A method, system, and apparatus for dynamically assessing delivery capacity of a delivery service provider

Abstract

The application provides a method, system and device for dynamically evaluating the delivery capacity of a delivery service provider, the method comprising: acquiring external data and delivery data of the delivery service provider in real time, determining multi-dimensional feature data and a historical performance score of the delivery service provider corresponding to a preset prediction time window according to the cumulative external data and the cumulative delivery data; converting the multi-dimensional feature data into a feature vector and inputting the feature vector into a transport capacity prediction model to obtain a transport capacity prediction result corresponding to the delivery service provider, and determining a transport capacity score corresponding to the delivery service provider according to the transport capacity prediction result; and determining a delivery capacity score corresponding to the delivery service provider according to the historical performance score and the transport capacity score. The application can dynamically evaluate the real-time delivery capacity of the delivery service provider, so that the delivery scheduling can be performed according to the real-time delivery capacity of different delivery service providers, the delivery success rate is improved, the delivery demand of users can be better met, the user experience is improved, and the delivery service quality of the delivery service provider is improved.

Description

Technical Field

[0001] This application relates to the field of computer information processing technology, and in particular to a technology for dynamically evaluating the delivery capabilities of delivery service providers. Background Technology

[0002] In the field of digital operations for instant retail, instant retail service aggregation platforms typically connect with multiple delivery service providers, and existing order allocation strategies usually employ static weight allocation. With the rapid development of the instant delivery market and intensified competition among various delivery platforms, the delivery capabilities of each platform vary across different times and regions. Meanwhile, consumers' delivery demands and requirements for delivery service quality are increasing, rendering the traditional order scheduling method based on static weight allocation inadequate.

[0003] Therefore, determining the real-time delivery capabilities of each delivery service provider in order to accurately, efficiently, and flexibly schedule order deliveries to meet dynamic order delivery needs is a technical problem that urgently needs to be solved. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method, system, and device for dynamically evaluating the delivery capabilities of delivery service providers.

[0005] According to one aspect of this application, a method for dynamically assessing the delivery capabilities of a delivery service provider is provided, wherein the method includes:

[0006] Real-time acquisition of external data affecting order delivery and delivery data of delivery service providers, and determination of the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the cumulative external data and cumulative delivery data up to the present.

[0007] The multi-dimensional feature data is converted into feature vectors and input into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window. Based on the capacity prediction result, the capacity score of the delivery service provider corresponding to the preset prediction time window is determined.

[0008] Based on the historical performance score and the capacity score, the delivery capacity score of the delivery service provider corresponding to the preset forecast time window is determined.

[0009] Optionally, the multi-dimensional feature data includes at least: order acceptance rate, average response time, current load rate, regional coverage, time characteristics, weather information, and traffic information.

[0010] Optionally, determining the historical performance score of the delivery service provider corresponding to a preset prediction time window includes:

[0011] Based on the accumulated delivery data, the delivery success rate score, delivery time score, and user satisfaction score of the delivery service provider corresponding to the preset prediction time window are determined. The delivery success rate score, delivery time score, and user satisfaction score are calculated based on the first preset weight to obtain the historical performance score of the delivery service provider corresponding to the preset prediction time window.

[0012] Optionally, determining the historical performance score of the delivery service provider corresponding to a preset prediction time window further includes:

[0013] Based on the accumulated delivery data, the cost-effectiveness score and reliability score of the delivery service provider corresponding to the preset forecast time window are determined, wherein...

[0014] Based on a first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the historical performance score of the delivery service provider corresponding to a preset prediction time window, including:

[0015] Based on the first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the performance score of the delivery service provider corresponding to the preset prediction time window. The performance score, cost-effectiveness score, and reliability score are then used as the historical performance score of the delivery service provider corresponding to the preset prediction time window.

[0016] Optionally, the feature vector includes at least the following elements: delivery service provider identifier, order acceptance rate, average response time, current load rate, regional coverage, time period performance, weather impact factor, and traffic impact factor.

[0017] Optionally, the construction of the capacity prediction model includes:

[0018] Acquire historical delivery data from multiple delivery service providers and corresponding historical external data affecting delivery. Based on the historical delivery data, the historical external data, and the preset prediction time window, determine several historical multi-dimensional feature data and actual transportation capacity scores corresponding to the preset prediction time window.

[0019] Each historical multi-dimensional feature data is transformed into a feature vector, and its corresponding actual capacity score is used as its true value as a sample. This process is repeated for each historical multi-dimensional feature data and its actual capacity score to construct a sample set.

[0020] The sample set is divided into a training set, a test set, and a validation set. An LSTM neural network is trained to obtain a capacity prediction model.

[0021] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0022] For multiple delivery service providers, obtain the capacity forecast result for each delivery service provider corresponding to the preset forecast time window, and predict the distribution of delivery demand corresponding to the preset forecast time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers.

[0023] Based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window and the distribution of delivery demand, identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window, and adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window according to the identification results and preset adjustment strategies.

[0024] Optionally, the step of predicting the distribution of delivery demand corresponding to the preset prediction time window based on the acquired cumulative external data and the cumulative delivery data of all delivery service providers includes:

[0025] Based on the accumulated delivery data of all delivery service providers, a time series forecasting model is used to predict the basic distribution of delivery demand corresponding to the preset forecasting time window.

[0026] Based on the acquired cumulative external data, determine the external influencing factors corresponding to the preset prediction time window;

[0027] Based on the basic distribution of delivery demand and the external influencing factors, determine the distribution of delivery demand corresponding to the preset prediction time window.

[0028] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0029] Obtain the order information of the orders to be delivered corresponding to the preset prediction time window, and determine the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information;

[0030] Based on the order characteristics, adjust the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area, and determine the delivery service provider to execute the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

[0031] Optionally, before adjusting the delivery capacity score of each delivery service provider corresponding to the delivery time period and delivery area based on the order characteristics, the method for dynamically evaluating the delivery capacity of delivery service providers further includes:

[0032] Based on the cumulative delivery data of each delivery service provider, the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of the delivery service provider for the corresponding delivery time period and delivery area are determined. The regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score are calculated based on a second preset weight to obtain the regional matching score of the delivery service provider for the corresponding delivery time period and delivery area. Based on the regional matching score, the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area is adjusted.

[0033] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0034] Based on the execution results of the delivery task, update the cumulative delivery data of the corresponding delivery service provider.

[0035] According to another aspect of this application, a system for dynamically assessing the delivery capabilities of a delivery service provider is provided, wherein the system includes:

[0036] The first module is used to acquire external data affecting order delivery and delivery data of delivery service providers in real time, and determine the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the cumulative external data and cumulative delivery data up to the present.

[0037] The second module is used to convert the multi-dimensional feature data into feature vectors and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and to determine the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result.

[0038] The third module is used to determine the delivery capacity score of the delivery service provider corresponding to the preset prediction time window based on the historical performance score and the capacity score.

[0039] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0040] The fourth module is used to obtain the capacity prediction results of each delivery service provider corresponding to the preset prediction time window for multiple delivery service providers, and to predict the distribution of delivery demand corresponding to the preset prediction time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers.

[0041] The fifth module is used to identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window and the distribution of delivery demand, and to adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window based on the identification results and preset adjustment strategies.

[0042] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0043] The sixth module is used to obtain the order information of the orders to be delivered corresponding to the preset prediction time window, and to determine the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information.

[0044] The eighth module is used to adjust the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area according to the order characteristics, and to determine the delivery service provider to perform the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

[0045] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0046] The seventh module is used to determine the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of each delivery service provider based on the cumulative delivery data of each delivery service provider for the delivery time period and delivery area. Based on a second preset weight, the module calculates the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score to obtain the regional matching score of the delivery service provider for the delivery time period and delivery area. Based on the regional matching score, the module adjusts the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area.

[0047] According to another aspect of this application, a computer-readable medium is provided, wherein computer-readable instructions are stored on the medium, which are executed by a processor to implement part or all of any of the above methods.

[0048] According to another aspect of this application, an apparatus for dynamically assessing the delivery capabilities of a delivery service provider is provided, wherein the apparatus includes:

[0049] One or more processors; and a memory storing computer-readable instructions that, when executed, cause the processor to perform some or all of the operations described above.

[0050] Compared with existing technologies, this application provides a method, system, and device for dynamically evaluating the delivery capacity of a delivery service provider. The method includes: A) acquiring external data affecting order delivery and delivery data of the delivery service provider in real time, and determining the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to a preset prediction time window based on the accumulated external data and accumulated delivery data up to the present; B) converting the multi-dimensional feature data into feature vectors and inputting them into a capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and determining the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result; C) determining the delivery capacity score of the delivery service provider corresponding to the preset prediction time window based on the historical performance score and the capacity score. This application determines the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the acquired accumulated data up to the present, inputs it into a capacity prediction model, obtains the capacity prediction result of the delivery service provider, and thus determines the capacity score corresponding to the prediction time window. It can dynamically assess the real-time delivery capabilities of delivery service providers, so that when scheduling order delivery, delivery can be scheduled according to the real-time delivery capabilities of different delivery service providers, which improves the delivery success rate, better meets user delivery needs, improves user experience, and improves the delivery service quality of delivery service providers. Attached Figure Description

[0051] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0052] Figure 1 A schematic diagram illustrating a method for dynamically assessing the delivery capabilities of a delivery service provider according to one aspect of this application is shown.

[0053] Figure 2 A schematic diagram of an apparatus for dynamically assessing the delivery capabilities of a delivery service provider, according to another aspect of this application, is shown.

[0054] The same or similar reference numerals in the accompanying drawings represent the same or similar parts. Detailed Implementation

[0055] The present application will now be described in further detail with reference to the accompanying drawings.

[0056] In a typical configuration of various embodiments of this application, the method execution entity, each trusted party of the system, and / or each module of the device may include one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0057] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0058] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include non-transitory computer-readable media, such as modulated data signals and carrier waves.

[0059] To further illustrate the technical means adopted and the effects achieved in this application, the technical solution of this application will be clearly and completely described below in conjunction with the accompanying drawings and preferred embodiments.

[0060] Figure 1 The diagram illustrates a method for dynamically assessing the delivery capabilities of a delivery service provider according to one aspect of this application, wherein one embodiment of the method includes:

[0061] S101 acquires external data affecting order delivery and delivery data of delivery service providers in real time, and determines the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the cumulative external data and cumulative delivery data up to the present.

[0062] S102 converts the multi-dimensional feature data into feature vectors and inputs them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and determines the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result.

[0063] S103 determines the delivery capacity score of the delivery service provider corresponding to the preset prediction time window based on the historical performance score and the capacity score.

[0064] In the application scenario of this application, the aggregation platform provided by the digital operation service provider for the catering and retail industry can provide digital operation services to numerous merchants and their stores connected to different instant retail service platforms, and connect with several delivery service providers. Merchants' orders are allocated to a single delivery service provider through delivery scheduling, and the delivery service provider completes the order delivery. Traditional delivery scheduling typically involves pre-assessing the delivery capacity of each delivery service provider and assigning a fixed weight to each provider based on the pre-assessed capacity. However, when scheduling order delivery, the allocation of these fixed weights to each delivery service provider suffers from problems such as low matching accuracy, low efficiency, and consequently, a poor user experience.

[0065] This application provides a technical solution for dynamically evaluating the delivery capabilities of delivery service providers. It can dynamically assess the real-time delivery capabilities of delivery service providers, and during order delivery scheduling, dynamically determine weights based on the real-time delivery capabilities of different delivery service providers. This can better meet user delivery needs, improve user experience, and enhance the delivery service quality of delivery service providers. This application is implemented through an aggregation platform 100 provided by a digital operation service provider for catering and retail. The aggregation platform includes computer equipment and / or a cloud with the necessary hardware and software environment deployed. The computer equipment includes, but is not limited to, personal computers, laptops, industrial computers, servers, network hosts, single network servers, or network server clusters. The cloud consists of a large number of computers or network servers based on cloud computing, where cloud computing is a type of distributed computing, consisting of a virtual supercomputer composed of a group of loosely coupled computers. Here, the computer equipment and / or cloud are merely examples; other existing or future equipment and / or resource sharing platforms applicable to this application should also be included within the scope of protection of this application, and are hereby incorporated by reference.

[0066] In this embodiment, in step S101, the aggregation platform 100 can acquire external data affecting order delivery and delivery data of delivery service providers in real time, and determine the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the accumulated external data and accumulated delivery data up to the present.

[0067] Factors such as weather and traffic can affect order delivery. The aggregation platform 100 can obtain real-time weather forecasts and traffic data from third-party data sources through relevant interfaces, serving as external data. The aggregation platform 100 can also obtain real-time delivery data from delivery service providers through interfaces with their systems. This delivery data includes, but is not limited to: the number of currently online delivery personnel, the number of orders awaiting processing, the number of orders currently in transit, the order response time for the last N orders (which can be set to seconds), the number of rejected / timeout orders for the last N orders, the current location distribution data of delivery personnel, the platform's real-time price quotes, and the platform's current service status (normal / busy / overwhelmed with orders), where N is a preset natural number. The aggregation platform 100 can process the accumulated external data (including real-time and historical data) to determine the multi-dimensional characteristic data of the delivery service provider corresponding to a preset prediction time window and its historical performance score based on past delivery results. The preset forecast time window is a time window that covers one or more delivery periods and includes at least the following relevant information: start time, end time, and time granularity. The time granularity is a forecast frequency or time point that is predetermined according to actual business needs, such as forecasting once every 15 minutes (or 30 minutes, 1 hour, etc.).

[0068] Continuing in this embodiment, in step S102, the aggregation platform 100 can convert the multi-dimensional feature data into feature vectors and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and determine the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result.

[0069] The aggregation platform 100, after determining the multi-dimensional feature data of the delivery service provider corresponding to the preset prediction time window and its historical performance score corresponding to previous delivery results, can convert the multi-dimensional feature data into feature vectors that meet the input requirements of the capacity prediction model and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window. Based on this capacity prediction result, the capacity score of the delivery service provider corresponding to the preset prediction time window can be determined. The capacity prediction result of the delivery service provider corresponding to the preset prediction time window, output by the capacity prediction model, may include the predicted order acceptance rate, response time, and capacity saturation. These can be processed to obtain corresponding scores, which are then weighted to obtain the capacity score. Specifically, the predicted order acceptance rate, response time, and capacity saturation can be standardized / normalized, converted into standard scores using appropriate mapping functions, and then weighted to obtain the capacity score. For example, the following mapping functions can be used to standardize / normalize the predicted order acceptance rate, predicted response time, and predicted saturation to obtain the corresponding scores:

[0070] Order acceptance rate score = min(predicted order acceptance rate / 0.95, 1.0)

[0071] Response time score = max(0, 1.0 - (predicted response time - 300) / 1800)

[0072] Saturation score = max(0, 1.0 - predicted saturation)

[0073] If the capacity prediction model outputs the following results: predicted order acceptance rate 0.88, predicted response time 520 (seconds), and predicted saturation 0.72, then, according to the above mapping function, the corresponding order acceptance rate score is 0.926, the response time score is 0.878, and the saturation score is 0.38. If the corresponding weights are preset to 0.4, 0.3, and 0.3 respectively (which can be set according to the cumulative delivery data analysis results), then after weighted calculation, the capacity score is 0.749.

[0074] Continuing in this embodiment, in step S103, after the delivery service provider obtains its capacity score corresponding to the preset prediction time window, the aggregation platform 100 can perform a weighted calculation (the sum of each weight is 1) based on the historical performance score determined in step S101 and the capacity score to obtain the delivery service provider's delivery capacity score corresponding to the preset prediction time window.

[0075] This embodiment allows for the dynamic evaluation of delivery service providers' real-time delivery capabilities. When scheduling order deliveries, the delivery service provider to be assigned to the order can be dynamically determined based on the real-time delivery capabilities of different delivery service providers and the delivery scheduling strategy. This can better meet users' delivery needs, improve user experience, and enhance the delivery service quality of delivery service providers.

[0076] Optionally, in step S101, the multi-dimensional feature data includes at least: order acceptance rate, average response time, current load rate, regional coverage, time characteristics, weather information, and traffic information.

[0077] In this optional embodiment, the aggregation platform 100 can process the acquired cumulative external data (including real-time and historical data) up to the present to determine the multi-dimensional feature data of the delivery service provider corresponding to the preset prediction time window. The multi-dimensional feature data may include at least: order acceptance rate, average response time, current load rate, regional coverage, time characteristics, weather information, traffic information, etc. For example, based on the accumulated external data (including real-time and historical data) and the related data obtained from processing it, the following can be calculated within the historical time window corresponding to the preset prediction time window: Order acceptance rate = (number of successfully accepted orders / total number of allocated orders) x 100%; Average response time = Sum(order acceptance time - order placement time) / number of successfully accepted orders; Single-signature load rate = number of orders in delivery / total number of deliverable orders (i.e., number of online delivery personnel x maximum number of delivery orders per delivery personnel); Area coverage = number of successfully delivered areas / total number of delivery areas; Time characteristics include time period coding (identifying the time period type, which can be one-hot ([morning peak, noon peak, evening peak, ...))). Off-peak hours, weekday codes (identifying date type, can be one-hot([Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday])), holiday codes (identifying whether it is a holiday, can be set to 1 for holidays, 0 otherwise, or vice versa); weather information includes weather codes (can be one-hot([Heavy rain (e.g., rainfall greater than 10 mm per hour), Heavy rain (e.g., rainfall greater than 5 mm per hour), Normal])) or directly preset weather factors (e.g., heavy rain is 0.75, greater than is 0.85, normal is 1.0); traffic information includes traffic codes (can be one-hot encoded according to the degree of congestion, e.g., one-hot([Severe congestion (red), moderate congestion (yellow), Normal])).

[0078] Optionally, in step S101, determining the historical performance score of the delivery service provider corresponding to a preset prediction time window includes:

[0079] Based on the accumulated delivery data, the delivery success rate score, delivery time score, and user satisfaction score of the delivery service provider corresponding to the preset prediction time window are determined. The delivery success rate score, delivery time score, and user satisfaction score are calculated based on the first preset weight to obtain the historical performance score of the delivery service provider corresponding to the preset prediction time window.

[0080] In this optional embodiment, the aggregation platform 100 can track the entire lifecycle (or process) of order delivery, process the accumulated delivery data related to the order, and determine the delivery service provider's scores across multiple dimensions corresponding to a preset prediction time window. These scores include delivery success rate, delivery time, and user satisfaction. Based on pre-set weights for each dimension, the platform performs a weighted calculation on the obtained delivery success rate, delivery time, and user satisfaction scores to obtain the delivery service provider's historical performance score corresponding to the preset prediction time window. Specifically, the delivery success rate score can be determined based on all completed delivery order data in the accumulated delivery data; the delivery time score can be determined based on successfully delivered order data in the accumulated delivery data; and the user satisfaction score can be determined based on user reviews of all orders in the accumulated delivery data. The platform can comprehensively preset the first weights for each dimension based on delivery business experience, customer and / or order nature. For example, delivery success rate is a core indicator directly affecting business performance, so its weight should be set to the highest; delivery time also directly affects competitiveness, so its weight should also be set relatively high; while the weight for user satisfaction can be set relatively low. In addition, if the proportion of high-value orders is relatively large, the weight of delivery success rate and user satisfaction can be appropriately increased; if the proportion of fragile orders is relatively large, the weight of delivery time and user satisfaction can be appropriately increased.

[0081] For example, the delivery success rate score = number of successfully delivered orders / total number of allocated orders x 100%; first calculate the average delivery time = Sum (delivery time - order time) / number of successfully delivered orders, and then perform a linear mapping based on a preset threshold. For example, the score for delivery time within 30 minutes is 1, the score for delivery time exceeding 60 minutes is 0, and the score for delivery time between 30 and 60 minutes is the result of the linear mapping of the average delivery time based on a preset linear function, for example, (average delivery time - 30) / (60 - 30) x 100%; if user rating data is usually in a 5-point scale, first calculate the average user satisfaction = Sum (user rating per order) / total number of orders, and then calculate the user satisfaction score = average user satisfaction / 5. Assuming the pre-set weights for delivery success rate, delivery time, and user satisfaction are 0.5, 0.3, and 0.2 respectively (the sum of the weights for each dimension should be 1), and the delivery success rate score, delivery time score, and user satisfaction score are a, b, and c respectively, the weighted calculation of these scores yields the historical performance score of the delivery service provider corresponding to the pre-set prediction time window, as follows:

[0082] Historical performance score = 0.5a + 0.3b + 0.2c

[0083] Optionally, in step S101, determining the historical performance score of the delivery service provider corresponding to the preset prediction time window further includes:

[0084] Based on the accumulated delivery data, the cost-effectiveness score and reliability score of the delivery service provider corresponding to the preset forecast time window are determined, wherein...

[0085] Based on a first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the historical performance score of the delivery service provider corresponding to a preset prediction time window, including:

[0086] Based on the first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the performance score of the delivery service provider corresponding to the preset prediction time window. The performance score, cost-effectiveness score, and reliability score are then used as the historical performance score of the delivery service provider corresponding to the preset prediction time window.

[0087] In this optional embodiment, the aggregation platform 100 can further process the cumulative delivery data related to orders to determine the cost-effectiveness score and reliability score of the delivery service provider in the cost-effectiveness dimension and reliability dimension, respectively, corresponding to a preset prediction time window. Specifically, based on the pre-set first preset weights corresponding to each dimension, the obtained delivery success rate score, delivery time score, and user satisfaction score are weighted and calculated to obtain the performance score of the delivery service provider corresponding to the preset prediction time window. This performance score, cost-effectiveness score, and reliability score are then combined to form the historical performance score of the delivery service provider corresponding to the preset prediction time window. Furthermore, based on the cumulative delivery data related to orders and the results of its processing, scores for several dimensions, such as average delivery cost, average delivery time, success rate, service stability, anomaly rate, complaint rate, and recovery capability, can be determined. These scores are then weighted and calculated according to preset weights corresponding to each dimension to obtain the cost-effectiveness score and reliability score. Specifically, the average delivery cost score can be calculated based on the delivery costs of all successful delivery orders corresponding to the preset forecast time window in the accumulated delivery data; the average delivery time score can be calculated based on the delivery time of all successful delivery orders corresponding to the preset forecast time window in the accumulated delivery data; the standard deviation of delivery time can also be calculated based on the delivery time of all successful delivery orders corresponding to the preset forecast time window in the accumulated delivery data; and then, based on the average delivery time and the standard deviation of delivery time, the coefficient of variation of delivery time can be calculated (the smaller the value, the higher the stability); and based on the coefficient of variation of delivery time, the service stability score can be calculated (score range: 0-1, the smaller the value, the higher the stability). (The closer to 1, the more stable); Based on all successfully delivered orders and unsuccessfully delivered orders corresponding to the preset prediction time window in the accumulated delivery data, a success rate score is calculated; Abnormal events in the accumulated delivery data are also identified (such as: orders whose delivery time exceeds a preset threshold (e.g., 90 minutes), rejected orders, abnormally canceled orders, etc.), and the recovery time for each abnormal event is calculated (the time from the occurrence of the abnormality to the restoration of normal service; the shorter the recovery time, the stronger the recovery capability; the sign of the restoration of normal service can be set to a preset number of consecutive orders being successfully delivered), and a recovery capability score is calculated (score range: 0-1, the closer to 1, the stronger the recovery capability).

[0088] For example, the cost-effectiveness score can be determined based on three dimensions: average delivery cost, average delivery time, and success rate. Assuming the preset weights for each dimension are 0.4, 0.35, and 0.25 (the sum of these weights is 1), and the average delivery cost score is d, the average delivery time score is e, and the success rate score is f, then the weighted cost-effectiveness score is as follows:

[0089] Cost-effectiveness score = 0.4d + 0.35e + 0.25f

[0090] As an example, referring to a Service Level Agreement (SLA), a reliability score is determined based on three dimensions: success rate, service stability, and resilience. Assuming the preset weights for each dimension are 0.4, 0.3, and 0.3 (the sum of these weights is 1), and the success rate score is g, the service stability score is h (service stability score = max(0,1 - delivery time coefficient of variation), where the delivery time coefficient of variation = standard deviation of delivery time / average delivery time), and the resilience score is i (recovery capability score = max(0,1 - average recovery time / baseline recovery time), where the baseline recovery time is a preset value, e.g., 90 minutes), the weighted reliability score is as follows:

[0091] Reliability score = 0.4g + 0.3h + 0.3i

[0092] Optionally, in step S102, the feature vector includes at least the following elements: delivery service provider identifier, order acceptance rate, average response time, current load rate, regional coverage, time period performance, weather impact factor, and traffic impact factor.

[0093] In this optional embodiment, after determining the multi-dimensional feature data and historical performance scores of the delivery service provider corresponding to the preset prediction time window, the aggregation platform 100 can preprocess the obtained multi-dimensional feature data and convert it into a feature vector that matches the capacity prediction model. The feature vector includes at least the following elements: element values ​​corresponding to the identifiers that correspond one-to-one with the delivery service providers, element values ​​corresponding to the order acceptance rate, element values ​​corresponding to the average response time, element values ​​corresponding to the current load rate, element values ​​corresponding to the regional coverage, element values ​​corresponding to the time period performance associated with the time characteristics, element values ​​corresponding to the weather impact factor, and element values ​​corresponding to the traffic impact factor.

[0094] This process can divide a day into multiple time periods, such as a breakfast period, or, based on accumulated delivery data, time periods such as breakfast 07:00-10:00, morning off-peak period 10:00-11:00, lunch 11:00-14:00, afternoon off-peak period 14:00-17:00, dinner 17:00-21:00, and nighttime off-peak period 21:00-22:00. Based on time characteristics, and combined with the accumulated delivery data of the delivery service provider for different time periods, time-related indicators are determined within a preset time range (e.g., the past 30 days up to the present), such as time-period order acceptance rate, average response time, success rate, and average delivery time. Then, weighted calculations are performed according to preset weights to obtain a time-period performance score (which can reflect the historical performance of the delivery service provider in a specific time period), thereby determining the element values ​​in the feature vector corresponding to the time-period performance.

[0095] The dimensions involved in the obtained multi-dimensional feature data are the same as those of the samples used to train the capacity prediction model, and the elements of the transformed feature vector also correspond to the elements of the feature vector transformed from the samples used to train the capacity prediction model.

[0096] Optionally, in step S102, the construction of the capacity prediction model includes:

[0097] Acquire historical delivery data from multiple delivery service providers and corresponding historical external data affecting delivery. Based on the historical delivery data, the historical external data, and the preset prediction time window, determine several historical multi-dimensional feature data and actual transportation capacity scores corresponding to the preset prediction time window.

[0098] Each historical multi-dimensional feature data is transformed into a feature vector, and its corresponding actual capacity score is used as its true value as a sample. This process is repeated for each historical multi-dimensional feature data and its actual capacity score to construct a sample set.

[0099] The sample set is divided into a training set, a test set, and a validation set. An LSTM neural network is trained to obtain a capacity prediction model.

[0100] In this optional embodiment, the aggregation platform 100 can acquire historical delivery data from multiple delivery service providers and corresponding historical external data affecting delivery. Based on the acquired historical delivery data, historical external data, and a preset prediction time window, it can obtain several historical multi-dimensional feature data and actual capacity scores corresponding to the preset prediction time window. The actual capacity includes three dimensions: order acceptance rate, response time, and capacity saturation. The actual capacity score is obtained by weighting the order acceptance rate score, response time score, and capacity saturation score. Each historical multi-dimensional feature data is converted into a feature vector, and its corresponding actual capacity score is used as its ground truth value to construct samples for training, testing, and / or validation. This process iterates through each historical multi-dimensional feature data and its actual capacity score to construct a sample set. This sample set is further divided into a training set, a test set, and a validation set to train an LSTM (Long Short-Term Memory) neural network to obtain a capacity prediction model. The specific methods for dividing the constructed sample set and training, testing, and validating the LSTM neural network to obtain the capacity prediction model are existing technologies and will not be elaborated here.

[0101] For multiple delivery service providers, the distribution of delivery demand corresponding to a preset forecast time window can be predicted. By combining the distribution of delivery demand with the capacity forecast results of each delivery service provider, the delivery capacity score of each delivery service provider corresponding to the preset forecast time window can be dynamically adjusted to further improve the accuracy and efficiency of delivery matching.

[0102] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0103] S104 For multiple delivery service providers, obtain the capacity prediction result of each delivery service provider corresponding to the preset prediction time window, and predict the distribution of delivery demand corresponding to the preset prediction time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers.

[0104] S105 identifies delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window based on the capacity forecast results of all delivery service providers and the distribution of delivery demand corresponding to the preset forecast time window, and adjusts the delivery capacity score of each delivery service provider corresponding to the preset forecast time window according to the identification results and preset adjustment strategies.

[0105] In this optional embodiment, in step S104, for multiple delivery service providers, the aggregation platform 100 can obtain the capacity prediction result of each delivery service provider corresponding to the preset prediction time window, and can also predict the distribution of delivery demand corresponding to the preset prediction time window based on the acquired cumulative external data and the cumulative delivery data of all delivery service providers.

[0106] Continuing in this optional embodiment, in step S105, the aggregation platform 100 can identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window based on the capacity forecast results and predicted delivery demand distribution of all delivery service providers corresponding to the preset forecast time window, and adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window according to the identification results and preset adjustment strategies.

[0107] Specifically, based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window, the available delivery capacity for the corresponding time period and region can be determined. Based on the distribution of delivery demand corresponding to the preset forecast time window, the delivery demand for the corresponding time period and region can be determined. Based on the ratio of delivery demand to available delivery capacity, high-risk time periods and / or high-risk regions with unbalanced loads can be determined (for example, if the preset threshold is 0.8, then if the ratio of delivery demand to available delivery capacity for a certain time period or region exceeds 0.8, then that time period and / or region can be determined as a high-risk time period and / or high-risk region).

[0108] To avoid large fluctuations in delivery capacity before and after adjustment, a smoothing function can be used to process the adjusted delivery capacity score, resulting in a new adjusted delivery capacity score. For example, an exponential moving average smoothing function can be used, with a smoothing exponent of 0.3. The exponential moving average smoothing function can be expressed as follows:

[0109] New delivery capacity score = Smoothing index x Adjusted delivery capacity score + (1 - Smoothing index) x Original delivery capacity score

[0110] The system can be set to automatically determine the risk level, affected time period and / or region, and expected overload volume based on the identification results and preset rules, and then calculate the adjustment amount and generate the adjustment strategy. The system can also set the adjustment effective time (for example, adjusting according to the adjustment strategy 15 minutes before the high-risk period). The preset rules can consider the following factors: basic adjustment amount (which can be determined according to the risk level), delivery service provider factor (which can be determined based on the delivery service provider's historical performance), and overload factor (which can be determined based on the overload volume; the larger the overload volume, the larger the adjustment range). The system can also combine the following factors to further adjust the delivery capacity score of each delivery service provider, such as the delivery service provider's delivery capacity score, load rate, regional matching degree, business strategy, etc.

[0111] It can also record the reasons for adjustments for subsequent audit traceability. For example, it can record the triggering reasons for adjustments to facilitate problem investigation; and record the effective time of adjustments for time sequence control to ensure that adjustments are executed at preset time points before risk periods, providing a warm-up period.

[0112] In this optional embodiment, based on the cumulative data, the distribution of delivery demand corresponding to the preset forecast time window and the capacity forecast results of all delivery service providers can be predicted to perform predictive load balancing and intervene in advance accordingly to adjust the delivery capacity of each delivery service provider corresponding to the preset forecast time window. This ensures that only delivery service providers that are truly capable of coping with high-risk periods and / or high-risk areas participate in load adjustment, avoiding excessive adjustment of delivery service providers with weak delivery capacity that could lead to delivery service interruptions. This can further improve the delivery success rate and enhance the overall stability of the delivery system.

[0113] Optionally, the step of predicting the distribution of delivery demand corresponding to the preset prediction time window based on the acquired cumulative external data and the cumulative delivery data of all delivery service providers includes:

[0114] Based on the accumulated delivery data of all delivery service providers, a time series forecasting model is used to predict the basic distribution of delivery demand corresponding to the preset forecasting time window.

[0115] Based on the acquired cumulative external data, determine the external influencing factors corresponding to the preset prediction time window;

[0116] Based on the basic distribution of delivery demand and the external influencing factors, determine the distribution of delivery demand corresponding to the preset prediction time window.

[0117] In this optional embodiment, the aggregation platform 100 can extract time dimension information from the accumulated delivery data of all delivery service providers using a time series prediction model, such as the Autoregressive Integrated Moving Average (ARIMA) model, and aggregate it according to the time granularity within a preset prediction time window to obtain several continuous time series data. This time series data can accurately reflect the time distribution characteristics of delivery demand, and the corresponding true values ​​can be determined from the accumulated delivery data. The time series data and the corresponding true values ​​can be used as samples to construct and train the ARIMA to obtain a predicted delivery demand distribution model. The predicted delivery demand distribution model obtained after training, testing, and verification can accurately predict the distribution of delivery demand according to the time granularity within a preset prediction time window. Key steps may include:

[0118] 1. Time field extraction: Extract time information from historical order creation time, expected time, order acceptance time, completion time, and other relevant fields in the cumulative delivery data;

[0119] 2. Constructing the time series: Select a time period with the same time granularity as the preset prediction time window to construct the time series;

[0120] 3. Data aggregation: Analyze metrics such as order quantity, average duration, and success rate for each time period to construct time series data;

[0121] 4. Sequence imputation: Handles missing time points to ensure sequence continuity;

[0122] 5. Outlier handling: Identify and correct outlier data points.

[0123] In addition to time series data on delivery demand, multi-dimensional time series data can be constructed as needed, such as time series data on delivery services and time series data on delivery efficiency.

[0124] Specifically, a trained time series prediction model can be used to accurately predict the distribution of delivery demand within a preset prediction time window based on time granularity, thus obtaining the basic distribution of delivery demand corresponding to that preset prediction time window. The aggregation platform 100 can also determine the external influencing factors corresponding to the preset prediction time window based on acquired accumulated external data, such as historical weather and traffic data for the same time period. These external influencing factors are then used to correct the obtained basic distribution of delivery demand, resulting in the final delivery demand distribution corresponding to the preset prediction time window.

[0125] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0126] S106 Obtains the order information of the orders to be delivered corresponding to the preset prediction time window, and determines the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information;

[0127] S108 adjusts the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area based on the order characteristics, and determines the delivery service provider to execute the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

[0128] In this optional embodiment, when the aggregation platform 100 receives a delivery order from a merchant that corresponds to the preset predicted time window, in step S106, it obtains the order information of the delivery order and determines the delivery time period, delivery area, and order characteristics of the delivery order based on the order information. The order information may include: basic information (e.g., order number, store information, order time, etc.), pickup information (e.g., latitude and longitude + detailed address, contact person, contact number, etc.), receiving information (e.g., latitude and longitude + detailed address, recipient, contact number, etc.), product information (e.g., product list, total value, product type, weight, etc.), timeliness requirements (e.g., expected delivery time, latest delivery time, whether it is urgent, etc.), and cost information (e.g., single item amount, total amount, delivery fee limit, etc.). Continuing with this optional embodiment, in step S107, the aggregation platform 100 can adjust the weights of various dimensions related to the delivery time and delivery area of ​​the order to be delivered based on the order characteristics of the order to be delivered, thereby adjusting the delivery capacity score of each delivery service provider, and determining the delivery service provider to perform the delivery task of the order to be delivered based on the adjusted delivery capacity score of each delivery service provider and the preset selection strategy. Order characteristics may include: time period characteristics, urgency level, expected delivery time, delivery distance, product type / weight, delivery fee limit, etc. For example, for urgent or peak-period orders to be delivered, the weight of the delivery service provider's capacity dimension in the corresponding time period and delivery area can be increased according to the corresponding preset rules; for orders to be delivered of fragile or perishable goods, or for high-value orders to be delivered for high-value goods or key customers, the weight of the delivery service provider's performance, reliability, and other dimensions in the corresponding time period and delivery area can be increased according to the corresponding preset rules. The preset selection strategies may include: the optimal delivery service provider selection strategy, which selects the delivery service provider with the highest dynamically adjusted delivery capacity score to execute the order to be delivered; the multi-delivery service provider selection strategy, which selects multiple delivery service providers with the highest dynamically adjusted delivery capacity scores, sends them sequentially or simultaneously, and the delivery service provider that responds first executes the order to be delivered; and the load balancing selection strategy, which calculates a comprehensive score based on the dynamically adjusted delivery capacity and load rate of each delivery service provider, and selects the delivery service provider with the highest comprehensive score to execute the order to be delivered.

[0129] In this optional embodiment, the weights of various dimensions related to the delivery time and delivery area of ​​the order to be delivered can be dynamically adjusted based on the order characteristics of the order to be delivered, thereby adjusting the delivery capability score of each delivery service provider, so as to finally determine the most suitable delivery service provider to perform the delivery task of the order to be delivered, so as to further improve the delivery success rate.

[0130] Optionally, prior to step S108, the method for dynamically assessing the delivery capabilities of a delivery service provider further includes:

[0131] S107 determines the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of each delivery service provider based on the cumulative delivery data of each delivery service provider for the delivery time period and delivery area. Based on the second preset weight, the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score are calculated to obtain the regional matching score of the delivery service provider for the delivery time period and delivery area. Based on the regional matching score, the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area is adjusted.

[0132] In this optional embodiment, when adjusting the delivery capacity score of the delivery service provider and the delivery time and delivery area corresponding to the order to be delivered, the regional matching degree dimension of the delivery area is also taken into consideration. In step S107, the aggregation platform 100 can determine the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of each delivery service provider corresponding to the delivery time and delivery area based on the cumulative delivery data of each delivery service provider. Then, it performs a weighted calculation of the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of the delivery service provider according to a second preset weight, to obtain the regional matching score of each delivery service provider for the corresponding delivery time and delivery area. Based on the corresponding regional matching score, the delivery capacity score of each delivery service provider corresponding to the delivery time and delivery area is adjusted. Specifically, the historical performance score obtained in step S101, the capacity score obtained in step S102, and the regional matching score obtained in step S107 are weighted and calculated according to the preset weights corresponding to each score (the sum of all weights is 1), to obtain the delivery capacity score of each delivery service provider corresponding to the delivery time and delivery area.

[0133] Optionally, the method for dynamically assessing the delivery capabilities of delivery service providers further includes:

[0134] S109 updates the cumulative delivery data of the corresponding delivery service provider based on the execution result of the delivery task.

[0135] In this optional embodiment, after a delivery person from a relevant delivery service provider completes a delivery task, the delivery service provider's system sends the corresponding delivery task execution result to the aggregation platform 100. The aggregation platform 100 can update the delivery service provider's cumulative delivery data to update the historical performance score of the delivery service provider. It can also construct new samples for iterative training of capacity prediction models and time series prediction models. The delivery task execution result may include: delivery service provider identifier, order characteristics, response time, delivery time, user satisfaction, cost-effectiveness, etc.

[0136] The method for dynamically evaluating the delivery capabilities of delivery service providers, provided by the above embodiments and / or optional embodiments, can dynamically evaluate the delivery capabilities of each delivery service provider based on multiple dimensions such as capacity forecasting, historical performance, and regional matching. This allows for efficient and flexible delivery scheduling based on the dynamic delivery capabilities of different delivery service providers, improving delivery efficiency and success rate. Furthermore, it can predict delivery demand distribution using a time-series forecasting model at a time granularity, and combine this with the capacity forecasts of each delivery service provider to perform predictive load balancing and dynamically adjust the delivery capabilities of each provider accordingly. This avoids periods and / or areas of concentrated capacity overflow, improving delivery success rate and overall stability of the delivery system. It can also automatically accumulate data to continuously optimize capacity forecasting and delivery demand distribution forecasting, reducing the rate of manual intervention.

[0137] From the production environment of the aggregation platform 100, 50 representative restaurant chain stores and 8 mainstream delivery service providers were selected for a 3-month comparative verification covering different time characteristics. Compared with the delivery scheduling scheme that sets fixed weights based on the historical performance of each delivery service provider, the delivery scheduling scheme that dynamically adjusts the weights and evaluates the delivery capability scores of each delivery service provider according to this application can improve delivery efficiency by 20-40%, improve the overall stability of the delivery system by more than 60%, increase the delivery success rate to 99.5%, and reduce the manual intervention rate by more than 90%.

[0138] Figure 2 The diagram illustrates a system for dynamically assessing the delivery capabilities of a delivery service provider according to another aspect of this application, wherein, in one embodiment, the system includes:

[0139] The first module 210 is used to acquire external data affecting order delivery and delivery data of delivery service providers in real time, and determine the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the cumulative external data and cumulative delivery data up to the present.

[0140] The second module 220 is used to convert the multi-dimensional feature data into feature vectors and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and to determine the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result.

[0141] The third module 230 is used to determine the delivery capacity score of the delivery service provider corresponding to the preset prediction time window based on the historical performance score and the capacity score.

[0142] In this system embodiment, the system is deployed in the aggregation platform 100.

[0143] Through the first module 210 of this system, weather forecasts, traffic data, and other information provided by third-party data sources can be acquired in real time as external data. Furthermore, through interfaces with the relevant systems of delivery service providers, related delivery data can be obtained from them in real time. The system can also process the accumulated external data (including real-time and historical data) to determine the multi-dimensional characteristic data of the delivery service provider corresponding to a preset prediction time window and its historical performance score based on past delivery results. The preset prediction time window covers one or more delivery periods and includes at least the following information: start time, end time, and time granularity. The time granularity is a prediction frequency or time point predetermined according to actual business needs, such as predicting every 15 minutes (or 30 minutes, 1 hour, etc.).

[0144] After determining the multi-dimensional feature data of the delivery service provider corresponding to the preset prediction time window and its historical performance score corresponding to previous delivery results, the second module 220 of the system can convert the multi-dimensional feature data into feature vectors that meet the input requirements of the capacity prediction model and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window. Based on this capacity prediction result, the capacity score of the delivery service provider corresponding to the preset prediction time window can be determined. The capacity prediction result of the delivery service provider corresponding to the preset prediction time window output by the capacity prediction model may include the predicted order acceptance rate, response time, and capacity saturation. These can be processed to obtain corresponding scores, and then weighted to obtain the capacity score. Specifically, the predicted order acceptance rate, response time, and capacity saturation can be standardized / normalized, converted into standard scores through a corresponding mapping function, and then weighted to obtain the capacity score.

[0145] The third module 230 of the system can perform a weighted calculation (the sum of each weight is 1) on the historical performance score obtained through the first module 210 and the capacity score obtained through the second module to obtain the delivery capacity score of the delivery service provider corresponding to the preset prediction time window.

[0146] Through this system implementation, the real-time delivery capabilities of delivery service providers can be dynamically evaluated. When scheduling order delivery, the delivery service provider to perform the delivery task of the order to be delivered can be dynamically determined based on the real-time delivery capabilities of different delivery service providers and the delivery scheduling strategy. This can better meet users' delivery needs, improve user experience, and improve the delivery service quality of delivery service providers.

[0147] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0148] The fourth module 240 is used to obtain the capacity prediction result of each delivery service provider corresponding to the preset prediction time window for multiple delivery service providers, and predict the distribution of delivery demand corresponding to the preset prediction time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers.

[0149] The fifth module 250 is used to identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window and the distribution of delivery demand, and to adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window based on the identification results and preset adjustment strategies.

[0150] For multiple delivery service providers, in this optional embodiment, the fourth module 240 of the system can obtain the capacity forecast result for each delivery service provider corresponding to a preset forecast time window. It can also predict the distribution of delivery demand corresponding to the preset forecast time window based on the acquired cumulative external data and the cumulative delivery data of all delivery service providers. Continuing in this optional embodiment, the fifth module 250 of the system can identify potentially unbalanced delivery periods and areas corresponding to the preset forecast time window based on the capacity forecast results and predicted delivery demand distribution of all delivery service providers corresponding to the preset forecast time window. Based on the identification results and a preset adjustment strategy, it can adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window.

[0151] Through this optional embodiment, the distribution of delivery demand corresponding to a preset forecast time window can be predicted based on accumulated data and the capacity forecast results of all delivery service providers. Predictive load balancing can then be performed, and the delivery capacity of each delivery service provider corresponding to the preset forecast time window can be adjusted accordingly, thereby further improving the delivery success rate and the overall stability of the delivery system.

[0152] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0153] The sixth module 260 is used to obtain the order information of the orders to be delivered corresponding to the preset prediction time window, and to determine the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information;

[0154] The eighth module 280 is used to adjust the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area according to the order characteristics, and to determine the delivery service provider to perform the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

[0155] In this optional embodiment, when a merchant sends a delivery order corresponding to the preset predicted time window, the sixth module 260 of the system can also obtain the order information of the delivery order and determine the delivery time period, delivery area, and order characteristics of the delivery order based on the order information. Continuing in this optional embodiment, the eighth module 280 of the system can adjust the weights of various dimensions related to the delivery time period and delivery area of ​​the delivery order based on the order characteristics of the delivery order, thereby adjusting the delivery capability score of each delivery service provider. Based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy, the delivery service provider to execute the delivery task of the delivery order is determined.

[0156] In this optional embodiment, the weights of various dimensions related to the delivery time and delivery area of ​​the order to be delivered can be dynamically adjusted based on the order characteristics of the order to be delivered, thereby adjusting the delivery capability score of each delivery service provider, so as to finally determine the most suitable delivery service provider to perform the delivery task of the order to be delivered, so as to further improve the delivery success rate.

[0157] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0158] The seventh module 270 is used to determine the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of each delivery service provider corresponding to the delivery time period and delivery area based on the cumulative delivery data of each delivery service provider. It also calculates the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score based on a second preset weight to obtain the regional matching score of the delivery service provider corresponding to the delivery time period and delivery area. Based on the regional matching score, it adjusts the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area.

[0159] In this optional embodiment, the seventh module 270 of the system can also determine the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score for each delivery service provider corresponding to the delivery time period and delivery area based on the cumulative delivery data of each delivery service provider. Then, based on a second preset weight, the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score corresponding to each delivery service provider are weighted and calculated to obtain the regional matching score for each delivery service provider for the corresponding delivery time period and delivery area. Based on the corresponding regional matching score, the delivery capacity score of each delivery service provider corresponding to the delivery time period and delivery area is adjusted. Specifically, the historical performance score obtained through the first module 210, the capacity score obtained through the second module 220, and the regional matching score obtained through the seventh module 270 are weighted and calculated according to the preset weights corresponding to each score (the sum of all weights is 1) to obtain the delivery capacity score of each delivery service provider corresponding to the delivery time period and delivery area.

[0160] In this optional embodiment, when adjusting the delivery time period and delivery capacity score corresponding to the delivery area and the delivery service provider and the order to be delivered, the regional matching degree dimension of the delivery area is also taken into consideration, which can further improve the accuracy of scheduling and further improve the delivery success rate.

[0161] Optionally, the system for dynamically evaluating the delivery capabilities of delivery service providers further includes:

[0162] The ninth module 290 is used to update the cumulative delivery data of the corresponding delivery service provider based on the execution result of the delivery task.

[0163] In this optional embodiment, after a delivery person from a relevant delivery service provider completes a delivery task, the delivery service provider's system sends the corresponding delivery task execution result to the aggregation platform 100. Through the system's ninth module 290, this result can be updated to the delivery service provider's cumulative delivery data. This is used to update the historical performance score of the delivery service provider and to construct new samples for iterative training of capacity prediction models and time series prediction models. The delivery task execution result may include: delivery service provider identifier, order characteristics, response time, delivery time, user satisfaction, cost-effectiveness, etc.

[0164] In the above system embodiments and / or optional embodiments, any parts of the system not mentioned are the same as those in the aforementioned related method embodiments and / or optional embodiments, and will not be repeated here.

[0165] The above-described system embodiments and / or optional embodiments provide a system for dynamically evaluating the delivery capabilities of delivery service providers. This system can dynamically evaluate the delivery capabilities of each delivery service provider based on multiple dimensions such as capacity forecasting, historical performance, and regional matching. This allows for efficient and flexible delivery scheduling based on the dynamic delivery capabilities of different delivery service providers, improving delivery efficiency and success rate. Furthermore, it can predict the distribution of delivery demand corresponding to preset forecast time windows, and, combined with the capacity forecasts of each delivery service provider, perform predictive load balancing and dynamically adjust the delivery capabilities of each provider accordingly. This avoids periods and / or areas of concentrated capacity overflow, improving delivery success rate and overall stability of the delivery system. It can also automatically accumulate data to continuously optimize capacity forecasting and delivery demand distribution forecasting, reducing the rate of manual intervention.

[0166] According to another aspect of this application, a computer-readable medium is also provided, on which computer-readable instructions are stored, which can be executed by a processor to implement some or all of the foregoing method embodiments and / or optional embodiments.

[0167] It should be noted that the method embodiments and / or optional embodiments in this application do not strictly limit the order of execution of each step, as long as the method embodiments and / or optional embodiments can solve the defects existing in the prior art, achieve the inventive purpose of this application, and obtain beneficial effects. The method embodiments and / or optional embodiments in this application can be implemented in software and / or combinations of software and hardware. The software program involved in this application can be executed by a processor to implement the steps or functions of the above embodiments. Similarly, the software program of this application (including related data structures) can be stored in a computer-readable recording medium.

[0168] Furthermore, part or all of this application can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. The program instructions invoking the methods of this application may be stored in a fixed or removable recording medium, and / or transmitted via data streams in broadcast or other signal carrying media, and / or stored in the working memory of a computer device operating according to the program instructions.

[0169] According to another aspect of this application, an apparatus for dynamically assessing the delivery capabilities of a delivery service provider is also provided. The apparatus includes: a memory for storing computer-readable instructions and one or more processors for executing the computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, the apparatus is triggered to run part or all of the methods and / or technical solutions of the foregoing embodiments.

[0170] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0171] In this application, when terms such as "upper," "lower," "left," "right," "front," "rear," "top," "bottom," "inner," "outer," "middle," "vertical," "horizontal," "lateral," and "longitudinal" are used, the indicated orientation and / or positional relationship is based on the orientation and / or positional relationship shown in the accompanying drawings. These terms are primarily for the purpose of better describing this application and its embodiments, and are not intended to limit the indicated device, element, or component to having a specific orientation, or to be constructed and operated in a specific orientation. Furthermore, some of the above terms, in addition to indicating orientation or positional relationship, can also be used to indicate other meanings; for example, the term "upper" can also be used in some cases to indicate a certain dependency or connection relationship. Those skilled in the art can understand the specific meaning of these terms in this application according to the specific circumstances.

[0172] Furthermore, the terms "installation," "setup," "equipped with," "connection," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral structure; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection via an intermediate medium; and they can refer to an internal connection between two devices, components, or constituent parts. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0173] Furthermore, the terms "first," "second," etc., are primarily used to distinguish different devices, units, modules, elements, circuits, or components (which may be the same or different in specific type and construction), and are not intended to indicate or imply the relative importance, order, and / or quantity of the indicated devices, units, modules, elements, circuits, or components. Unless otherwise stated, "a plurality of" means two or more.

[0174] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in the apparatus claims may also be implemented by a single unit or device through software and / or hardware.

Claims

1. A method for dynamically evaluating the delivery capabilities of delivery service providers, characterized in that, The method includes: The system acquires external data affecting order delivery and delivery data of delivery service providers in real time. Based on the cumulative external data and cumulative delivery data up to the present, it determines the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window. The multi-dimensional feature data includes at least: order acceptance rate, average response time, current load rate, regional coverage, time characteristics, weather information, and traffic information. The multi-dimensional feature data is converted into feature vectors and input into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window. Based on the capacity prediction result, the capacity score of the delivery service provider corresponding to the preset prediction time window is determined. Based on the historical performance score and the capacity score, the delivery capacity score of the delivery service provider corresponding to the preset forecast time window is determined; For multiple delivery service providers, obtain the capacity forecast result for each delivery service provider corresponding to the preset forecast time window, and predict the distribution of delivery demand corresponding to the preset forecast time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers. Based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window and the distribution of delivery demand, identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window, and adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window according to the identification results and preset adjustment strategies. Obtain the order information of the orders to be delivered corresponding to the preset prediction time window, and determine the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information; Based on the order characteristics, adjust the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area, and determine the delivery service provider to execute the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

2. The method according to claim 1, characterized in that, Determining the historical performance score of the delivery service provider corresponding to a preset forecast time window includes: Based on the accumulated delivery data, the delivery success rate score, delivery time score, and user satisfaction score of the delivery service provider corresponding to the preset prediction time window are determined. The delivery success rate score, delivery time score, and user satisfaction score are calculated based on the first preset weight to obtain the historical performance score of the delivery service provider corresponding to the preset prediction time window.

3. The method according to claim 2, characterized in that, Determining the historical performance score of the delivery service provider corresponding to the preset prediction time window also includes: Based on the accumulated delivery data, determine the cost-effectiveness score and reliability score of the delivery service provider corresponding to the preset forecast time window, wherein... Based on a first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the historical performance score of the delivery service provider corresponding to a preset prediction time window, including: Based on the first preset weight, the delivery success rate score, delivery time score, and user satisfaction score are calculated to obtain the performance score of the delivery service provider corresponding to the preset prediction time window. The performance score, cost-effectiveness score, and reliability score are then used as the historical performance score of the delivery service provider corresponding to the preset prediction time window.

4. The method according to claim 1, characterized in that, The feature vector includes at least the following elements: delivery service provider identifier, order acceptance rate, average response time, current load rate, regional coverage, time period performance, weather impact factor, and traffic impact factor.

5. The method according to claim 1, characterized in that, The construction of the capacity prediction model includes: Acquire historical delivery data from multiple delivery service providers and corresponding historical external data affecting delivery. Based on the historical delivery data, the historical external data, and the preset prediction time window, determine several historical multi-dimensional feature data and actual transportation capacity scores corresponding to the preset prediction time window. Each historical multi-dimensional feature data is transformed into a feature vector, and its corresponding actual capacity score is used as its true value as a sample. This process is repeated for each historical multi-dimensional feature data and its corresponding actual capacity score to construct a sample set. The sample set is divided into a training set, a test set, and a validation set. An LSTM neural network is trained to obtain a capacity prediction model.

6. The method according to claim 1, characterized in that, The step of predicting the distribution of delivery demand corresponding to the preset prediction time window based on the acquired cumulative external data and the cumulative delivery data of all delivery service providers includes: Based on the accumulated delivery data of all delivery service providers, a time series forecasting model is used to predict the basic distribution of delivery demand corresponding to the preset forecasting time window. Based on the acquired cumulative external data, determine the external influencing factors corresponding to the preset prediction time window; Based on the basic distribution of delivery demand and the external influencing factors, determine the distribution of delivery demand corresponding to the preset prediction time window.

7. The method according to claim 1, characterized in that, Before adjusting the delivery capacity score of each delivery service provider corresponding to the delivery time period and delivery area based on the order characteristics, the method further includes: Based on the cumulative delivery data of each delivery service provider, the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of the delivery service provider for the corresponding delivery time period and delivery area are determined. The regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score are calculated based on a second preset weight to obtain the regional matching score of the delivery service provider for the corresponding delivery time period and delivery area. Based on the regional matching score, the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area is adjusted.

8. The method according to claim 1, characterized in that, The method further includes: Based on the execution results of the delivery task, update the cumulative delivery data of the corresponding delivery service provider.

9. A system for dynamically evaluating the delivery capabilities of delivery service providers, characterized in that, The system includes: The first module is used to acquire external data affecting order delivery and delivery data of delivery service providers in real time, and determine the multi-dimensional feature data and historical performance score of the delivery service provider corresponding to the preset prediction time window based on the cumulative external data and cumulative delivery data up to the present. The multi-dimensional feature data includes at least: order acceptance rate, average response time, current load rate, regional coverage, time characteristics, weather information, and traffic information. The second module is used to convert the multi-dimensional feature data into feature vectors and input them into the capacity prediction model to obtain the capacity prediction result of the delivery service provider corresponding to the preset prediction time window, and to determine the capacity score of the delivery service provider corresponding to the preset prediction time window based on the capacity prediction result. The third module is used to determine the delivery capacity score of the delivery service provider corresponding to the preset prediction time window based on the historical performance score and the capacity score. The fourth module is used to obtain the capacity prediction results of each delivery service provider corresponding to the preset prediction time window for multiple delivery service providers, and to predict the distribution of delivery demand corresponding to the preset prediction time window based on the obtained cumulative external data and the cumulative delivery data of all delivery service providers. The fifth module is used to identify delivery periods and delivery areas with potential load imbalances corresponding to the preset forecast time window based on the capacity forecast results of all delivery service providers corresponding to the preset forecast time window and the distribution of delivery demand, and to adjust the delivery capacity score of each delivery service provider corresponding to the preset forecast time window based on the identification results and the preset adjustment strategy. The sixth module is used to obtain the order information of the orders to be delivered corresponding to the preset prediction time window, and to determine the delivery time period, delivery area and order characteristics of the orders to be delivered based on the order information. The eighth module is used to adjust the delivery capability score of each delivery service provider corresponding to the delivery time period and delivery area according to the order characteristics, and to determine the delivery service provider to perform the delivery task of the order to be delivered based on the adjusted delivery capability score of each delivery service provider and the preset selection strategy.

10. The system according to claim 9, characterized in that, The system also includes: The seventh module is used to determine the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score of each delivery service provider based on the cumulative delivery data of each delivery service provider for the delivery time period and delivery area. Based on a second preset weight, the module calculates the regional delivery personnel density score, regional coverage score, regional historical performance score, and regional adaptation score to obtain the regional matching score of the delivery service provider for the delivery time period and delivery area. Based on the regional matching score, the module adjusts the delivery capability score of the delivery service provider corresponding to the delivery time period and delivery area.

11. A computer-readable medium, characterized in that, It stores computer-readable instructions that are executed by a processor to implement part or all of the method as claimed in any one of claims 1 to 8.

12. A device for dynamically evaluating the delivery capabilities of a delivery service provider, characterized in that, The device includes: One or more processors; and A memory storing computer-readable instructions, which, when executed, cause the processor to perform some or all of the operations of the method as described in any one of claims 1 to 8.

Images