present disclosure, based on the aforementioned scheme, the bipartite graph weight information includes feature weight values corresponding to order features of each dimension, and the edge weight values corresponding to each edge in the target bipartite graph are calculated according to the bipartite graph weight information and the pre-configured weight values of each delivery personnel, including: for an order task unit, obtaining feature data corresponding to order features of each dimension of the order task unit according to the order information; calculating the weight value of the order task unit based on the feature data and feature weight values corresponding to the order features of each dimension; traversing all the order task units to obtain the weight value of each order task unit; calculating the edge weight values corresponding to each edge in the target bipartite graph based on the weight value of each order task unit and the weight value of each delivery personnel.
[0013] According to some embodiments of the present disclosure, based on the aforementioned scheme, the method further includes: pre-configuring a weight value for each delivery personnel, the pre-configuring a weight value for each delivery personnel including: obtaining delivery information of each delivery personnel; wherein the delivery information includes basic information and / or service information; and configuring a weight value corresponding to each delivery personnel based on the delivery information.
[0014] According to a second aspect of an embodiment of the present disclosure, an order distribution device is provided, comprising: an identification module, for identifying a target business scenario corresponding to a site based on the order information of the site, and extracting clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario; a clustering module, for calculating the non-Euclidean metric distance between each order according to the order information and the clustering algorithm weight information, so as to perform clustering analysis according to the non-Euclidean metric distance to obtain a clustering result; a division module, for dividing the clustering result according to the task division method to obtain an order task unit composed of orders; a matching module, for constructing a weighted bipartite graph according to the order task unit, the delivery personnel of the site and the bipartite graph weight information, and calculating the matching result between the order task unit and the delivery personnel based on the weighted bipartite graph.
[0015] According to a third aspect of an embodiment of the present disclosure, there is provided a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the order allocation method in the above embodiment is implemented.
[0016] According to the fourth aspect of an embodiment of the present disclosure, an electronic device is provided, characterized in that it includes: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the order allocation method as in the above-mentioned embodiment.
[0017] The exemplary embodiments of the present disclosure may have some or all of the following beneficial effects:
[0018] In the technical solutions provided by some embodiments of the present disclosure, the corresponding target business scenario is first identified through the order information at the site to extract the clustering algorithm weight information, task division method and bipartite graph weight information corresponding to the target scenario; then, according to the extracted information, the non-Euclidean distance between the orders is used for clustering and then divided to obtain the order task unit; finally, the order task unit is matched with the delivery personnel, and then the order is delivered according to the matching result. Based on the above method, multiple sets of solutions are formulated for different business scenarios. On the one hand, different business scenarios correspond to different clustering algorithm weight information, and the distance between orders is measured by non-Euclidean distance, and then different degrees of weighted clustering are implemented for orders with different dimensional information, which improves the accuracy of order clustering and improves the rationality of subsequent order task unit division; on the other hand, different business scenarios correspond to different task division methods, and then multiple combinations can be used for division, which improves the diversity of order task unit division; on the other hand, different business scenarios correspond to different bipartite graph weight information for matching order task units and delivery personnel, which improves the rationality of order allocation.
[0019] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:
[0021] Figure 1 A schematic diagram schematically illustrates a flow chart of an order allocation method in an exemplary embodiment of the present disclosure;
[0022] Figure 2 A schematic diagram of a process of calculating a non-Euclidean distance in an exemplary embodiment of the present disclosure is schematically shown;
[0023] Figure 3 A schematic diagram schematically illustrates a flow chart of a cluster analysis method in an exemplary embodiment of the present disclosure;
[0024] Figure 4 A schematic diagram of a process of constructing a weighted bipartite graph in an exemplary embodiment of the present disclosure is schematically shown;
[0025] Figure 5 A schematic diagram schematically shows the composition of an order distribution device in an exemplary embodiment of the present disclosure;
[0026] Figure 6 A schematic diagram schematically illustrates a computer-readable storage medium in an exemplary embodiment of the present disclosure;
[0027] Figure 7 The structure diagram of a computer system of an electronic device in an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION
[0028] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more comprehensive and complete and will fully convey the concept of the example embodiments to those skilled in the art.
[0029] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the present disclosure.
[0030] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.
[0031] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.
[0032] As the "last mile" of the logistics system, the overall capacity of the distribution link directly affects customer loyalty to the company.
[0033] In the existing technology, the division of delivery areas in actual scenarios and the matching of couriers and areas mainly rely on manual designation or adjustment by managers such as station managers of the business department. This method completely relies on manual experience, resulting in the inability to meet the scheduling requirements of accuracy and timeliness.
[0034] At present, there are few related technical studies on the allocation and binding of couriers and road areas (grids). In the existing technology, the map corresponding to the delivery area of the delivery station is divided into multiple grids, and then the corresponding orders with addresses in each grid are counted. According to the geographical coordinates of the center point of each grid, the binding relationship between the courier and the grid is calculated to determine the binding relationship between the courier and the grid with the goal of minimizing the delivery distance of each courier and minimizing the difference in order quantities between couriers.
[0035] Since the algorithm uses the spatial geographic features of orders as input and takes the shortest distance and the smallest order quantity difference as the goal, the algorithm is related to spatial path planning, so the following problems still exist:
[0036] 1. Single decision input: Only the order space location and quantity information are considered. The singleness of the decision input information determines that the ceiling height of the algorithm's capabilities is insufficient.
[0037] 2. Unreasonable task division: When the delivery tasks are divided according to the order quantity and the center coordinates of the orders in the grid, the impact of other dimensions of the order information on the rationality of the division results is not considered; for example, there are two grids A and B, and the order quantity and center coordinate distance between the two grids are not much different, but the average weight or timeliness requirements of the orders in area A are significantly higher than those in area B. The existing technology will not identify the differences between the tasks corresponding to grids A and B. Therefore, from the perspective of business scenarios, such task division results cannot guarantee equality and are not reasonable.
[0038] 3. Single task division method: As mentioned above, the existing technology can only divide the tasks as evenly as possible based on the spatial location and order quantity information, but cannot achieve a variety of division methods; for example, give priority to other special types of orders such as those with high timeliness requirements or orders that have been delayed for an extended period; or evenly distribute overweight items, as well as other task division methods that are suitable for a variety of business scenarios.
[0039] 4. Unreasonable binding between tasks and couriers: When binding grids and courier tasks, on the one hand, the use of multi-dimensional order information is not taken into account, and on the other hand, accurate and personalized scheduling plans are not made based on the differences between different couriers. For example, couriers with physical preferences and younger age are given priority in assigning large tasks, or couriers with a good historical reputation are given priority in matching high-quality tasks, while couriers with a bad reputation are given priority in matching smaller tasks, etc.
[0040] 5. Lack of adaptability to business development: As business development becomes more diversified and services become more precise, the information dimension channels for order binding will increase in the future, and the order types and order channel sources will become more diversified. Current technology is unable to identify multi-dimensional needs and provide personalized delivery services.
[0041] 6. Lack of incentives for delivery services: When the existing technology assigns tasks, it does not take into account the courier’s historical service quality. All couriers have equal rights when receiving tasks, and the assignment results are random, so it does not play an incentive role.
[0042] Based on the many shortcomings of the prior art, the present disclosure provides an order allocation method that can comprehensively measure order information to reasonably and diversely divide tasks. The implementation details of the technical solution of the embodiment of the present disclosure are described in detail below.
[0043] Figure 1 The following is a schematic diagram showing a flow chart of an order allocation method in an exemplary embodiment of the present disclosure. Figure 1 As shown, the order allocation method includes steps S1 to S3:
[0044] Step S101, identifying a target business scenario corresponding to the site based on the order information of the site, and extracting clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario;
[0045] Step S102, calculating the non-Euclidean distance between two orders according to the order information and the clustering algorithm weight information, so as to perform cluster analysis according to the non-Euclidean distance to obtain a clustering result;
[0046] Step S103, dividing the tasks according to the task division method based on the clustering result to obtain order task units consisting of orders;
[0047] Step S104, constructing a weighted bipartite graph according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, and calculating the matching result between the order task unit and the delivery personnel based on the weighted bipartite graph.
[0048] In the technical solutions provided by some embodiments of the present disclosure, the corresponding target business scenario is first identified through the order information at the site to extract the clustering algorithm weight information, task division method and bipartite graph weight information corresponding to the target scenario; then, according to the extracted information, the non-Euclidean distance between the orders is used for clustering and then divided to obtain the order task unit; finally, the order task unit is matched with the delivery personnel, and then the order is delivered according to the matching result. Based on the above method, multiple sets of solutions are formulated for different business scenarios. On the one hand, different business scenarios correspond to different clustering algorithm weight information, and the distance between orders is measured by non-Euclidean distance, and then different degrees of weighted clustering are implemented for orders with different dimensional information, which improves the accuracy of order clustering and improves the rationality of subsequent order task unit division; on the other hand, different business scenarios correspond to different task division methods, and then multiple combinations can be used for division, which improves the diversity of order task unit division; on the other hand, different business scenarios correspond to different bipartite graph weight information for matching order task units and delivery personnel, which improves the rationality of order allocation.
[0049] Below, each step of the order allocation method in this example implementation will be described in more detail with reference to the accompanying drawings and embodiments.
[0050] In step S101, a target business scenario corresponding to the site is identified based on the order information of the site, and clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario are extracted.
[0051] Specifically, multiple business scenarios of the site are pre-configured, as well as the clustering algorithm weight information w corresponding to each business scenario. 1, task division method and bipartite graph weight information w 2 The clustering algorithm weight information is used for order clustering analysis in step S102, the task division method is used for order task unit division in step S103, and the bipartite graph weight information is used for construction of a weighted bipartite graph in step S104.
[0052] For example, an order waybill includes three order feature dimensions: the door-to-door service appointment timestamp t, geographic location s, and customer tag l.
[0053] If the task volume of the day's pending fulfillment tasks is moderate and there are a large number of special delivery orders such as VIP customers, expedited orders, and fresh food orders, and the customer's delivery experience needs to be guaranteed, then configure the business scenario to prioritize customer experience. In the scenario with priority to customer experience:
[0054] 1) Clustering algorithm weight information:
[0055] All orders are near the same coordinate point, but the timeliness requirements vary greatly. Some require delivery within 1 hour, such as stranded orders, while others can be delivered within a day. Therefore, the weight of the waybill reservation service time information is greater than the weight of other waybill information. The weight information is configured as
[0056] Among them, the first digit in the superscript of w represents the scenario, the second digit represents the clustering algorithm weight information or the bipartite graph weight information, and the subscript represents the order feature dimension, and the following are the same.
[0057] 2) Task division method:
[0058] The arrival order pool contains many orders with high timeliness requirements, so it is configured to adopt a biased partitioning method. The biased partitioning method means that each task unit only contains one or a few types of orders, ensuring that special waybills are assigned to couriers with high energy efficiency with a higher priority to ensure the timeliness of fulfillment.
[0059] 3) Bipartite graph weight information:
[0060] When calculating the task unit weight based on the waybill information, we should consider assigning a higher weight to the time dimension information of special waybill and ordinary waybill to ensure that special waybill or waybill that is about to exceed the time limit and default is assigned first. The weight information is
[0061] For example, if the order information is analyzed to find that most of the tasks to be fulfilled are ordinary waybills, and the collection and delivery efficiency needs to be maximized, then the target business scenario is identified as the collection and delivery efficiency priority scenario. In the collection and delivery efficiency priority scenario:
[0062] 1) Clustering algorithm weight information:
[0063] All orders must be delivered within one day, but the distribution of geographical coordinates varies greatly. Therefore, the weight of the waybill spatial distance information is greater than the weight of other waybill information. The weight information is configured as follows:
[0064] 2) Task division method:
[0065] The arrival order pool contains a certain number of overweight items. It is obviously inappropriate to divide them into one or a few task units. Therefore, an unbiased division method is configured. The unbiased division method means that each task unit evenly contains all types of orders, ensuring even distribution of tasks and avoiding waste of transportation capacity.
[0066] 3) Bipartite graph weight information:
[0067] When calculating the task unit weight based on the waybill information, a higher weight is assigned to the time dimension information and spatial distance information of the waybill. The weight information is:
[0068] It should be noted that the clustering algorithm weight information, task division method and bipartite graph weight information corresponding to the above-mentioned business scenarios are only exemplary. In actual application, the weight parameters corresponding to the scenarios can also be adjusted according to the actual operational results.
[0069] In one embodiment of the present disclosure, when the system background assigns the order pool to each operating site, the order information of each order in the order pool is obtained, and the order information is summarized and analyzed to select a suitable target business scenario. The order information includes multi-dimensional features, such as the order's collection address, delivery address, time wave, product type, task type, channel source, object characteristics, etc.
[0070] The specific implementation of obtaining order information can be achieved by regularly scheduling SQL (Structured Query Language) and Spark (computing engine) retrieval tasks offline in the data warehouse, and binding the obtained order information to each order through the ID. Summarize and analyze the order information, such as counting the number, mean, and historical year-on-year information of each information dimension of the order.
[0071] By analyzing the order information, the corresponding target business scenario is identified, and then the clustering algorithm weight information, task division method and bipartite graph weight information under the target business scenario are extracted.
[0072] Based on the above method, the multi-dimensional feature information of the order is considered for subsequent conversion into multi-channel structured data, and multiple sets of weighting schemes are formulated according to multiple business scenarios, and different degrees of weighting are implemented on information of different dimensions, so that the division results are most applicable to the business.
[0073] In step S102, the non-Euclidean distance between each order is calculated according to the order information and the clustering algorithm weight information, so as to perform cluster analysis according to the non-Euclidean distance to obtain a clustering result.
[0074] In one embodiment of the present disclosure, the system backend uses the non-Euclidean DBSCAN density clustering algorithm to cluster the orders in the site order pool, that is, the non-Euclidean distance is used when calculating the distance matrix between orders.
[0075] In the prior art, the physical distance between two orders is often considered, that is, the longitude and latitude coordinates of the pickup and delivery of the orders are considered. However, in the actual order delivery process, more attention is paid to the delivery time. The time difference of 1 hour and the distance difference of 1 km cannot be treated the same. Therefore, the present invention comprehensively considers the multi-dimensional characteristics of the order to calculate the non-Euclidean distance.
[0076] Figure 2 The flowchart of a method for calculating a non-Euclidean distance in an exemplary embodiment of the present disclosure is schematically shown. Figure 2 As shown, calculating the non-Euclidean distance in step S102 specifically includes the following steps:
[0077] Step S201, obtaining feature data corresponding to each dimension of order features of each order according to the order information;
[0078] Step S202, calculating order distance data corresponding to order features of each dimension between two orders based on the feature data;
[0079] Step S203, calculating the non-Euclidean distance according to the order distance data and the distance weight value corresponding to each dimension of the order feature.
[0080] Specifically, for example, order features may include three dimensions: the door-to-door service appointment timestamp t, the geographic location s, and the customer tag l. In step S201, feature data corresponding to each order in the order pool may be obtained based on the order information.
[0081] Then, in step S202, the distance between two orders needs to be determined. Taking order a and order b as an example, the order distance data corresponding to each dimension of the order features between the two orders are calculated, as shown in formula (1) to formula (3):
[0082] x t =t a -t b (1)
[0083]
[0084] x l =tl -t l (3)
[0085] Finally, in step S203, the clustering algorithm weight information includes the distance weight values corresponding to each dimension of the order feature. The order distance data corresponding to each dimension of the order feature and the distance weight value are multiplied and summed to obtain the non-Euclidean distance, as shown in formula (4):
[0086]
[0087] Among them, w j is the distance weight value corresponding to each dimension of order features, extracted according to the target business scenario; j are the order distance data corresponding to each dimension of order features; j represents the three dimensions of door-to-door service appointment timestamp t, geographic location s, and customer label l.
[0088] In one embodiment of the present disclosure, cluster analysis specifically includes the following steps:
[0089] Step (1), obtaining a distance matrix according to the non-Euclidean distance between two orders;
[0090] Step (2): optimizing the hyperparameters in the clustering model based on the distance matrix to obtain the clustering results corresponding to the optimal hyperparameter values.
[0091] Specifically, based on the non-Euclidean distance between orders, it is necessary to use the non-Euclidean DBSACN density clustering algorithm to tune the parameters in order to obtain reasonable clustering results corresponding to the optimal parameters.
[0092] Step (2) specifically includes: configuring an initial optimal silhouette coefficient, an initial list index value and a hyperparameter candidate list; wherein the hyperparameter candidate list includes a neighborhood radius and a minimum number of neighborhood samples; selecting a target hyperparameter from the hyperparameter candidate list according to the initial list index value, and inputting the target hyperparameter and the distance matrix into the clustering model to obtain the cluster center coordinates output by the clustering model; calculating the current silhouette coefficient based on the cluster center coordinates, and when the current silhouette coefficient is greater than the optimal silhouette coefficient, updating the initial optimal silhouette coefficient to the current silhouette coefficient, and recording the clustering result; updating the initial list index value, and repeating the above clustering and judgment process until the hyperparameter candidate list is completely traversed to output the clustering result.
[0093] Figure 3 A schematic diagram of a cluster analysis method in an exemplary embodiment of the present disclosure is shown schematically. Figure 3 The process of step (2) is explained in detail.
[0094] Step S301, first initialize the parameters of the clustering model, including the initial optimal silhouette coefficient S max , initial list index i, hyperparameter candidate list {Eps}, {MinPts}. The hyperparameter candidate list {Eps}, {MinPts} includes the initial values of the specified hyperparameters Eps (epsilon, neighborhood radius) and MinPts (min-points, minimum number of neighborhood samples).
[0095] Step S302, determine whether the hyperparameter candidate list is completely traversed, if so, the loop tuning ends and jumps directly to step S309, if not, execute step S303;
[0096] Step S303: According to the position indicated by the initial list index i, the corresponding hyperparameters in the hyperparameter candidate list are sequentially selected and input into the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering model. At the same time, the order distance matrix is input into the DBSCAN clustering model.
[0097] Step S304, the DBSCAN clustering model performs clustering calculations, completes the cluster center coordinates produced in the current round i, and records the clustering results. The DBSCAN clustering model is a density-based clustering algorithm. The main idea is to define the concept of core elements, that is, samples with a distance around the sample greater than a specified threshold and a number greater than N, and then solve all core elements that can be directly reached by density and all elements in their neighborhood to obtain a cluster, delete the clustered samples, repeat the previous step, and gradually obtain all clusters.
[0098] Step S305, calculating the corresponding silhouette coefficient S based on the clustering result i .
[0099] Step S306: i With S max Compare, if S i >S max , then jump to step S307, if S i ≤S max , then execute step S302 and continue the cycle;
[0100] Step S307, S i >S max When S max =S i , and record the parameters of the current round and the clustering results.
[0101] Step S308: the list index i is incremented by a step size, ie, i=i+1, and clustering is repeated.
[0102] Until it is determined in S302 that all parameters in the hyperparameter candidate list have been selected, that is, the hyperparameter candidate list is completely traversed, the optimal parameter adjustment result and the corresponding clustering result can be output.
[0103] In step S103, the task division method is used to divide the tasks into order task units based on the clustering result.
[0104] Specifically, the task division method can be obtained according to the target business scenario. In step S103, an order list corresponding to each class is obtained based on the clustering result, and then the orders are divided according to the corresponding task division method to generate multiple order task units.
[0105] Based on the above-mentioned steps S102 and 103, the diversity of task division can be achieved. On the one hand, the introduction of clustering algorithm weight information can make the order weight calculation results different, and the same division method can bring different results; on the other hand, the introduction of task division method can make it possible to group in a variety of combinations under the same weighting method when the sorting results of the order list are the same according to the weights, and still obtain different division results.
[0106] It should be noted that the algorithm part of the clustering algorithm is only one implementation of the present invention. Those skilled in the art should know that it can be replaced by a general clustering algorithm with better effect, but it must be used in conjunction with the non-Euclidean metric distance and clustering algorithm weight information.
[0107] In step S104, a weighted bipartite graph is constructed according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, and a matching result between the order task unit and the delivery personnel is calculated based on the weighted bipartite graph.
[0108] Figure 4 The following is a schematic diagram of a flow chart of a method for constructing a weighted bipartite graph in an exemplary embodiment of the present disclosure. Figure 4 As shown, constructing a weighted bipartite graph in step S104 specifically includes the following steps:
[0109] Step S401, creating a left node according to the order task unit, and creating a right node according to the delivery personnel;
[0110] Step S402, generating an adjacency matrix based on the matching degree between the order task unit and the delivery personnel to create an edge between the left node and the right node to obtain an initial bipartite graph;
[0111] Step S403, traversing all elements in the adjacency matrix in turn, and when it is found that there is no possible match between the order task unit and the delivery personnel, pruning the initial bipartite graph to obtain a target bipartite graph;
[0112] Step S404, calculating the edge weight values corresponding to each edge in the target bipartite graph according to the bipartite graph weight information and the pre-configured weight values of the delivery personnel, so as to obtain the weighted bipartite graph.
[0113] Specifically, assuming there are M order task units, M left nodes are created, and N delivery personnel are created, N right nodes are created, and then the left nodes and the right nodes are connected to obtain edges, thereby obtaining an initial bipartite graph consisting of nodes and edges.
[0114] Then generate the adjacency matrix Q[m×n], Q mn Represents the matching degree between the mth order task unit and the nth delivery personnel, m∈M, n∈N.
[0115] Then, all elements of the adjacency matrix are traversed in turn to check whether there is a basic possibility for a match between a certain order task unit and a certain delivery personnel. For example, if a certain delivery personnel is too far away from a certain task unit, or the task unit shift does not match the delivery personnel shift, the element will be set to zero, which plays the role of bipartite graph pruning.
[0116] Finally, the edge weight values corresponding to the edges in the pruned target bipartite graph are calculated. Specifically, the edge weight values calculated in step S404 include the following steps:
[0117] Step (1), for an order task unit, obtaining feature data corresponding to each dimension of order features of the order task unit according to the order information;
[0118] Step (2), calculating the weight value of the order task unit based on the feature data and feature weight values corresponding to each dimension of the order feature;
[0119] Step (3), traversing all the order task units to obtain the weight value of each order task unit;
[0120] Step (4), based on the weight value of each order task unit and the weight value of each delivery personnel, the edge weight value corresponding to each edge in the target bipartite graph is calculated.
[0121] Specifically, steps (1) to (3) calculate the weight value w of each order task unit m m The bipartite graph weight information includes the feature weight values corresponding to each dimension of the order feature. m See formula (5) for details:
[0122]
[0123] in, are the feature weight values corresponding to the order features of each dimension, t i 、s i , l i The feature data respectively correspond to the order features of each dimension of the order task unit, and the average value can be calculated based on the feature data of each order in the order task unit.
[0124] In addition, it is also necessary to obtain the weight value w of each delivery personnel n n In one embodiment of the present disclosure, the method further includes: pre-configuring a weight value of each delivery personnel, including: obtaining delivery information of each delivery personnel; wherein the delivery information includes basic information and / or service information; and configuring a weight value corresponding to each delivery personnel based on the delivery information.
[0125] Specifically, basic information refers to the static portrait of the delivery personnel, such as age, gender, education, physical strength and other information that is not easy to change; while service information includes information such as the delivery personnel's service quality, delivery rate, false delivery, customer complaint rate, etc. during the delivery process. The recording of service information requires a higher update frequency, and the implementation method is similar to order collection.
[0126] After obtaining the weight value w of each order task unit m m And the weight value w of each delivery person n n After that, calculate the edge weight value w mn , which can be w mn =w m +w n , or Then we get a weighted bipartite graph.
[0127] In this way, when matching tasks, on the one hand, the basic information of the delivery personnel is taken into consideration. For example, different weights will be assigned to full-time or part-time employees from a business perspective, and static portraits such as type and age will also be taken into consideration. On the other hand, the service information of the delivery personnel will also be taken into consideration, so that the historical service quality data of the delivery personnel will be converted into the weights used in task allocation. The higher the historical service quality, the higher the weight, and the greater the advantage in getting a good task when allocating tasks. The lower the historical service quality, the lower the weight, and the more difficult it is to get a task when it is allocated, which plays an incentive role in encouraging couriers to improve service quality as much as possible.
[0128] Finally, a suitable matching algorithm can be used based on the weighted bipartite graph to calculate the matching result, such as the Kuhn-Munkres algorithm. However, the bipartite graph matching algorithm in the present disclosure is not limited to the Kuhn-Munkres algorithm, and can be replaced by a general and more effective graph matching algorithm.
[0129] Among them, the matching result is the matching result between the order task unit and the delivery personnel. Based on the matching result, the corresponding order delivery task can be generated, and the delivery personnel completes the delivery of the order task unit according to the matching result.
[0130] Based on the above method, since the weight of the order task unit and the weight of the delivery personnel are taken into account in the matching process, a bipartite graph is generated and the graph matching algorithm is used to achieve the optimal matching of task packages and transportation capacity. The matching result is the most reasonable allocation result given after comprehensive consideration of multi-dimensional information.
[0131] It should be noted that if a site has new business scenarios or requirements, it only needs to design a new weighted template to act on the task division results, and the requirements can be met without changing the existing algorithms or strategies, which is scalable.
[0132] In summary, in the related technical research in the field of logistics terminal distribution, the order allocation method provided by the present disclosure proposes:
[0133] 1) Clustering algorithm based on non-Euclidean metric with business-oriented weighted input: Compared with traditional clustering algorithms, the scenario-oriented weighting scheme makes the input information more realistically reflect the business scenario, and the algorithm mechanism based on non-Euclidean metric makes the algorithm results more reasonable.
[0134] 2) Task division methods based on multi-dimensional spatiotemporal information: Innovatively proposed a variety of task division methods based on spatiotemporal information, which are highly novel and adaptable to complex business scenarios.
[0135] 3) Task allocation mechanism based on bipartite graph matching: Innovative application of graph matching algorithm; Reasonable introduction of courier portrait information into the allocation mechanism, which has good novelty.
[0136] 4) The capacity incentive effect of automatic task allocation: The task allocation result depends on the courier’s historical service quality. The better the historical evaluation is, the higher the service quality is, and the higher the weight is. When assigning tasks, there is a greater possibility of assigning high-quality tasks, and vice versa. After long-term operation, it will motivate couriers to continuously improve their service quality and actively enhance their service awareness.
[0137] Figure 5 The following is a schematic diagram showing the composition of an order distribution device in an exemplary embodiment of the present disclosure. Figure 5 As shown, the order distribution device 500 may include an identification module 501, a clustering module 502, a division module 504 and a matching module 505. Among them:
[0138] An identification module 501 is used to identify a target business scenario corresponding to the site based on the order information of the site, and extract clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario;
[0139] A clustering module 502, configured to calculate the non-Euclidean distance between two orders according to the order information and the clustering algorithm weight information, so as to perform cluster analysis according to the non-Euclidean distance to obtain a clustering result;
[0140] A division module 503, configured to divide the tasks into order task units consisting of orders based on the clustering result and in accordance with the task division method;
[0141] The matching module 504 is used to construct a weighted bipartite graph according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, and calculate the matching result between the order task unit and the delivery personnel based on the weighted bipartite graph.
[0142] According to an exemplary embodiment of the present disclosure, the clustering algorithm weight information includes distance weight values corresponding to each dimension of order features, and the clustering module 502 is used to obtain feature data corresponding to each dimension of order features of each order according to the order information; wherein the order information includes one or more of the order's collection address, delivery address, time wave, product type, task type, channel source, and object characteristics; based on the feature data, the order distance data corresponding to each dimension of order features between two orders are calculated; the non-Euclidean distance is calculated according to the order distance data corresponding to each dimension of order features and the distance weight values.
[0143] According to an exemplary embodiment of the present disclosure, the clustering module 502 is used to obtain a distance matrix based on the non-Euclidean distance between two orders; and to tune the hyperparameters in the clustering model based on the distance matrix to obtain the clustering results corresponding to the optimal hyperparameter values.
[0144] According to an exemplary embodiment of the present disclosure, the clustering module 502 is used to configure an initial optimal silhouette coefficient, an initial list index value and a hyperparameter candidate list; wherein the hyperparameter candidate list includes a neighborhood radius and a minimum number of neighborhood samples; a target hyperparameter is selected from the hyperparameter candidate list according to the initial list index value, and the target hyperparameter and the distance matrix are input into the clustering model to obtain the cluster center coordinates output by the clustering model; a current silhouette coefficient is calculated based on the cluster center coordinates, and when the current silhouette coefficient is greater than the optimal silhouette coefficient, the initial optimal silhouette coefficient is updated to the current silhouette coefficient, and the clustering result is recorded; the initial list index value is updated, and the above clustering and judgment process is repeated until the hyperparameter candidate list is completely traversed to output the clustering result.
[0145] According to an exemplary embodiment of the present disclosure, the matching module 504 is used to create a left node according to the order task unit and a right node according to the delivery personnel; generate an adjacency matrix based on the matching degree between the order task unit and the delivery personnel to create an edge between the left node and the right node to obtain an initial bipartite graph; traverse all elements in the adjacency matrix in turn, and when it is checked that there is no possible match between the order task unit and the delivery personnel, prune the initial bipartite graph to obtain a target bipartite graph; calculate the edge weight values corresponding to each edge in the target bipartite graph according to the bipartite graph weight information and the pre-configured weight values of the delivery personnel to obtain the weighted bipartite graph.
[0146] According to an exemplary embodiment of the present disclosure, the bipartite graph weight information includes feature weight values corresponding to each dimension of order features, and the matching module 504 is used to obtain, for an order task unit, feature data corresponding to each dimension of order features of the order task unit according to the order information; calculate the weight value of the order task unit based on the feature data and feature weight values corresponding to each dimension of order features; traverse all the order task units to obtain the weight value of each order task unit; calculate the edge weight value corresponding to each edge in the target bipartite graph based on the weight value of each order task unit and the weight value of each delivery personnel.
[0147] According to an exemplary embodiment of the present disclosure, the matching module 504 also includes a delivery personnel weight unit, which is used to obtain delivery information of each delivery personnel; wherein the delivery information includes basic information and / or service information; and the weight values corresponding to each delivery personnel are configured based on the delivery information.
[0148] The specific details of each module in the above-mentioned order allocation device 500 have been described in detail in the corresponding order allocation method, so they will not be repeated here.
[0149] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.
[0150] In an exemplary embodiment of the present disclosure, a storage medium capable of implementing the above method is also provided. Figure 6 A schematic diagram schematically illustrates a computer-readable storage medium in an exemplary embodiment of the present disclosure, such as Figure 6 As shown, a program product 600 for implementing the above method according to an embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a mobile phone. However, the program product of the present disclosure is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.
[0151] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided. Figure 7 The structure diagram of a computer system of an electronic device in an exemplary embodiment of the present disclosure is schematically shown.
[0152] It should be noted that Figure 7 The computer system 700 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.
[0153] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 702 or the program loaded from the storage part 708 to the random access memory (RAM) 703. In the RAM 703, various programs and data required for system operation are also stored. The CPU 701, the ROM 702 and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.
[0154] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed so that a computer program read therefrom is installed into the storage section 708 as needed.
[0155] In particular, according to an embodiment of the present disclosure, the process described below with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 709, and / or installed from a removable medium 711. When the computer program is executed by a central processing unit (CPU) 701, various functions defined in the system of the present disclosure are executed.
[0156] It should be noted that the computer-readable medium shown in the embodiment of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system, device or device or used in combination with it. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.
[0157] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0158] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, and the units described may also be arranged in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.
[0159] As another aspect, the present disclosure further provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the above embodiment.
[0160] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.
[0161] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation of the present disclosure.
[0162] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure.
[0163] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. An order allocation method, characterized in that: include: Identify the target business scenario corresponding to the site based on the order information of the site, and extract clustering algorithm weight information, task division method and bipartite graph weight information of the target business scenario; Calculating the non-Euclidean distance between two orders according to the order information and the clustering algorithm weight information, so as to perform cluster analysis according to the non-Euclidean distance to obtain a clustering result; Based on the clustering result, the order task units consisting of orders are obtained by dividing the tasks in the task division method; A weighted bipartite graph is constructed according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, and a matching result between the order task unit and the delivery personnel is calculated based on the weighted bipartite graph.
2. The order allocation method according to claim 1, characterized in that: The clustering algorithm weight information includes distance weight values corresponding to order features of each dimension, and the non-Euclidean distance between two orders is calculated according to the order information and the clustering algorithm weight information, including: Obtaining feature data corresponding to each dimension of order features of each order according to the order information; wherein the order information includes one or more of the order's collection address, delivery address, time wave, product type, task type, channel source, and object characteristics; Based on the feature data, calculate the order distance data corresponding to the order features of each dimension between two orders; The non-Euclidean distance is calculated based on the order distance data and the distance weight value corresponding to each dimension of the order feature.
3. The order allocation method according to claim 1, characterized in that: The clustering analysis is performed according to the non-Euclidean distance to obtain a clustering result, including: Obtaining a distance matrix according to the non-Euclidean distance between two orders; The hyperparameters in the clustering model are tuned based on the distance matrix to obtain the clustering results corresponding to the optimal hyperparameter values.
4. The order allocation method according to claim 3, characterized in that: The step of optimizing the hyperparameters in the clustering model based on the distance matrix to obtain the optimal values of the hyperparameters and the clustering results corresponding to the optimal values of the hyperparameters includes: Configure the initial optimal silhouette coefficient, the initial list index value and the hyperparameter candidate list; wherein the hyperparameter candidate list includes the neighborhood radius and the minimum number of neighborhood samples; Selecting a target hyperparameter from the hyperparameter candidate list according to the initial list index value, and inputting the target hyperparameter and the distance matrix into the clustering model to obtain cluster center coordinates output by the clustering model; Calculating a current silhouette coefficient based on the cluster center coordinates, and when the current silhouette coefficient is greater than the optimal silhouette coefficient, updating the initial optimal silhouette coefficient to the current silhouette coefficient, and recording a clustering result; The initial list index value is updated, and the above clustering and judgment process is repeated until the hyperparameter candidate list is completely traversed to output the clustering result.
5. The order allocation method according to claim 1, characterized in that: The step of constructing a weighted bipartite graph according to the order task unit, the delivery personnel of the site and the bipartite graph weight information includes: Creating a left node according to the order task unit, and creating a right node according to the delivery personnel; Generate an adjacency matrix based on the matching degree between the order task unit and the delivery personnel to create an edge between the left node and the right node to obtain an initial bipartite graph; Traversing all elements of the adjacency matrix in sequence, and when it is detected that there is no possible match between the order task unit and the delivery personnel, pruning the initial bipartite graph to obtain a target bipartite graph; The edge weight values corresponding to the edges in the target bipartite graph are calculated according to the bipartite graph weight information and the pre-configured weight values of the delivery personnel to obtain the weighted bipartite graph.
6. The order distribution method according to claim 5, characterized in that: The bipartite graph weight information includes feature weight values corresponding to order features of each dimension, and the edge weight values corresponding to each edge in the target bipartite graph are calculated according to the bipartite graph weight information and the pre-configured weight values of each delivery personnel, including: For an order task unit, obtaining feature data corresponding to each dimension of order features of the order task unit according to the order information; Calculate the weight value of the order task unit based on the feature data and feature weight values corresponding to each dimension of the order feature; Traversing all the order task units to obtain a weight value of each order task unit; The edge weight values corresponding to the edges in the target bipartite graph are calculated based on the weight values of the order task units and the weight values of the delivery personnel.
7. The order distribution method according to claim 5, characterized in that: The method further includes: pre-configuring a weight value of each delivery personnel, wherein the pre-configuring a weight value of each delivery personnel includes: Obtaining the delivery information of each delivery person; wherein the delivery information includes basic information and / or service information; The weight values corresponding to the respective delivery personnel are configured based on the delivery information.
8. An order distribution device, characterized in that: include: An identification module, used to identify a target business scenario corresponding to the site based on the site's order information, and extract clustering algorithm weight information, task division method, and bipartite graph weight information of the target business scenario; A clustering module, used for calculating the non-Euclidean distance between two orders according to the order information and the clustering algorithm weight information, so as to perform cluster analysis according to the non-Euclidean distance to obtain a clustering result; A division module, used for dividing the tasks according to the task division method based on the clustering result to obtain order task units composed of orders; A matching module is used to construct a weighted bipartite graph according to the order task unit, the delivery personnel of the site and the weight information of the bipartite graph, and calculate the matching result between the order task unit and the delivery personnel based on the weighted bipartite graph.
9. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the order allocation method according to any one of claims 1 to 7 is implemented.
10. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the order allocation method according to any one of claims 1 to 7.
