Logistics transportation method and system based on carpooling, electronic equipment and storage medium

Abstract

The invention discloses a carpooling-based logistics transportation method and system, electronic equipment and a storage medium, and the method comprises the steps: carrying out the normalization calculation of data in a cargo transportation demand data set and data in a vehicle data set, and obtaining a normalized cargo transportation demand data set and a normalized vehicle data set; extracting spatio-temporal features between data in the normalized cargo transportation demand data set and data in the normalized vehicle data set; calculating a multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatial-temporal characteristics; clustering data in the normalized cargo transportation demand data set according to the multi-dimensional weighted distance to obtain a demand clustering result; carpooling matching is carried out on all cargo transportation demands of each category in the demand clustering result, and a target vehicle is obtained; and determining an optimal route according to the cargo transportation demand in the demand clustering result and the target vehicle. The vehicle resource utilization rate can be improved, and the logistics transportation efficiency is improved.

Description

Technical Field

[0001] The present application relates to the technical field of logistics transportation, and in particular, to a logistics transportation method, system, electronic device and storage medium based on carpooling. Background Art

[0002] In traditional logistics transportation methods, many vehicles operate without being fully loaded, resulting in high transportation costs. The vehicle empty load rate is high and the resource utilization rate is low. Secondly, due to the lack of an effective demand matching mechanism, many transportation vehicles are not fully utilized, causing waste of transportation resources. Empty vehicles not only increase transportation costs but also cause unnecessary pollution to the environment. In traditional logistics transportation methods, the transportation routes of goods are often not optimized, resulting in long transportation times and low efficiency. At the same time, the goods of multiple users are transported separately, lacking intensive management, which further reduces the transportation efficiency. And due to the lack of real-time monitoring and feedback mechanisms, it is difficult for users to timely understand the transportation status during the transportation of goods, lacking a sense of security and satisfaction. In addition, problems that may occur during transportation cannot be solved in a timely manner, affecting the user experience.

[0003] In summary, the main problems in traditional logistics transportation methods are low vehicle resource utilization rate and low logistics transportation efficiency. Summary of the Invention

[0004] The present application aims to provide a logistics transportation method, system, electronic device and storage medium based on carpooling, which can improve the vehicle resource utilization rate and the logistics transportation efficiency.

[0005] In a first aspect, an embodiment of the present application provides a logistics transportation method based on carpooling, the method comprising:

[0006] Obtain a cargo transportation demand data set and a vehicle data set, wherein the cargo transportation demand data set contains cargo transportation demand data submitted by users related to the cargo, and the vehicle data set contains data related to the vehicle;

[0007] Perform normalization calculations on the data in the cargo transportation demand data set and the data in the vehicle data set to obtain a normalized cargo transportation demand data set and a normalized vehicle data set;

[0008] Extract the spatio-temporal features between each data in the normalized cargo transportation demand data set and the normalized vehicle data set, wherein the spatio-temporal features are used to characterize the spatio-temporal relationship between the cargo transportation demand and the vehicle;

[0009] Calculate the multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatio-temporal features;

[0010] Cluster the data in the normalized freight transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result;

[0011] Perform carpool matching on all freight transportation demands in each category of the demand clustering result to obtain target vehicles;

[0012] Determine the optimal route according to the freight transportation demands in the demand clustering result and the target vehicles.

[0013] Compared with the prior art, the first aspect of the present application has the following beneficial effects:

[0014] This method obtains a freight transportation demand dataset and a vehicle dataset. Among them, the freight transportation demand dataset contains freight transportation demand data submitted by users related to goods, and the vehicle dataset contains data related to vehicles. Perform normalization calculations on the data in the freight transportation demand dataset and the data in the vehicle dataset to obtain a normalized freight transportation demand dataset and a normalized vehicle dataset. By normalizing the data, the effectiveness and reasonableness of the data can be ensured, and errors caused by static standardization can be avoided. Then extract the spatio-temporal features between each data in the normalized freight transportation demand dataset and the normalized vehicle dataset. Among them, the spatio-temporal features are used to characterize the spatio-temporal relationship between the freight transportation demand and the vehicle. By extracting the spatio-temporal relationship between the freight transportation demand and the vehicle, a good data foundation is laid for later calculations. Then, according to the spatio-temporal features, calculate the multi-dimensional weighted distance between the freight transportation demand and the vehicle. According to the multi-dimensional weighted distance, cluster the data in the normalized freight transportation demand dataset to obtain a demand clustering result. By clustering a series of freight transportation demands, a good data foundation is laid for later matching of suitable vehicles. Perform carpool matching on all freight transportation demands in each category of the demand clustering result to obtain target vehicles. Determine the optimal route according to the freight transportation demands in the demand clustering result and the target vehicles. By matching vehicles for the clustered freight transportation demands, the empty load rate can be reduced and the resource utilization rate can be improved, and the transportation route can be better optimized.

[0015] In some embodiments, the performing normalization calculations on the data in the freight transportation demand dataset and the data in the vehicle dataset to obtain a normalized freight transportation demand dataset and a normalized vehicle dataset includes:

[0016]

[0017] Among them, Norm(x i ) represents the normalized freight transportation demand dataset, x i represents the i-th freight transportation demand data, μ t,sDenotes the mean over time t and space s, δ t Denotes the time-related deviation value, σ t,s Denotes the standard deviation over time t and space s, β denotes the adjustment coefficient for controlling the data normalization effect, Norm(x j ) Denotes the normalized vehicle dataset, x j Denotes the j-th vehicle data.

[0018] In some embodiments, extracting the spatio-temporal features between each data in the normalized freight transport demand dataset and the normalized vehicle dataset includes:

[0019]

[0020] Wherein, Feature ij Denotes the spatio-temporal feature between the i-th freight transport demand data and the j-th vehicle data extracted, W t,s,k Denotes the adaptive convolution kernel weight for the k-th time period t and space region s, * denotes the convolution operator, Norm(x i ) Denotes the normalized freight transport demand dataset, Norm(x j ) Denotes the normalized vehicle dataset, Input represents the input, K represents the number of convolution kernels, T represents the time length considered, x t,s Denotes the freight transport demand data for the time period t and space region s, Pooling represents the global pooling operation, Pooling(x t,s ) Denotes performing the global pooling operation on the freight transport demand data for different time periods t and space regions s.

[0021] In some embodiments, calculating the multi-dimensional weighted distance between the freight transport demand and the vehicle based on the spatio-temporal features includes:

[0022]

[0023] Wherein, D ij Denotes the multi-dimensional weighted distance between the i-th freight transport demand and the j-th vehicle, M represents the number of spatio-temporal features, w k Denotes the weighting coefficient for the k-th spatio-temporal feature, x ik Denotes the value of the i-th freight transport demand data on the k-th spatio-temporal feature, x jk Denotes the value of the j-th vehicle data on the k-th spatio-temporal feature, σ k Denotes the standard deviation of the k-th spatio-temporal feature, λ represents the weighting coefficient, t i Denotes the time information of the i-th freight transport demand data, t jRepresents the time information of the j-th vehicle data, and ΔT represents the normalization coefficient of time.

[0024] In some embodiments, clustering the data in the normalized cargo transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result, including:

[0025]

[0026] Among them, G k Represents the centroid of the k-th cluster, and G k Represents the set of cargo transportation demand points of the k-th cluster, y i Represents the feature vector of the i-th cargo transportation demand point, D ij Represents the multi-dimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle, and θ k Represents the adaptive threshold of the k-th cluster.

[0027] In some embodiments, performing carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles, including:

[0028] Construct a target fitness function;

[0029] Based on the target fitness function, use a genetic algorithm to perform carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles.

[0030] In some embodiments, constructing the target fitness function includes:

[0031]

[0032] Among them, f fitness (x) represents the target fitness function, Weight(i) represents the weight of the i-th cargo transportation demand, D ij (i, x) represents the multi-dimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle in the vehicle set x that meets the constraint conditions, γ represents the penalty factor, Penalty(x) represents the penalty term for violating the constraint conditions, and N represents the total number of cargo transportation demands in each category of the clustering result.

[0033] Second, the embodiments of the present application also provide a logistics transportation system based on carpooling, and the system includes:

[0034] A data acquisition unit, configured to acquire a cargo transportation demand dataset and a vehicle dataset, where the cargo transportation demand dataset includes cargo transportation demand data submitted by users related to cargo, and the vehicle dataset includes data related to vehicles;

[0035] A first calculation unit for performing normalization calculation on the data in the cargo transportation demand dataset and the data in the vehicle dataset to obtain a normalized cargo transportation demand dataset and a normalized vehicle dataset;

[0036] A feature extraction unit for extracting spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset, where the spatio-temporal features are used to characterize the spatio-temporal relationship between the cargo transportation demand and the vehicle;

[0037] A second calculation unit for calculating the multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatio-temporal features;

[0038] A demand clustering unit for clustering the data in the normalized cargo transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result;

[0039] A vehicle matching unit for performing carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles;

[0040] A route optimization unit for determining an optimal route according to the cargo transportation demands in the demand clustering result and the target vehicles.

[0041] In a third aspect, an embodiment of the present application further provides an electronic device, including at least one control processor and a memory for communicatively connecting with the at least one control processor; the memory stores instructions executable by the at least one control processor, and the instructions are executed by the at least one control processor to enable the at least one control processor to execute a carpool-based logistics transportation method as described above.

[0042] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions for causing a computer to execute a carpool-based logistics transportation method as described above.

[0043] It can be understood that the beneficial effects of the above second aspect to the fourth aspect compared with the related art are the same as those of the first aspect compared with the related art, and reference can be made to the relevant descriptions in the first aspect, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and / or additional aspects and advantages of the present application will become apparent and easy to understand from the description of the embodiments in conjunction with the following drawings, where:

[0045] Figure 1It is a schematic flowchart of an embodiment of the carpool-based logistics transportation method provided by this application;

[0046] Figure 2 It is a schematic flowchart of the logistics transportation method in the best embodiment of the carpool-based logistics transportation method provided by this application;

[0047] Figure 3 It is a schematic structural diagram of an embodiment of the carpool-based logistics transportation system provided by this application. Detailed Description of the Embodiment

[0048] The embodiments of this application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application and should not be construed as a limitation of this application.

[0049] In the description of this application, if the first, second, etc. are described, it is only for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0050] In the description of this application, it should be understood that for the orientation description, such as up, down, etc., the indicated orientation or positional relationship is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing this application and simplifying the description, rather than indicating or implying that the indicated device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0051] In the description of this application, it should be noted that unless otherwise clearly defined, words such as setting, installation, connection, etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in this application in combination with the specific content of the technical solution.

[0052] First, several nouns involved in this application are analyzed:

[0053] Carpooling: It can refer to multiple cargo demands sharing a transportation vehicle, and through optimizing the transportation route, achieving the efficient transportation of goods, similar to the concept of sharing rides in personnel transportation.

[0054] Freight demand: It can refer to the information about cargo transportation submitted by users, including cargo type, quantity, size, starting point, destination, and time requirements, etc.

[0055] User terminal: It can refer to the application or web interface where users submit freight demands and view the transportation status.

[0056] Back-end management system: It can be a system used to handle users' freight transportation demands, perform intelligent matching and scheduling management.

[0057] Driver app: It can be a mobile application for drivers to receive transportation tasks and view transportation routes.

[0058] Intelligent matching algorithm: It can be an algorithm that optimizes the matching of freight transportation demands and transportation resources through big data analysis and machine learning technologies, improving transportation efficiency.

[0059] Real-time monitoring: It can utilize GPS and Internet of Things technologies to conduct real-time tracking and monitoring of the transportation process to ensure transportation safety and efficiency.

[0060] Evaluation and feedback mechanism: It can be that after users complete the transportation of goods, they evaluate the transportation process and provide feedback information for system optimization and service quality improvement.

[0061] Generative AI: It can create new data, such as text, images, audio, etc., rather than just analyzing existing data.

[0062] Quantum Computing: It can more efficiently find the optimal carpooling plans and routes, optimizing route planning, vehicle scheduling, etc.

[0063] Augmented Reality (AR) and Virtual Reality (VR) technologies: They can be used in logistics transportation applications to achieve efficient training, real-time tracking, and immersive customer service, comprehensively improving logistics efficiency and customer satisfaction.

[0064] Internet of Things (IoT) sensor technology: It can be used to monitor the status of goods, environmental conditions, and vehicle performance in real time.

[0065] 5G communication technology: It can provide high-speed and low-latency data transmission to support real-time communication and monitoring.

[0066] In traditional logistics transportation methods, many vehicles operate without being fully loaded, resulting in high transportation costs. The vehicle empty load rate is high, and the resource utilization rate is low. Secondly, due to the lack of an effective demand matching mechanism, many transportation vehicles are not fully utilized, causing waste of transportation resources. Empty vehicles not only increase transportation costs but also cause unnecessary pollution to the environment. In traditional logistics transportation methods, the transportation routes of goods are often not optimized, resulting in longer transportation times and lower efficiency. At the same time, the goods of multiple users are transported separately, lacking intensive management, further reducing transportation efficiency. And due to the lack of real-time monitoring and feedback mechanisms, users have difficulty understanding the transportation status in a timely manner during the transportation of goods, lacking a sense of security and satisfaction. In addition, problems that may occur during transportation cannot be solved in a timely manner, affecting the user experience.

[0067] In summary, in the traditional logistics transportation mode, there are mainly problems of low utilization rate of vehicle resources and low logistics transportation efficiency.

[0068] To solve the above problems of low utilization rate of vehicle resources and low logistics transportation efficiency, this application proposes a logistics transportation method, system, electronic device and storage medium based on carpooling.

[0069] Refer to Figure 1 , an embodiment of this application provides a logistics transportation method based on carpooling. The method includes the following steps:

[0070] Step S100: Obtain a cargo transportation demand dataset and a vehicle dataset. Among them, the cargo transportation demand dataset contains cargo transportation demand data related to the cargo submitted by users, and the vehicle dataset contains data related to vehicles;

[0071] Step S200: Perform normalization calculations on the data in the cargo transportation demand dataset and the data in the vehicle dataset to obtain a normalized cargo transportation demand dataset and a normalized vehicle dataset;

[0072] Step S300: Extract the spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset. Among them, the spatio-temporal features are used to characterize the spatio-temporal relationship between the cargo transportation demand and the vehicle;

[0073] Step S400: Calculate the multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatio-temporal features;

[0074] Step S500: Cluster the data in the normalized cargo transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result;

[0075] Step S600: Perform carpooling matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles;

[0076] Step S700: Determine the optimal route according to the cargo transportation demands and target vehicles in the demand clustering result.

[0077] In this embodiment, by obtaining a cargo transportation demand dataset and a vehicle dataset, where the cargo transportation demand dataset contains cargo transportation demand data related to the cargo submitted by users, and the vehicle dataset contains data related to vehicles, normalization calculations are performed on the data in the cargo transportation demand dataset and the data in the vehicle dataset to obtain a normalized cargo transportation demand dataset and a normalized vehicle dataset. By normalizing the data, the effectiveness and rationality of the data can be ensured, and errors caused by static standardization can be avoided. Then, the spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset are extracted, where the spatio-temporal features are used to characterize the spatio-temporal relationship between the cargo transportation demand and the vehicle. By extracting the spatio-temporal relationship between the cargo transportation demand and the vehicle, a good data foundation is laid for later calculations. Then, according to the spatio-temporal features, the multi-dimensional weighted distance between the cargo transportation demand and the vehicle is calculated. According to the multi-dimensional weighted distance, the data in the normalized cargo transportation demand dataset are clustered to obtain a demand clustering result. By clustering a series of cargo transportation demands, a good data foundation is laid for later matching suitable vehicles. Carpooling matching is performed on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles. According to the cargo transportation demands and the target vehicles in the demand clustering result, the optimal route is determined. By matching vehicles for the clustered cargo transportation demands, the empty load rate can be reduced and the resource utilization rate can be improved, and the transportation route can be better optimized.

[0078] The above-mentioned obtaining of the cargo transportation demand dataset and the vehicle dataset can be based on the cargo transportation demand data related to the cargo submitted by users, including cargo type, cargo quantity, cargo size, cargo starting location, cargo destination, and time requirements, etc., and based on the data related to vehicles submitted by vehicle owners, including the current load of the vehicle, vehicle location, and vehicle estimated arrival time, etc., to screen out the required cargo transportation demand dataset and vehicle dataset.

[0079] The above-mentioned extraction of the spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset can be to automatically extract the spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset by using a spatio-temporal weighted convolutional neural network. This spatio-temporal weighted convolutional neural network combines an adaptive convolutional kernel (i.e., the weight of the convolutional kernel will be adjusted according to the change of the time step) and global spatio-temporal aggregation (i.e., global pooling is adopted) on the basis of a convolutional neural network (CNN). The specific network structure is not specifically described and limited in this embodiment.

[0080] Performing clustering on the data in the normalized cargo transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result may involve using a clustering method based on "adaptive density" and "adaptive centroid" to dynamically adjust the clustering according to multiple dimensions such as multi-dimensional weighted distance, cargo type, and time window, so as to obtain the demand clustering result. It is also possible to use existing K-means clustering algorithms and DBSCAN clustering algorithms in the prior art, and this embodiment does not make specific limitations.

[0081] Performing carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles may involve using a genetic algorithm to perform carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles.

[0082] Determining the optimal route according to the cargo transportation demands and target vehicles in the demand clustering result may involve generating multiple initial routes based on the cargo transportation demands and target vehicles in the demand clustering result; then using a path optimization algorithm to determine the optimal route for the multiple initial routes. The path optimization algorithm can be a heuristic algorithm, a genetic algorithm, etc., and this embodiment does not make specific limitations.

[0083] In some embodiments, performing normalization calculations on the data in the cargo transportation demand dataset and the data in the vehicle dataset to obtain a normalized cargo transportation demand dataset and a normalized vehicle dataset includes:

[0084]

[0085] where Norm(x i ) represents the normalized cargo transportation demand dataset, x i represents the i-th cargo transportation demand data, μ t,s represents the mean value in time t and space s, δ t represents the deviation value related to time, σ t,s represents the standard deviation in time t and space s, β represents an adjustment coefficient for controlling the data normalization effect, Norm(x j ) represents the normalized vehicle dataset, and x j represents the j-th vehicle data.

[0086] In this embodiment, by performing normalization calculations on the data in the cargo transportation demand dataset and the data in the vehicle dataset, the effectiveness and rationality of the data can be ensured, and errors caused by static standardization can be avoided. By dynamically detecting and correcting the error distribution and outliers, a good data foundation is laid for subsequent vehicle matching, thereby improving the accuracy of vehicle matching.

[0087] In some embodiments, spatio-temporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset are extracted, including:

[0088]

[0089] Among them, Feature ij represents the spatio-temporal feature between the i-th cargo transportation demand data and the j-th vehicle data extracted, W t,s,k represents the adaptive convolution kernel weight for the k-th time period t and spatial region s, * represents the convolution operator, Norm(x i ) represents the normalized cargo transportation demand dataset, Norm(x j ) represents the normalized vehicle dataset, Input represents the input, K represents the number of convolution kernels, T represents the time length considered, x t,s represents the cargo transportation demand data for the time period t and spatial region s, Pooling represents the global pooling operation, Pooling(x t,s ) represents the global pooling operation on the cargo transportation demand data for different time periods t and spatial regions s.

[0090] In this embodiment, by comprehensively considering the spatio-temporal information between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset, the spatio-temporal relationship between the cargo transportation demand and the vehicle is further extracted, laying a good data foundation for subsequent calculations.

[0091] In some embodiments, according to the spatio-temporal features, the multi-dimensional weighted distance between the cargo transportation demand and the vehicle is calculated, including:

[0092]

[0093] Among them, D ij represents the multi-dimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle, M represents the number of spatio-temporal features, w k represents the weighting coefficient for the k-th spatio-temporal feature, x ik represents the value of the i-th cargo transportation demand data on the k-th spatio-temporal feature, x jk represents the value of the j-th vehicle data on the k-th spatio-temporal feature, σ k represents the standard deviation of the k-th spatio-temporal feature, λ represents the weight coefficient, t i represents the time information of the i-th cargo transportation demand data, t j represents the time information of the j-th vehicle data, ΔT represents the time normalization coefficient.

[0094] In some embodiments, clustering the data in the normalized freight transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result, including:

[0095]

[0096] where G k represents the centroid of the k-th cluster, G k represents the set of freight transportation demand points of the k-th cluster, y i represents the feature vector of the i-th freight transportation demand point, D ij represents the multi-dimensional weighted distance between the i-th freight transportation demand and the j-th vehicle, θ k represents the adaptive threshold of the k-th cluster.

[0097] In this embodiment, the clustering method based on "adaptive density" and "adaptive centroid" can dynamically adjust the clustering according to multiple dimensions such as cargo characteristics, time windows, and transportation demands, obtaining a relatively good freight transportation demand clustering result, thereby optimizing the transportation route, reducing the vehicle empty load rate, and improving the utilization rate of vehicle resources.

[0098] In some embodiments, performing carpool matching on all freight transportation demands in each category of the demand clustering result to obtain target vehicles, including:

[0099] Constructing a target fitness function;

[0100] Based on the target fitness function, using a genetic algorithm to perform carpool matching on all freight transportation demands in each category of the demand clustering result to obtain target vehicles.

[0101] In this embodiment, by using the local search and global search mechanisms in the genetic algorithm, this optimization algorithm can quickly jump out of the local optimum and find the global optimum matching of vehicles and freight transportation demands, thereby improving the utilization rate of vehicle resources and reducing the vehicle empty load rate.

[0102] In some embodiments, constructing a target fitness function, including:

[0103]

[0104] where f fitness (x) represents the target fitness function, Weight(i) represents the weight of the i-th freight transportation demand, D ij (i, x) represents the multi-dimensional weighted distance between the i-th freight transportation demand and the j-th vehicle in the set x of vehicles that meet the constraint conditions, γ represents a penalty factor, Penalty(x) represents the penalty term for violating the constraint conditions, and N represents the total number of freight transportation demands in each category of the clustering result.

[0105] In this embodiment, the above-mentioned objective fitness function comprehensively considers the constraint conditions, combines the penalty term, can better conform to the vehicle matching, so as to obtain the optimal vehicle matching result, improve the vehicle resource utilization rate, and reduce the vehicle empty load rate.

[0106] For the convenience of those skilled in the art to understand, the following provides a set of best embodiments:

[0107] In the traditional logistics transportation mode, many vehicles operate without being fully loaded, resulting in high transportation costs. The vehicle empty load rate is high and the resource utilization rate is low. Secondly, due to the lack of an effective demand matching mechanism, many transport vehicles are not fully utilized, resulting in a waste of transport resources. Empty vehicles not only increase the transportation cost, but also cause unnecessary pollution to the environment. In the traditional logistics transportation mode, the transportation routes of goods are often not optimized, resulting in a long transportation time and low efficiency. At the same time, the goods of multiple users are transported separately, lacking intensive management, which further reduces the transportation efficiency. Due to the lack of real-time monitoring and feedback mechanism, it is difficult for users to understand the transportation status in a timely manner during the goods transportation process, lacking a sense of security and satisfaction. In addition, the problems that may occur during the transportation process cannot be solved in a timely manner, affecting the user experience.

[0108] To solve the above existing problems, this embodiment integrates the freight demands of multiple users, shares a transport vehicle, reduces the empty load rate, and significantly reduces the transportation cost. The vehicle runs fully loaded, and the resource utilization rate is greatly improved. The intelligent matching algorithm can optimize the matching of freight demand and transport resources (i.e., vehicles) according to factors such as cargo information, time requirements, and geographical location, improve the utilization rate of vehicles, and reduce resource waste. The carpooling logistics system can dynamically adjust the transportation route according to the real-time traffic conditions and cargo demands, shorten the transportation time, and improve the transportation efficiency. The goods of multiple users can achieve efficient intensive transportation through a reasonable carpooling plan. The system provides a real-time monitoring and feedback mechanism, and users can view the transportation status in real time through the application, and understand the current location and estimated arrival time of the goods. After the user completes the goods transportation, they can make evaluations and feedback, and the system optimizes the service quality according to the feedback, improving user satisfaction. And by reducing the operation of empty vehicles, the carbon emissions and environmental pollution during the transportation process are reduced, which has a positive significance for environmental protection.

[0109] This embodiment proposes a logistics transportation method based on carpooling, which realizes efficient carpooling transportation by integrating freight demand and transport resources. Refer to Figure 2 , the method of this embodiment includes the following contents:

[0110] Figure 2Among them, the regional-branch (PC side) part creates a carpooling order branch and stocks goods at the headquarters to obtain a created carpooling order. The supply chain (PC side) part accepts the carpooling order and judges whether the information in the carpooling order is within the configured list. If it is not within the carpooling list, it is shipped conventionally through the regional-branch (PC side) part. While waiting for departure, fill in the shipping information for each vehicle, start the vehicle through online operation. After successful departure, transport the goods. Wait for receipt after arriving at the destination. After successful receipt, it means the transportation is completed, and freight settlement and user evaluation can be carried out.

[0111] If it is judged that the information in the carpooling order is within the configured list, intelligent shipping is carried out. First, use the clustering algorithm to cluster the freight demand, then select a logistics company, give relevant information to the logistics company (such as filling in the expected delivery time and logistics cost), and push it to the logistics company through SMS notification. After receiving the notice, the logistics company matches the driver and the vehicle, uses the intelligent matching algorithm to match the vehicle for the clustered freight demand, and sends the vehicle dispatch information to the factory-supplier. The factory-supplier part conducts transportation calculation review. If the review is passed, the driver is notified to accept the order by SMS. If the review fails, the logistics company needs to re-dispatch the vehicle until the factory-supplier review is passed. If the driver refuses to accept the order or cancels the order before picking up the goods, the logistics company needs to re-dispatch the vehicle. If the driver successfully accepts the order, drive to the factory warehouse to pick up the goods. The vehicle will have its position judged by GPS, and then scan the code to enter the yard. If there are many vehicles, queuing is required (system queuing, large-screen device displays the queuing progress). After the goods are loaded, sign and upload the waybill, and the vehicle can leave the yard. Before the vehicle leaves the yard, the consignee will be notified by SMS that the goods have been loaded.

[0112] After the vehicle finishes picking up the goods, it shows that the goods are in transit. The transportation process is monitored in real time through GPS, and the optimal path is calculated in real time according to the road conditions using quantum computing until the goods are delivered. After the goods are delivered, unload the goods. After uploading the receiving note and inspection attachment together, it means the transportation is completed, and freight settlement and user evaluation can be carried out.

[0113] The specific details include the following steps:

[0114] 1. User-side freight demand submission module.

[0115] The user submits the freight demand through the mobile application or web interface, filling in detailed information such as the type, quantity, size, starting location, and destination of the goods.

[0116] 2. Back-end management system.

[0117] (1) Demand processing module: Receive and store the freight demand submitted by the user, and preprocess and classify the demand.

[0118] (2) Matching algorithm module: Using an intelligent matching algorithm, it matches suitable transportation vehicles and carpooling plans according to factors such as cargo information, time requirements, and geographical location. Specifically:

[0119] 1) Data preprocessing: Adaptive deviation normalization.

[0120] Collect the freight demand data and vehicle information submitted by users, and automatically adjust the preprocessing strategy according to the characteristics of the data distribution (such as the imbalance of demand data), so as to ensure the consistency of data in different regions, time periods, and vehicle resources in subsequent calculations. This method significantly improves the matching accuracy by dynamically detecting and correcting error distributions and outliers. The calculation formula for adaptive deviation normalization is:

[0121]

[0122]

[0123] Among them, Norm(x i ) represents the normalized cargo transportation demand data set, x i represents the i-th cargo transportation demand data, and this cargo transportation demand data can be the cargo weight, cargo volume, or timeliness. μ t,s represents the mean value at time t and space s, which represents the average value within this region (or time period). δ t represents the deviation value related to time (considering the fluctuation of timeliness requirements), σ t,s represents the standard deviation at time t and space s, β represents the adjustment coefficient for controlling the data normalization effect, Norm(x j ) represents the normalized vehicle data set, x j represents the j-th vehicle data, and this vehicle data can be the current load of the vehicle, vehicle location, and estimated arrival time of the vehicle, etc. exp represents the exponential function, represents the standardized squared difference, which represents the deviation degree of the current data point from the mean value. Here, the squared difference is used as the input value, and after exponential decay, the influence is reduced.

[0124] This step enables the normalization of carpooling data to adaptively adjust when dealing with time changes and demand fluctuations, thus ensuring the effectiveness and rationality of the data and avoiding errors caused by static standardization.

[0125] 2) Feature extraction: Spatiotemporal similarity weighted feature extraction.

[0126] Design an information extraction method that combines spatio-temporal similarity, and use a spatio-temporal weighted convolutional neural network to automatically extract the spatio-temporal features of regional demands to achieve the matching between vehicles and demands. This method uses a local adaptive convolutional kernel (LACN) and a global spatio-temporal aggregation (GTA) algorithm, and can extract fine features from multiple dimensions. Spatio-temporal weighted convolutional neural network (CNN).

[0127] The above spatio-temporal weighted convolutional neural network is a neural network based on the convolutional neural network (CNN). The convolutional neural network (CNN) is a deep learning algorithm, usually used for feature extraction of data such as images, videos, and audios. CNN gradually extracts features from low-level to high-level through convolutional operations, pooling operations, and fully connected layers. In traditional CNN, convolutional operations mainly focus on spatial information.

[0128] The spatio-temporal weighted convolutional neural network combines spatio-temporal feature extraction. Its core idea is to weight the traditional convolutional kernel, not only perform convolutional operations on spatial information, but also add sensitivity to data in the time dimension. This enhancement enables the model to process dynamic time-series data and adapt to changing spatio-temporal environments. The convolution and weighting process is as follows:

[0129] Spatial convolution: The same as traditional CNN, use a convolutional kernel to extract features in the spatial dimension.

[0130] Spatio-temporal weighting: For each moment in time-series data, the weight of the convolutional kernel is adjusted according to the change of time steps. This weight is dynamic and related to the data change within the time window.

[0131] The expression of the spatio-temporal weighted convolutional neural network is:

[0132]

[0133] Among them, Feature ij represents the spatio-temporal feature between the i-th cargo transportation demand data and the j-th vehicle data extracted, W t,s,k represents the adaptive convolutional kernel weight of the k-th time period t and spatial region s (considering spatio-temporal relationship), * represents the convolution operator, indicating that the convolutional kernel W t,s,k is convolved with the input data Input(Norm(x i ),Norm(x j ))), Norm(x i ) represents the normalized cargo transportation demand data set, Norm(x j ) represents the normalized vehicle data set, Input represents the input, K represents the number of convolutional kernels, indicating how many different convolutional kernels are used to extract features, T represents the considered time length (such as one day, one week, etc.), xt,s The freight transportation demand data representing the time period t and the spatial region s, and Pooling represents the global pooling operation. Pooling(x t,s ) represents performing the global pooling operation on the freight transportation demand data for different time periods t and spatial regions s, that is, performing weighted averaging on the demand information for different time periods t and spatial regions s.

[0134] This step combines the convolutional network with the global spatio-temporal information. Through multi-level information fusion, the spatio-temporal relationship between vehicle matching and demand features is extracted, providing a more accurate input for the subsequent steps.

[0135] 3) Distance calculation: Intelligent weighted similarity calculation.

[0136] A "multi-dimensional weighted distance function" is proposed. By adjusting the weights of different attributes, the optimal calculation method is automatically selected to more accurately consider the matching factors between demand and vehicle (such as time window, distance, load capacity, etc.). The distance calculation formula is:

[0137]

[0138] where D ij represents the multi-dimensional weighted distance between the i-th freight transportation demand and the j-th vehicle, M represents the number of spatio-temporal features, w k represents the weighted coefficient of the k-th spatio-temporal feature, x ik represents the value of the i-th freight transportation demand data on the k-th spatio-temporal feature, x jk represents the value of the j-th vehicle data on the k-th spatio-temporal feature, σ k represents the standard deviation of the k-th spatio-temporal feature, λ represents the weight coefficient, t i represents the time information of the i-th freight transportation demand data, t j represents the time information of the j-th vehicle data, and ΔT represents the time normalization coefficient, usually referring to the time span, which is used to normalize the time difference and avoid the time difference occupying too large a weight in the calculation.

[0139] This step makes the distance calculation more in line with the actual scenario by introducing multi-dimensional weighting and time difference adjustment, considering the complex associations between multiple factors.

[0140] 4) Demand clustering: Adaptive clustering method.

[0141] A clustering method based on "adaptive density" and "adaptive centroid" is proposed, which can dynamically adjust the clustering according to multiple dimensions such as cargo characteristics, time window, and transportation demand, so as to optimize the transportation route, reduce the empty load rate, and improve resource utilization rate. The clustering calculation formula is:

[0142]

[0143] Among them, G k represents the centroid of the k-th cluster, and G k represents the set of cargo transportation demand points of the k-th cluster. y i represents the feature vector of the i-th cargo transportation demand point, including information such as the spatial location, time window, and cargo type of the cargo transportation demand point. D ij represents the multi-dimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle. θ k represents the adaptive threshold of the k-th cluster, which is dynamically adjusted according to the data density and is not specifically limited in this embodiment.

[0144] This clustering method can be dynamically adjusted according to the changes in data, avoiding the problem of fixed number of clusters in traditional methods and having stronger adaptability.

[0145] 5) Vehicle matching and optimization: Global adaptive optimization algorithm.

[0146] Vehicle matching not only depends on simple distance and demand matching, but also needs to consider multiple constraints, such as the real-time availability of vehicles, capacity limits, time windows, etc. An adaptive genetic algorithm based on local priority and global optimization is proposed, which can quickly find the optimal match under complex constraints. Fitness function calculation formula:

[0147]

[0148] Among them, f fitness (x) represents the objective fitness function, which represents the matching quality, that is, the effect of the cargo transportation demand and vehicle matching. Weight(i) represents the weight of the i-th cargo transportation demand (indicating the relative importance of this demand). D ij (i, x) represents the multi-dimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle in the set x of vehicles that meet the constraint conditions. γ represents the penalty factor, which is used to penalize the matching scheme that violates the constraints. Penalty(x) represents the penalty term for violating the constraint conditions. N represents the total number of cargo transportation demands in each category of the clustering result.

[0149] By using the local search and global search mechanisms in the genetic algorithm, this optimization algorithm can quickly jump out of the local optimum and find the global optimum vehicle and cargo transportation demand matching.

[0150] (3) Scheduling management module: Schedules and manages the successfully matched transportation tasks, generates the optimal transportation route, and assigns the tasks to suitable drivers, focusing on task assignment and the generation of the preliminary route.

[0151] (4) Data Analysis Module: Analyze and monitor the data during the transportation process in real time to ensure safe and efficient transportation.

[0152] 3. Driver-side Task Receiving and Execution Module.

[0153] The driver receives transportation tasks through the mobile application, views detailed cargo information and the optimal transportation route, and transports according to the task requirements.

[0154] 4. Generative AI.

[0155] In the logistics transportation system, Generative AI (including ChatGPT, DALL-E) can be used for automatically generating transportation reports, customer feedback analysis, and predictive maintenance suggestions, etc. Applications: (1) Intelligent matching: Quickly and accurately match goods, considering travel needs, vehicle information, and real-time road conditions. (2) Route optimization: Generate the optimal driving route for the driver, comprehensively considering factors such as time, distance, and fuel consumption. (3) Personalized service: Provide personalized recommendations based on historical cargo data and preferences, such as driver selection, vehicle type, etc. (4) Voice interaction: Achieve smooth voice communication with the cargo owner, answer questions, and provide real-time information. (4) Safety monitoring: Real-time monitor the vehicle status and driver behavior, and give early warnings of potential safety hazards. (5) Operational optimization: Analyze operational data, optimize service areas, time periods, and matching algorithms to improve efficiency and satisfaction.

[0156] 5. Quantum Computing.

[0157] In the logistics transportation system, Quantum Computing can be used for optimizing transportation routes, predicting cargo demand, and improving logistics efficiency, etc. Applications: (1) Route optimization: Quickly find the optimal carpooling route to reduce driving time and costs. (2) Demand prediction: Predict cargo and driver demand by analyzing data to optimize capacity management. (3) Intelligent scheduling: Real-time monitor information, dynamically adjust matching and route strategies to improve service efficiency. (4) Logistics improvement: Optimize the logistics network, improve vehicle utilization rate, and reduce costs. (5) Security and privacy: Use quantum encryption technology to ensure information security and detect network threats.

[0158] 6. Augmented Reality (AR) and Virtual Reality (VR).

[0159] In the logistics transportation system, AR and VR can be used for training drivers, simulating transportation scenarios, providing real-time cargo tracking and visualization, etc. Applications: AR (1) Quick identification: Helps cargo owners and drivers quickly find each other and improve the efficiency of carpooling. (2) Cargo tracking: Displays the location and status of the cargo in real time, increasing transparency. (3) Driving assistance: Provides information such as road conditions and traffic signs, enhancing driving safety. Applications: VR (1) Driver training: Simulates the driving environment and improves driving skills. (2) Experience preview: Allows cargo owners to preview the carpooling route and increases trust. (3) Entertainment interaction: Provides an immersive entertainment experience and increases the fun of carpooling.

[0160] 7. Internet of Things (IoT) sensor technology.

[0161] It can monitor the status of goods such as temperature and humidity in real time to ensure the safety and quality of goods during transportation; monitor environmental conditions such as air pressure and light to provide precise environmental control for the transportation of special goods; monitor vehicle performance such as fuel consumption and tire pressure in real time to prevent potential failures and ensure transportation safety.

[0162] 8. 5G communication technology.

[0163] Provides ultra-high-speed data transmission and low-latency communication to ensure real-time navigation, vehicle monitoring and efficient scheduling, optimizing the user experience and driving safety. Applications: (1) Cargo tracking and monitoring: Utilizes the high-speed data transmission provided by 5G to achieve high-precision vehicle positioning and real-time navigation. (2) Seamless docking and communication: Vehicle and platform data are synchronized in real time, and cargo owners, platforms and drivers can communicate instantly. (3) Improve efficiency and safety: Automated management, reduce human errors, safety warnings, and ensure transportation safety.

[0164] 9. Real-time monitoring and feedback.

[0165] The system monitors the transportation process in real time, and users and drivers can view the transportation status through the application. After the transportation is completed, the user makes an evaluation and feedback, and the system optimizes and improves the service according to the feedback. Application scenarios: (1) After the goods are delivered, the system notifies the user to evaluate. The user logs in to the user terminal, finds the order in the "Completed" section, makes an evaluation based on service attitude, speed, safety, punctuality, etc., and can fill in written feedback. (2) The system provides evaluation options such as star ratings and satisfaction surveys. Users can upload photos or videos as supplementary evidence to comprehensively feedback on the transportation service.

[0166] 10. Combining dynamic freight settlement with smart contracts.

[0167] For the vehicles after intelligent matching and route optimization, the freight generated during the vehicle transportation process is calculated in the following way:

[0168] Dynamic Pricing Algorithm: Dynamically adjust freight based on real-time data (such as traffic flow, weather, delays, etc.). Multidimensional data (starting point, destination, type of goods, timeliness, etc.) in the transportation task generates a dynamic pricing coefficient through the model.

[0169]

[0170] Among them, F final represents the finally calculated freight, F base represents the base freight, which is based on the transportation distance and the quantity of goods, etc., and γ t represents the influence weight of each factor (such as traffic, time, weather, etc.) on the freight, and X t represents the value of the influencing factor, such as the degree of traffic congestion, weather conditions, etc.

[0171] Smart Contract Settlement: The smart contract can automatically trigger the freight settlement after all conditions are met, without manual intervention. The triggering conditions of various factors are preset in the contract. Once the conditions are met, the contract will automatically execute the payment and cost settlement.

[0172] Business Scenario Example: Suppose Company A needs to transport a batch of goods from Beijing to Shanghai, and the platform matches a suitable truck according to the goods demand. During the transportation process, due to traffic congestion on the way, the smart contract automatically calculates the corresponding delay cost and adds it to the total freight. After the transportation is completed, the smart contract will automatically calculate the final cost and record all relevant information (such as freight and delay cost, etc.) through the blockchain to ensure the transparency and accuracy of the whole process.

[0173] Form a closed loop for the above steps. The specific process is as follows:

[0174] 1. User Demand Submission and Intelligent Data Preprocessing.

[0175] (1) User Submits Demand: The user submits the freight demand through the platform (such as a mobile application or website), providing detailed goods information (such as type, quantity, size, starting location, destination, and timeliness requirements, etc.).

[0176] (2) Intelligent Data Preprocessing: The system uses the adaptive deviation normalization algorithm to dynamically adjust the preprocessing strategy according to the distribution characteristics of the data (such as the non-uniformity of the demand). This ensures the consistency of data in different regions, time periods, and vehicle resources in subsequent calculations, and avoids the impact of data anomalies on the matching results.

[0177] The system improves the data quality by dynamically detecting and correcting data anomalies and error distributions, thus ensuring the accuracy of subsequent matching.

[0178] 2. Based on the normalized data, intelligently generate a demand and carpooling matching plan.

[0179] (1) Intelligent matching algorithm: The system uses a multi-dimensional weighted distance function to score the matching degree between each demand and vehicle based on different attributes of the demand (such as cargo type, quantity, time requirement, geographical location, etc.). This algorithm can dynamically adjust the matching method according to the priority of the demand to ensure the selection of the most suitable vehicle for carpooling.

[0180] For example, when the demand involves timeliness requirements, the algorithm will increase the importance of the time window and preferentially match vehicles that can meet the timeliness requirements. At the same time, different matching parameters are weighted according to different cargo types (such as perishable goods, dangerous goods, etc.) to ensure transportation safety.

[0181] (2) Adaptive clustering method: The system dynamically identifies different demand groups through an adaptive clustering algorithm. Based on the clustering methods of adaptive density and adaptive centroid, it can intelligently adjust the number of clusters and the centroid position according to the data distribution and demand characteristics to achieve precise demand grouping.

[0182] The algorithm not only considers the similarity of geographical location and transportation demand, but also can further optimize clustering according to the special requirements of the cargo (such as transportation time window, special vehicle requirements, etc.) to ensure the optimality of the transportation plan.

[0183] (3) Dynamic route optimization: The system combines real-time traffic, weather and other information to generate the optimal route for each carpooling transportation task through a dynamic route optimization algorithm to ensure that the goods arrive at the destination on time and safely. By dynamically adjusting the route, the impact of factors such as congestion and bad weather on transportation efficiency is avoided.

[0184] 3. After vehicle matching and route optimization, transportation execution and real-time monitoring are carried out.

[0185] (1) Transportation task execution: After transportation starts, the vehicle executes the task according to the optimal carpooling plan. The driver drives according to the transportation route generated by the system. At the same time, the platform continues to track the transportation process.

[0186] (2) Real-time monitoring and feedback: The system collects transportation data (such as vehicle location, speed, temperature and humidity, etc.) in real time through Internet of Things devices and conducts real-time monitoring through the platform to ensure safety and efficiency during transportation. Users can view the status of the goods in real time to enhance the customer experience.

[0187] (3) Abnormality detection and automatic adjustment: When an abnormality occurs during transportation (such as sudden traffic conditions, weather changes, etc.), the system automatically adjusts the route or deploys standby vehicles to ensure the timely completion of the transportation task.

[0188] 4. Automatically settle the freight generated during vehicle transportation.

[0189] (1) Dynamic Pricing and Smart Contracts: After the transportation is completed, the system calculates the final transportation cost through a dynamic pricing algorithm based on factors such as transportation time, distance, and special requirements of the goods. The freight is automatically settled through a smart contract, ensuring transparent and automatic fund transfer among all participating parties (such as the platform, drivers, and users).

[0190] (2) The smart contract automatically calculates the cost and makes payments according to the preset terms and transportation data. At the same time, all transaction records are encrypted and stored by blockchain technology to ensure immutability and full transparency.

[0191] 5. User Feedback and Optimization Iteration.

[0192] (1) User Feedback Collection: After the transportation is completed, the platform collects user feedback on the transportation process, including transportation time, service quality, safety, etc.

[0193] (2) System Optimization: By analyzing user feedback and transportation data, intelligent algorithms continuously optimize links such as matching strategies, route planning, and freight settlement to improve the efficiency of the system and the user experience.

[0194] Compared with the prior art, the method of this embodiment has the following advantages:

[0195] 1. Improve the utilization rate of transportation resources, reduce the empty-haul rate, and reduce transportation costs.

[0196] 2. Optimize the transportation route, shorten the transportation time, and improve transportation efficiency.

[0197] 3. Provide convenient freight demand submission and real-time monitoring to enhance the user experience.

[0198] 4. Ensure the safety and efficiency of the transportation process through intelligent matching and real-time scheduling.

[0199] 5. Reduce the operation of empty vehicles, which helps to reduce carbon emissions during transportation and has a positive significance for environmental protection.

[0200] 6. Technological innovation, leading the future development of the logistics industry, integrating cutting-edge technologies such as generative artificial intelligence, quantum computing, AR / VR, etc., bringing innovative changes to the logistics industry and enhancing competitiveness.

[0201] Refer to Figure 3 , this embodiment of the present application also provides a carpool-based logistics transportation system, which includes a data acquisition unit 100, a first calculation unit 200, a feature extraction unit 300, a second calculation unit 400, a demand clustering unit 500, a vehicle matching unit 600, and a route optimization unit 700, where:

[0202] A data acquisition unit 100 is configured to acquire a cargo transportation demand dataset and a vehicle dataset. The cargo transportation demand dataset includes cargo transportation demand data related to cargo submitted by users, and the vehicle dataset includes data related to vehicles.

[0203] A first calculation unit 200 is configured to perform normalization calculations on the data in the cargo transportation demand dataset and the data in the vehicle dataset to obtain a normalized cargo transportation demand dataset and a normalized vehicle dataset.

[0204] A feature extraction unit 300 is configured to extract spatio-temporal features between each piece of data in the normalized cargo transportation demand dataset and the normalized vehicle dataset. The spatio-temporal features are used to characterize the spatio-temporal relationship between the cargo transportation demand and the vehicle.

[0205] A second calculation unit 400 is configured to calculate the multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatio-temporal features.

[0206] A demand clustering unit 500 is configured to cluster the data in the normalized cargo transportation demand dataset according to the multi-dimensional weighted distance to obtain a demand clustering result.

[0207] A vehicle matching unit 600 is configured to perform carpool matching on all cargo transportation demands in each category of the demand clustering result to obtain target vehicles.

[0208] A route optimization unit 700 is configured to determine an optimal route according to the cargo transportation demands and the target vehicles in the demand clustering result.

[0209] It should be noted that since a carpool-based logistics transportation system in this embodiment and the above-mentioned carpool-based logistics transportation method are based on the same inventive concept, the corresponding content in the method embodiment also applies to this system embodiment and will not be elaborated here.

[0210] An embodiment of the present application also provides an electronic device, including: at least one control processor and a memory communicatively connected to the at least one control processor.

[0211] The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include a high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0212] The non-transitory software program and instructions required to implement a carpool-based logistics transportation method of the above embodiments are stored in a memory, and when executed by a processor, implement a carpool-based logistics transportation method in the above embodiments. For example, execute the Figure 1 method steps S100 to S700 described above.

[0213] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0214] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed by one or more control processors, the one or more control processors can be caused to execute a carpool-based logistics transportation method in the above method embodiments. For example, execute the Figure 1 functions of method steps S100 to S700 described above.

[0215] Those of ordinary skill in the art can understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and their appropriate combinations. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, a communication medium typically includes computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium.

[0216] The above is a specific description of the preferred embodiments of the present application. However, the embodiments of the present application are not limited to the above-mentioned implementation manners. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the embodiments of the present application. These equivalent deformations or substitutions are all included within the scope defined by the claims of the embodiments of the present application.

[0217] The above has described the embodiments of the present application in detail with reference to the accompanying drawings. However, the present application is not limited to the above embodiments, and various changes can be made without departing from the purpose of the present application within the scope of knowledge possessed by those of ordinary skill in the art.

Claims

1. A logistics transportation method based on carpooling, characterized in that: The method comprises: Acquire a cargo transportation demand data set and a vehicle data set, wherein the cargo transportation demand data set includes cargo transportation demand data related to cargo submitted by a user, and the vehicle data set includes data related to vehicles; Performing normalization calculation on the data in the cargo transportation demand data set and the data in the vehicle data set to obtain a normalized cargo transportation demand data set and a normalized vehicle data set; Extracting the spatiotemporal features between each data in the normalized cargo transportation demand data set and the normalized vehicle data set, wherein the spatiotemporal features are used to characterize the spatiotemporal relationship between cargo transportation demand and vehicles; Calculating the multi-dimensional weighted distance between cargo transportation demand and vehicles according to the spatiotemporal characteristics; Clustering the data in the normalized cargo transportation demand data set according to the multi-dimensional weighted distance to obtain a demand clustering result; Carpooling and matching are performed for all cargo transportation demands of each category in the demand clustering results to obtain the target vehicle; An optimal route is determined according to the cargo transportation demand in the demand clustering result and the target vehicle.

2. The logistics transportation method based on carpooling according to claim 1 is characterized in that: The normalizing calculation of the data in the cargo transportation demand data set and the data in the vehicle data set to obtain a normalized cargo transportation demand data set and a normalized vehicle data set includes: Among them, Norm(x i ) represents the normalized cargo transportation demand dataset, x i represents the i-th cargo transportation demand data, μ t,s represents the mean over time t and space s, δ t represents the time-dependent deviation value, σ t,s represents the standard deviation in time t and space s, β represents the adjustment coefficient for controlling the normalization effect of data, Norm(x j ) represents the normalized vehicle dataset, x j Represents the jth vehicle data.

3. The logistics transportation method based on carpooling according to claim 1 is characterized in that: The extracting the spatiotemporal features between each data in the normalized cargo transportation demand dataset and the normalized vehicle dataset includes: Among them, Feature ij represents the spatiotemporal characteristics between the extracted i-th cargo transportation demand data and the j-th vehicle data, W t，s，k represents the adaptive convolution kernel weight of the kth time period t and spatial region s, * represents the convolution operator, and Norm(x i ) represents the normalized cargo transportation demand dataset, Norm(x j ) represents the normalized vehicle dataset, Input represents the input, K represents the number of convolution kernels, T represents the length of time considered, and x t，s represents the cargo transportation demand data in time period t and spatial region s, Pooling represents the global pooling operation, Pooling(x t,s ) represents the global pooling operation of cargo transportation demand data in different time periods t and spatial regions s.

4. The logistics transportation method based on carpooling according to claim 1 is characterized in that: The calculating, according to the spatiotemporal characteristics, the multi-dimensional weighted distance between the cargo transportation demand and the vehicle comprises: Among them, D ij represents the multidimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle, M represents the number of spatiotemporal features, and w k represents the weight coefficient of the kth spatiotemporal feature, x ik represents the value of the i-th cargo transportation demand data at the k-th spatiotemporal feature, x jk represents the value of the jth vehicle data on the kth spatiotemporal feature, σ k represents the standard deviation of the kth spatiotemporal feature, λ represents the weight coefficient, and t i represents the time information of the i-th cargo transportation demand data, t j represents the time information of the j-th vehicle data, and ΔT represents the normalization coefficient of time.

5. The logistics transportation method based on carpooling according to claim 1 is characterized in that: The step of clustering the data in the normalized cargo transportation demand data set according to the multi-dimensional weighted distance to obtain a demand clustering result includes: Among them, G k represents the center of gravity of the kth cluster, G k represents the set of cargo transportation demand points of the kth cluster, y i represents the characteristic vector of the i-th cargo transportation demand point, D ij represents the multidimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle, θ k represents the adaptive threshold of the k-th cluster.

6. The logistics transportation method based on carpooling according to claim 1 is characterized in that: The method of performing carpooling matching on all cargo transportation demands of each category in the demand clustering results to obtain a target vehicle includes: Construct the target fitness function; Based on the target fitness function, a genetic algorithm is used to perform carpooling matching on all cargo transportation demands of each category in the demand clustering results to obtain a target vehicle.

7. The logistics transportation method based on carpooling according to claim 6 is characterized in that: The constructing of the target fitness function comprises: Among them, f fitness (x) represents the target fitness function, Weight(i) represents the weight of the i-th cargo transportation demand, D ij (i, x) represents the multidimensional weighted distance between the i-th cargo transportation demand and the j-th vehicle in the vehicle set x that meets the constraint conditions, γ represents the penalty factor, Penalty(x) represents the penalty item for violating the constraint conditions, and N represents the total number of cargo transportation demands in each category in the clustering results.

8. A logistics transportation system based on carpooling, characterized in that: The system comprises: A data acquisition unit, used to acquire a cargo transportation demand data set and a vehicle data set, wherein the cargo transportation demand data set includes cargo transportation demand data related to cargo submitted by a user, and the vehicle data set includes data related to vehicles; A first calculation unit, configured to perform normalization calculation on the data in the cargo transportation demand data set and the data in the vehicle data set to obtain a normalized cargo transportation demand data set and a normalized vehicle data set; A feature extraction unit, used to extract the spatiotemporal features between each data in the normalized cargo transportation demand data set and the normalized vehicle data set, wherein the spatiotemporal features are used to characterize the spatiotemporal relationship between the cargo transportation demand and the vehicle; A second calculation unit is used to calculate the multi-dimensional weighted distance between the cargo transportation demand and the vehicle according to the spatiotemporal characteristics; A demand clustering unit, configured to cluster the data in the normalized cargo transportation demand data set according to the multi-dimensional weighted distance to obtain a demand clustering result; The vehicle matching unit is used to match all cargo transportation demands of each category in the demand clustering results to obtain the target vehicle; The route optimization unit is used to determine the optimal route according to the cargo transportation demand in the demand clustering result and the target vehicle.

9. An electronic device, characterized in that: It includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor to enable the at least one control processor to execute the carpooling-based logistics transportation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the carpooling-based logistics transportation method as described in any one of claims 1 to 7.

Images