A single-quantity balancing method for dispatching in a crowdsourced instant logistics scenario

Abstract

The application discloses a single-quantity balancing dispatching method in a crowd-sourced instant logistics scene. When a new order is generated, a nearby delivery personnel set is first screened according to the spatial position of the order, then the delivery cost of each delivery personnel in the set for delivering the order is calculated through path inference of the new order and the delivery personnel set, then K delivery personnel with the minimum delivery cost are selected for pre-dispatching, the delivery balancing score of the K delivery personnel is calculated and sorted, the first order in the pre-dispatching list of the delivery personnel with the highest score is selected for dispatching decision, and the pre-dispatching list of other delivery personnel including the order is updated until the distribution of all new orders is completed. The method guarantees the system efficiency and service quality to a certain extent, alleviates the contradiction between the single quantity and the transport capacity resources, realizes the stability and continuity of the order quantity of different delivery personnel individuals and different time periods, and is beneficial to the improvement of the entire platform ecology.

Description

A method for balanced order dispatch in a crowdsourced instant logistics scenario Technical Field

[0001] This invention belongs to the field of logistics, specifically relating to a method for balanced order dispatch in a crowdsourced instant logistics scenario. Background Technology

[0002] Instant logistics refers to a logistics model where service providers quickly respond to and fulfill users' unpredictable logistics needs. It typically involves point-to-point delivery without warehousing or transshipment. Compared to traditional logistics, instant logistics has three main characteristics: immediacy, regional focus, and point-to-point delivery. With Alibaba's introduction of the New Retail concept in 2016 and the development of users' local shopping habits, the demand for instant logistics has steadily increased, with food delivery and New Retail being the most typical examples. Unlike traditional express delivery, which involves "collection-transfer-transportation-delivery," instant logistics generally involves dedicated personnel and relies on crowdsourcing to improve efficiency.

[0003] Order dispatch is the most critical link in the entire logistics and delivery process. CN114819819A proposes a method for calculating delivery costs. To reduce delivery costs or improve system efficiency, most platform companies adopt a solution of maximizing system efficiency or minimizing delivery costs in order dispatch. However, this approach can easily lead to an uneven distribution of orders among riders in crowdsourcing scenarios. Unlike full-time riders, crowdsourced riders can choose not to accept orders. This uneven distribution of orders results in a poor rider experience in the short term and affects rider retention in the long term, ultimately impacting the overall health of the platform's ecosystem. Order dispatch balance is essentially a matter of fairness and stability in order dispatch. It is reflected not only in the equality of the number of orders dispatched among different riders but also in the stability and continuity of the number of orders dispatched to an individual rider over a given time. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method for balanced order dispatch in a crowdsourced instant logistics scenario.

[0005] The technical solution adopted in this invention is: a method for balanced order dispatch in a crowdsourced instant logistics scenario, the specific steps of which are as follows:

[0006] S1. Obtain a new order collection, order information, and receive real-time location information uploaded by delivery personnel;

[0007] S2. Iterate through the unassigned order set in chronological order of order placement time. Using the order's pickup address or user address as the center, select delivery personnel whose real-time geographical location is within a pre-set radius, while filtering out delivery personnel whose number of orders has reached the limit.

[0008] S3. Calculate the delivery cost for delivery personnel to deliver new orders and execute the order pre-dispatch operation;

[0009] S4. Calculate the balance score of the delivery personnel, select the first order in the pre-dispatch list of the delivery personnel with the highest balance score, perform the actual dispatch operation, and delete the order from the pre-dispatch list of other delivery personnel.

[0010] S5. Repeat step S4 until the set of delivery personnel that is not empty in the pre-dispatch list is empty, and complete the assignment of all new orders.

[0011] The beneficial effects of this invention are as follows: When a new order is generated, the method of this invention first filters out a set of nearby delivery personnel based on the spatial location of the order. Then, it calculates the delivery cost for each delivery person by performing path inference between the new order and each delivery person in the set. Next, it selects the K delivery personnel with the lowest delivery costs for pre-assignment, calculates and sorts the delivery balance score of these K delivery personnel, and selects the first order in the pre-assignment list of the rider with the highest score to execute the assignment decision. At the same time, it updates the pre-assignment lists of other delivery personnel including that order until all new orders are assigned. This method of the present invention, to a certain extent, ensures system efficiency and service quality, alleviates the contradiction between order volume and transportation resources, and achieves stability and continuity in the order volume of delivery personnel at different times and for different individuals, which is conducive to improving the entire platform ecosystem. Attached Figure Description

[0012] Figure 1 is a flowchart of a crowdsourced instant logistics scenario order balancing method according to the present invention. Detailed Implementation

[0013] To facilitate understanding of the technical content of the present invention by those skilled in the art, the method of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0014] This invention provides a method for balanced order dispatch in a crowdsourced instant logistics scenario. It employs a two-stage dispatch method with two objectives: First, during order and rider matching, the K riders with the lowest additional delivery costs are pre-dispatched, considering system efficiency and service quality goals. Second, riders are sorted by their dispatch balance score, and the rider with the highest balance score is selected for dispatch decision-making, considering the rider's order volume balance goal. This, to a certain extent, ensures the work enthusiasm of crowdsourced riders, especially newcomers, and in the long run, is beneficial to the healthy and stable development of the platform ecosystem. The main process modules include: information receiving module, rider selection module, pre-dispatching module, and dispatch decision module.

[0015] As shown in Figure 1, the flowchart of a crowdsourced instant logistics scenario order balancing dispatch method of the present invention is mainly divided into four layers:

[0016] The highest layer is the information receiving layer, which mainly receives new order information from the upstream system and the rider's real-time geographical location uploaded from the rider's APP. At the same time, it receives real-time delivery order information for each rider from the internal system of the server.

[0017] The second layer is the rider layer. This layer mainly iterates through the unassigned order set in chronological order of order placement time, selects riders whose real-time geographical location is within a pre-set radius using the order's pickup address or user address as the center, and filters out riders whose number of orders has reached the limit.

[0018] The third layer is the pre-dispatch layer, which mainly calculates the rider's delivery cost and executes the pre-dispatch operation to improve efficiency and service quality.

[0019] The fourth layer is the order dispatch decision layer. This layer mainly calculates the rider's balanced score, sorts riders by their balanced score, selects the rider with the highest balanced score to execute the order dispatch decision, and continues until all new orders are allocated, thereby improving the rider's order volume balance target.

[0020] Instant food delivery is an important scenario for instant logistics. For ease of explanation, the following will use food delivery services as an example, and the relevant definitions are as follows:

[0021] First, define the set of new pending orders generated within a certain period as O, denoted as o1o2...o n The set of riders filtered according to the pending orders is U. o ={u1u2...u m Let the set of rider u's working time periods be}.

[0022] Where n represents the number of new orders to be dispatched, and m represents the number of riders after filtering the orders to be dispatched. This indicates the rider's start and end times. This represents the rider's off-duty time. If the rider is currently on duty, the off-duty time is taken as the current time (now()). The actual number of dispatched orders is recorded as alreadyDispatchNum. u The target dispatch volume is goalDispatchNum u .

[0023] When rider u performs route planning for delivery order o, the method for calculating delivery cost is defined as follows:

[0024] Delivery cost

[0025] The crowdsourced rider order distribution method based on system efficiency of the present invention uses a server as the execution entity, and the specific steps are as follows:

[0026] Step 1: The server receives a new order set O. The order information includes the order placement time, merchant pickup address, user delivery address, platform-required delivery time, and delivery items. At the same time, the server receives the real-time location information uploaded by all riders.

[0027] Step 2: Iterate through the unallocated order set O in chronological order of order placement time, starting with order o. i Using the pickup address or user address as the center, select riders whose real-time geographic location is within a radius of p kilometers, while filtering out riders whose number of orders has reached the limit. Record the set of riders after filtering by order ox.

[0028] Where p represents a pre-set parameter, which can take the value 3.

[0029] Step 3: Execute the order. i The pre-dispatch operation is as follows:

[0030] Step 3.1: Traverse the rider set For each rider u and order o i Path inference is performed to calculate the rider's delivery route for the order. i Delivery costs Select the K riders with the lowest delivery costs for pre-assignment, and simultaneously allocate the orders to them. i Insert into the pre-dispatch list of the K riders, assuming the pre-dispatch list of rider u after execution.

[0031] Where K represents a pre-set parameter, which can take the value 3.

[0032] Step 3.2: Remove order o from the unassigned order set O. i Then repeat step 2 until the pre-dispatch operation of all orders in order set O is completed.

[0033] Step 4: Make the actual order dispatch decision for the set U of all riders whose pre-dispatch lists are not empty, as follows:

[0034] Step 4.1: Traverse the rider set U and calculate the balanced score for each rider u, as follows:

[0035] Step 4.1.1: First, we need to calculate the target dispatch volume for rider u as goalDispatchNum. u ;

[0036] The target order volume is a standard for measuring whether the current order volume of each rider is reasonable and for giving different levels of order allocation preference to riders with different order volumes. It is the average order volume that a rider should receive after a period of work, and it is also related to the order volume density during that period.

[0037] Based on the rider's average delivery volume over different time periods, we obtain the coordinates where the X-axis represents the time period and the Y-axis represents the average delivery volume within the corresponding time period. Then, we fit a polynomial function to obtain an order density function f(t) over time. Finally, using the definite integral method, the formula for calculating the target delivery volume for rider u is as follows:

[0038]

[0039] in, This indicates the rider's work hours.

[0040] Step 4.1.2: Calculate rider u's equal score;

[0041] Balanced Score u It measures the imbalance between the rider's actual order dispatch volume and the target order dispatch volume, and its calculation formula is as follows:

[0042]

[0043] Where μ represents a pre-set parameter and l represents a smoothing factor.

[0044] Step 4.2: Select the first order from the pre-dispatch list of the rider with the highest balanced score. Perform the actual order dispatch operation, and simultaneously process the order. Delete it from other riders' pre-assignment lists;

[0045] Step 4.3: Repeat step 4.1 until the set U of riders whose pre-dispatch list is not empty is empty.

[0046] The rationality of the order dispatch method definition. Consider the following initial scenario:

[0047] Suppose the server receives a new order o. Using the pickup or delivery address of order o as the center, select riders whose real-time geographical location is within a radius of p kilometers. Assume there are 4 riders u1, u2, u3, and u4 who meet the order acceptance criteria. Calculate the delivery cost for each rider delivering order o. Select the three riders with the lowest delivery costs for pre-dispatch orders. Let riders u1, u2, and u3 have the lowest delivery costs, then their pre-dispatch order lists are mayDispatchOrderList and mayDispatchOrderList, respectively. u1={o}、mayDispatchOrderList u2 ={o}、mayDispatchOrderList u3 ={o}, and then calculate the rider's balanced score according to step 4.1, denoted as Score. u1 Score u2 Score u3 Let Score be one of the scores. u3 If the rider has the highest score, select rider u3 with the highest balanced score, assign order o from their pre-dispatch list to rider u3 for delivery, and remove order o from the pre-dispatch lists of other riders, resulting in mayDispatchOrderList. u1 ={}、mayDispatchOrderList u2 ={}、mayDispatchOrderList u3 ={} means that all pre-dispatch lists are empty, and the dispatching process ends.

[0048] In summary, in the context of instant logistics order dispatch, the method of this invention not only considers system efficiency and service quality goals, namely the delivery cost for riders to accept new orders (reflected by the increased distance and increased timeout time), but also, to a certain extent, tilts order dispatch towards riders with a large difference between the target order volume and the actual dispatch volume, thereby achieving order volume balance. This not only ensures system efficiency and alleviates the contradiction between order volume and transportation resources, but also benefits the retention of crowdsourced riders, especially new ones, thus improving the entire platform ecosystem.

[0049] The above description of the embodiments is provided to enable those skilled in the art to understand and apply the present invention. It will be apparent to those skilled in the art that various modifications can be made to the above embodiments, and the general principles described herein can be applied to other embodiments without inventive effort. Therefore, the present invention is not limited to the above embodiments, and any improvements and modifications made to the present invention by those skilled in the art based on the disclosure thereof should be within the scope of protection of the present invention.

Claims

1. A method for balanced order dispatch in a crowdsourced instant logistics scenario, comprising the following steps: S1. Obtaining a new order set, order information, and receiving real-time geolocation information uploaded by delivery personnel; S2. Iterating through the unassigned order set in chronological order of order placement time, selecting delivery personnel whose real-time geolocation is within a pre-defined radius centered on the order's pickup address or user address, while filtering out delivery personnel whose order count has reached the limit; S3. Calculating the delivery cost for delivery personnel to deliver new orders and performing order pre-dispatch operations; Traversing the collection of delivery personnel For each of the delivery personnel With orders Perform route inference to calculate delivery personnel Deliver this order Delivery costs Choose the one with the lowest delivery cost. Delivery personnel pre-dispatch orders and simultaneously... Insert into From the list of pre-dispatch orders for delivery personnel, obtain the delivery personnel after execution. Pre-dispatch list ;in, This refers to pre-set parameters, such as delivery cost. The calculation method is as follows: = Delivery personnel's newly added route distance + newly added timeout time at each route node * node timeout penalty cost; remove unassigned order set. Orders in Then repeat the above steps until the order collection is complete. S4. Perform pre-dispatch operations on all orders; calculate the balance score of delivery personnel, select the first order in the pre-dispatch list of the delivery personnel with the highest balance score for actual dispatch operation, and delete the order from the pre-dispatch lists of other delivery personnel; based on the average number of delivery orders per delivery personnel in different historical time periods, obtain the coordinates of the X-axis (time period) and Y-axis (average number of delivery orders in the corresponding time period), and fit a polynomial function to obtain an order density function with respect to time. Calculate delivery personnel using definite integral method Target order volume The calculation formula is as follows: ;in, Indicates delivery personnel The working hours of the delivery personnel; The formula for calculating the balanced score is as follows: ;in, This indicates pre-set parameters. Represents the smoothing factor. S5. Repeat step S4 until the set of delivery personnel in the pre-dispatch list that is not empty is empty, and complete the allocation of all new orders.

Images