A logistics solution intelligent design system and method

Abstract

The present invention proposes a logistics solution intelligent design system and method, the method includes real-time acquisition of first logistics facility data, first transportation tool data, and first cargo data; using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data; based on historical logistics data and based on big data training, respectively constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and anomaly detection and early warning models; based on current logistics dynamic data and pending logistics order data, combining the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, generating a logistics solution for the pending logistics order. The solution of the present invention realizes the intelligence of various logistics decision-making links by combining artificial intelligence algorithm technology with big data, and can provide efficient and globally optimal logistics solutions.

Description

Technical Field

[0001] The present invention relates to the field of logistics technology, and in particular to a logistics solution intelligent design system and method. Background Art

[0002] Logistics is a cross-industry, cross-department, cross-regional, and highly permeable complex industry. With the upgrading of the industry and the development of productivity, modern logistics has an increasingly urgent demand for real-time, all-weather, continuous, punctual, and high-precision. In the case of complex and diverse transportation networks, the planning difficulty and selectivity of logistics solutions have also increased significantly. However, the existing logistics solution design system is not intelligent, efficient, and accurate enough. Therefore, how to design logistics solutions intelligently, quickly, and accurately is an urgent problem to be solved in modern logistics. Summary of the invention

[0003] Based on the above problems, the present invention proposes a logistics solution intelligent design system and method, which realizes the intelligence of various logistics decision-making links by combining artificial intelligence algorithm technology with big data, and can provide efficient and globally optimal logistics solutions.

[0004] In view of this, one aspect of the present invention proposes a logistics solution intelligent design system, comprising: a cloud server, an Internet of Things edge server communicatively connected to the cloud server, a first intelligent sensing terminal group arranged in a logistics facility, a second intelligent sensing terminal group arranged in a transportation vehicle, and a third intelligent sensing terminal group arranged in a cargo, wherein the Internet of Things edge server is communicatively connected to the first intelligent sensing terminal group, the second intelligent sensing terminal group, and the third intelligent sensing terminal group;

[0005] The cloud server is configured as follows:

[0006] The first intelligent sensing terminal group acquires first logistics facility data in real time, the second intelligent sensing terminal group acquires first transportation tool data in real time, and the third intelligent sensing terminal group acquires first cargo data in real time;

[0007] Using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data;

[0008] Based on historical logistics data and big data training, we build logistics demand forecasting models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models.

[0009] Based on the current logistics dynamic data and the logistics order data to be processed, combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, a logistics plan for the logistics order to be processed is generated.

[0010] Another aspect of the present invention provides a method for intelligently designing a logistics solution, comprising:

[0011] The first intelligent sensing terminal group set at the logistics facility obtains the first logistics facility data in real time, the second intelligent sensing terminal group set at the transportation tool obtains the first transportation tool data in real time, and the third intelligent sensing terminal group set at the cargo obtains the first cargo data in real time;

[0012] The cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data;

[0013] Based on historical logistics data and big data training, we build logistics demand forecasting models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models.

[0014] Based on the current logistics dynamic data and the logistics order data to be processed, combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, a logistics plan for the logistics order to be processed is generated.

[0015] Optionally, the step of acquiring first logistics facility data in real time by a first intelligent sensing terminal group arranged at a logistics facility, acquiring first transportation tool data in real time by a second intelligent sensing terminal group arranged at a transportation tool, and acquiring first cargo data in real time by a third intelligent sensing terminal group arranged at the cargo includes:

[0016] Acquire first attribute data, first status data and first environment data of the logistics facility;

[0017] Acquiring second attribute data, second state data, and second environment data of the transportation tool;

[0018] Acquiring third attribute data, third state data, and third environment data of the goods;

[0019] installing the first intelligent sensing terminal group in a first key area of ​​the logistics facility according to the first attribute data, the first status data, and the first environmental data;

[0020] Installing the second intelligent sensing terminal group in a second key area of ​​the transportation vehicle according to the first attribute data, the first state data and the first environment data;

[0021] The third intelligent sensing terminal group is installed in a third key area of ​​the cargo according to the third attribute data, the third state data and the third environment data.

[0022] Optionally, the cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data and the first cargo data to obtain current logistics dynamic data, including:

[0023] The cloud server uses data mining algorithms to analyze historical logistics facility data, historical transportation tool data, and historical cargo data, and constructs descriptive analysis models, diagnostic analysis models, and predictive analysis models;

[0024] The cloud server uses a big data processing framework to clean, deduplicate, and convert the first logistics facility data, the first transportation tool data, and the first cargo data, and converts the three types of heterogeneous data into input data that conforms to the input format of the descriptive analysis model, the diagnostic analysis model, and the predictive analysis model through data fusion technology;

[0025] The input data is input into the descriptive analysis model, the diagnostic analysis model and the predictive analysis model, and the analysis results are integrated to obtain current logistics dynamic data.

[0026] Optionally, the steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0027] Extracting historical logistics order data, historical cargo data, historical logistics facility data, and historical transportation tool data from the historical logistics data as a first basic data set;

[0028] Cleaning, format conversion, missing value processing, and outlier processing of the data in the first basic data set;

[0029] Fusing and associating the data in the first basic data set to construct a unified first training data set;

[0030] Extracting features of the dimensions of time, location, cargo attributes, and environmental factors from the first training data set as first features;

[0031] Adopting the time series forecasting model, an order forecasting model is trained based on the historical order quantity sequence to predict future order demand;

[0032] Using a machine learning model, combined with the first feature, to train a transportation demand prediction model for predicting transportation demand;

[0033] Constructing a multi-model integration framework, fusing the order prediction model and the transportation demand prediction model to obtain a first fusion model;

[0034] Splitting the first training data set into a first training set, a first validation set, and a first test set;

[0035] Training the first fusion model on the first training set, adjusting hyperparameters on the first validation set, and optimizing the structure of the first fusion model;

[0036] The optimized first fusion model is evaluated on the first test set, and the first fusion model is adjusted according to the evaluation result to obtain the logistics demand prediction model.

[0037] Optionally, the steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0038] extracting historical transportation route data from the historical logistics data;

[0039] After establishing a first association relationship between the historical logistics order data, the historical goods data, the historical logistics facility data, the historical transportation tool data and the historical transportation route data, the historical logistics order data are used as a second training data set;

[0040] constructing a first modeling algorithm for performing path planning;

[0041] Splitting the second training data set into a second training set, a second validation set, and a second test set;

[0042] Training the first modeling algorithm on the second training set to obtain a basic path planning model;

[0043] Adjusting hyperparameters on the second validation set to optimize the structure of the basic path planning model;

[0044] The optimized basic path planning model is evaluated on the second test set, and the basic path planning model is adjusted according to the evaluation result to obtain the logistics route planning model.

[0045] Optionally, the steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0046] Determine the first optimization goal of logistics and transportation;

[0047] Construct the first constraint;

[0048] Constructing a first optimization model according to the first optimization objective and the first constraint condition;

[0049] Generate a third training data set and a third test data set based on the historical logistics order data, the historical transportation tool data, the historical logistics facility data, and the historical route data;

[0050] Using the third training data set to train the first optimization model to obtain a basic transportation resource scheduling model;

[0051] The basic transportation resource scheduling model is evaluated on the third test data set, and the basic transportation resource scheduling model is adjusted according to the evaluation result to obtain the transportation resource scheduling model.

[0052] Optionally, the steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0053] Extracting historical status data in the historical transportation process from the historical logistics data;

[0054] Performing data cleaning, format conversion, missing value processing, and relationship analysis on the historical logistics order data, the historical goods data, the historical logistics facility data, the historical transportation tool data, the historical transportation route data, and the historical status data to construct a fourth data set;

[0055] Marking known abnormal cases in the fourth data set as supervised learning training samples;

[0056] extracting abnormality-related feature data related to the abnormal case from the fourth data set;

[0057] Build basic anomaly detection and early warning models;

[0058] The basic anomaly detection and early warning model is trained and tested using the data of the abnormal cases and the abnormality-related feature data to obtain the anomaly detection and early warning model.

[0059] Optionally, the step of generating a logistics plan for a logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, in combination with the logistics demand forecasting model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, includes:

[0060] Extracting key order information from the logistics order data to be processed;

[0061] Using natural language processing technology to structure the key information of the order;

[0062] Combining the key order information with the current logistics dynamic data, and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed;

[0063] Generate alternative route plans using the logistics route planning model according to the starting and ending locations, time windows, etc.;

[0064] Conduct multi-dimensional evaluation of each alternative route plan in terms of timeliness, cost, and abnormal risk;

[0065] Inputting the alternative route plan and the transportation demand into the transportation resource scheduling model to generate a vehicle scheduling plan that meets the transportation demand;

[0066] Based on the current logistics dynamic data, the abnormal risk of the alternative plan is analyzed using the abnormal detection and early warning model to obtain an abnormal analysis result, and the logistics plan for the logistics order to be processed is determined according to the abnormal analysis result.

[0067] Optionally, the step of combining the key order information with the current logistics dynamic data and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed includes:

[0068] Collect and extract current inventory data, current transportation capacity data, current weather data, and current traffic data from the current logistics dynamic data;

[0069] The current inventory data, the current transportation capacity data, the current weather data, the current traffic data and the key order information are integrated to construct a complete input feature vector;

[0070] The input feature vector is input into the logistics demand prediction model to obtain the corresponding transportation demand prediction value, which is the transportation demand corresponding to the logistics order to be processed.

[0071] According to the technical solution of the present invention, the intelligent design method of logistics scheme includes obtaining the first logistics facility data in real time through the first intelligent sensing terminal group set in the logistics facility, obtaining the first transportation tool data in real time through the second intelligent sensing terminal group set in the transportation tool, and obtaining the first cargo data in real time through the third intelligent sensing terminal group set in the cargo; the cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data and the first cargo data to obtain the current logistics dynamic data; according to the historical logistics data, based on big data training, respectively construct a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and anomaly detection and early warning model; according to the current logistics dynamic data and the logistics order data to be processed, the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model are combined to generate a logistics scheme for the logistics order to be processed. The scheme of the present invention realizes the intelligence of various logistics decision-making links by combining artificial intelligence algorithm technology with big data, and can provide efficient and globally optimal logistics schemes. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 is a schematic block diagram of a logistics solution intelligent design system provided by an embodiment of the present invention;

[0073] Figure 2 It is a flow chart of a method for intelligent design of logistics solutions provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0075] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0076] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0077] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0078] Refer to the following Figure 1 to Figure 2 To describe a logistics solution intelligent design system and method provided according to some embodiments of the present invention.

[0079] like Figure 1 As shown, one embodiment of the present invention provides a logistics solution intelligent design system, including: a cloud server, an Internet of Things edge server communicatively connected to the cloud server, a first intelligent sensing terminal group arranged in a logistics facility, a second intelligent sensing terminal group arranged in a transportation vehicle, and a third intelligent sensing terminal group arranged in a cargo, wherein the Internet of Things edge server is communicatively connected to the first intelligent sensing terminal group, the second intelligent sensing terminal group, and the third intelligent sensing terminal group;

[0080] The cloud server is configured as follows:

[0081] The first intelligent sensing terminal group acquires the first logistics facility data in real time, the second intelligent sensing terminal group acquires the first transportation tool data in real time, and the third intelligent sensing terminal group acquires the first cargo data in real time (the first or second or third intelligent sensing terminal group includes but is not limited to intelligent terminals such as sensors, RFID, and cameras);

[0082] Using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data (i.e., the three as a whole are logistics dynamic data) to obtain current logistics dynamic data;

[0083] Based on historical logistics data (including historical logistics order data, historical cargo data, historical logistics facility data, historical transportation tool data, historical transportation route data, historical status data in the historical transportation process, etc.), based on big data training, we build logistics demand prediction models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models.

[0084] Based on the current logistics dynamic data and the logistics order data to be processed (including the transportation object, starting location, transportation time, transportation timeliness, transportation cost, etc.), combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, a logistics plan for the logistics order to be processed is generated.

[0085] Using the technical solution of this embodiment, the intelligent design method of logistics plan includes obtaining the first logistics facility data in real time through the first intelligent sensing terminal group set in the logistics facility, obtaining the first transportation tool data in real time through the second intelligent sensing terminal group set in the transportation tool, and obtaining the first cargo data in real time through the third intelligent sensing terminal group set in the cargo; the cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data and the first cargo data to obtain the current logistics dynamic data; according to the historical logistics data, based on big data training, respectively construct a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and anomaly detection and early warning model; according to the current logistics dynamic data and the logistics order data to be processed, combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, generate a logistics plan for the logistics order to be processed. The scheme of the present invention combines artificial intelligence algorithm technology with big data to realize the intelligence of various logistics decision-making links, and can provide efficient and globally optimal logistics solutions.

[0086] It should be known that Figure 1 The block diagram of the logistics solution intelligent design system is shown for illustration only, and the number of modules shown therein does not limit the protection scope of the present invention.

[0087] See also Figure 2 Another embodiment of the present invention provides a method for intelligently designing a logistics solution, comprising:

[0088] The first logistics facility data is obtained in real time by a first intelligent sensing terminal group installed in the logistics facility, the first transportation tool data is obtained in real time by a second intelligent sensing terminal group installed in the transportation tool, and the first cargo data is obtained in real time by a third intelligent sensing terminal group installed in the cargo (the first or second or third intelligent sensing terminal group includes but is not limited to intelligent terminals such as sensors, RFID, and cameras);

[0089] The cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data (that is, the three as a whole are logistics dynamic data) to obtain current logistics dynamic data;

[0090] Based on historical logistics data (including historical logistics order data, historical cargo data, historical logistics facility data, historical transportation tool data, historical transportation route data, historical status data in the historical transportation process, etc.), based on big data training, we build logistics demand prediction models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models.

[0091] Based on the current logistics dynamic data and the logistics order data to be processed (including the transportation object, starting location, transportation time, transportation timeliness, transportation cost, etc.), combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, a logistics plan for the logistics order to be processed is generated.

[0092] In this embodiment, it also includes: developing an automated execution system, and applying automated warehouses, unmanned guided vehicles, automatic loading and unloading systems and other equipment to automate and intelligentize the logistics execution process; establishing a unified information platform to integrate information from various logistics companies and departments to achieve information interaction and sharing and global collaborative optimization; applying encryption technologies such as blockchain to improve system data security; building redundant backups to improve system robustness; establishing real-time monitoring and closed-loop control mechanisms to continuously optimize logistics plans and reduce overall logistics costs.

[0093] In this embodiment, logistics facilities generally include but are not limited to: storage facilities (ordinary warehouses, automated warehouses, cold storage, temporary storage points / transfer stations, etc.), loading and unloading facilities (loading and unloading platforms, conveyor belts, pallet forklifts, automatic guided vehicles (AGVs), bridge cranes, elevators, etc.), packaging facilities (packaging machines, strapping machines, label printers, etc.), transportation tools (trucks, trains, ships, aircraft, etc.), logistics parks (logistics centers, logistics bases, distribution centers, etc.), information systems (logistics management systems (WMS), transportation management systems (TMS), freight tracking systems, etc.), other auxiliary facilities (weighing equipment, monitoring equipment, road indicator lights, etc.), etc. The above facilities can all be connected to the Internet of Things to realize the automation, intelligence and visualization of logistics processes, thereby supporting efficient logistics operations.

[0094] In this embodiment, by combining artificial intelligence algorithm technology with big data, the intelligence of various logistics decision-making links is realized, and efficient and globally optimal logistics solutions can be provided.

[0095] In some possible implementations of the present invention, the step of acquiring first logistics facility data in real time by a first intelligent sensing terminal group arranged in a logistics facility, acquiring first transportation tool data in real time by a second intelligent sensing terminal group arranged in a transportation tool, and acquiring first cargo data in real time by a third intelligent sensing terminal group arranged in cargo includes:

[0096] Acquire first attribute data, first status data and first environment data of the logistics facility;

[0097] Acquire second attribute data, second state data and second environment data (including historical route data, historical route environment data, etc.) of the transportation tool;

[0098] Acquiring third attribute data, third state data, and third environment data of the goods;

[0099] The first intelligent sensing terminal group is installed in the first key area of ​​the logistics facility according to the first attribute data, the first state data and the first environmental data, specifically including: installing intelligent cameras in key areas of warehouses and cargo yards to monitor the storage and operation areas of goods in real time; installing visual sensors at loading and unloading conveyor lines, shelves, etc. to detect the flow direction and inventory of goods; installing RFID readers at entrances and exits to identify and record the entry and exit of vehicles and goods; deploying environmental monitoring sensors in warehouses, operation areas, etc. to detect environmental parameters such as temperature and humidity, air quality, noise, etc. These first intelligent sensing terminals (sensors) are connected to the local control system or other intelligent sensing terminals or servers or edge nodes of the logistics facility through wired or wireless means to upload and collect data in real time.

[0100] The second intelligent sensing terminal group is installed in the second key area of ​​the transportation vehicle according to the first attribute data, the first status data and the first environmental data, specifically including: installing RFID readers and environmental sensors in transportation compartments such as trucks and trains to identify and monitor the loaded cargo information; installing GPS positioning systems, driving recorders, etc. on vehicles to track transportation routes, driving speeds, driving behaviors and other data; installing dedicated temperature, pressure and other environmental sensing devices for special vehicles such as refrigerated trucks and hazardous chemicals trucks; the terminal equipment uploads the collected data to the transportation management platform or other intelligent sensing terminals or cloud servers or edge nodes in real time through the on-board wireless transmission module.

[0101] The third intelligent sensing terminal group is installed in the third key area of ​​the goods according to the third attribute data, the third status data and the third environmental data, specifically including: pasting an RFID electronic tag on the outer packaging of the goods to uniquely identify the ID information of each piece of goods; installing vibration, displacement and other sensors on fragile and valuable goods to monitor the impact of vibration during transportation; for special goods such as refrigerated and living objects, special sensors (third intelligent sensing terminals) such as temperature, humidity, and vital signs are implanted. The sensing terminal is equipped with a self-powered power supply and a wireless communication module, which can transmit the real-time status data of the goods to other intelligent sensing terminals or cloud servers or edge nodes.

[0102] In this embodiment, the above three types of intelligent sensing terminal groups can be interconnected through the Internet of Things technology to jointly build a comprehensive perception system for the entire logistics process. The cloud-based Internet of Things platform (or cloud server) will summarize and analyze all data to form a comprehensive real-time logistics monitoring and decision support.

[0103] In some possible implementations of the present invention, the cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data, including:

[0104] The cloud server uses data mining algorithms (such as machine learning, statistical analysis, etc.) to analyze historical logistics facility data, historical transportation tool data, and historical cargo data, and builds descriptive analysis models (including: association analysis, i.e., discovering association patterns between cargo, facilities, and transportation tools; path analysis, i.e., analyzing the flow paths and patterns of cargo in the logistics network; etc.), diagnostic analysis models (such as anomaly detection: discovering anomalies in cargo status, facility status, and transportation behavior based on the model), and predictive analysis models (such as time series prediction: predicting freight demand, inventory, and transportation capacity demand for a period of time in the future);

[0105] The cloud server uses a big data processing framework (such as Apache Spark) to clean, deduplicate, and convert the first logistics facility data, the first transportation tool data, and the first cargo data, and converts the three types of heterogeneous data into input data that conforms to the input format of the descriptive analysis model, the diagnostic analysis model, and the predictive analysis model through data fusion technology;

[0106] The input data is input into the descriptive analysis model, the diagnostic analysis model and the predictive analysis model, and the analysis results are integrated to obtain current logistics dynamic data.

[0107] In this embodiment, through the above-mentioned big data analysis technology, comprehensive logistics dynamic data can be extracted from multi-dimensional heterogeneous data, including overall operating conditions, potential problems, future trends, etc., to provide data support for formulating logistics intelligent decision-making solutions.

[0108] In some possible implementations of the present invention, the steps of constructing a logistics demand forecasting model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0109] Extracting historical logistics order data, historical cargo data, historical logistics facility data, and historical transportation tool data from the historical logistics data as a first basic data set;

[0110] Cleaning, format conversion, missing value processing, and outlier processing of the data in the first basic data set;

[0111] Fusing and associating the data in the first basic data set to construct a unified first training data set;

[0112] Extracting features of the dimensions of time, location, cargo attributes, and environmental factors from the first training data set as first features;

[0113] Use time series forecasting models (such as ARIMA, Prophet, etc.) to train an order forecasting model to predict future order demand based on historical order quantity sequences;

[0114] Using a machine learning model (such as XGBoost or a deep learning model) and combining the first feature to train a transportation demand prediction model for predicting transportation demand;

[0115] Constructing a multi-model integration framework, fusing the order prediction model and the transportation demand prediction model to obtain a first fusion model;

[0116] Splitting the first training data set into a first training set, a first validation set, and a first test set;

[0117] Training the first fusion model on the first training set, adjusting hyperparameters on the first validation set, and optimizing the structure of the first fusion model;

[0118] The optimized first fusion model is evaluated on the first test set, and the first fusion model is adjusted according to the evaluation result to obtain the logistics demand prediction model.

[0119] In this embodiment, through the above process, effective patterns and rules can be extracted from the multi-source heterogeneous big data within the enterprise, and an accurate logistics demand forecasting model can be constructed to provide important data support for the intelligent decision-making of the logistics system. The implementation process needs to deal with multiple challenges such as data quality, computing resources, and online and offline consistency.

[0120] In some possible implementations of the present invention, the steps of constructing a logistics demand forecasting model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0121] Extracting historical transportation route data (such as road, weather, traffic conditions, etc.) from the historical logistics data;

[0122] After establishing a first association relationship among the historical logistics order data (such as starting point, end point, time window, etc.), the historical cargo data (including cargo attribute data, such as weight, volume, special requirements, etc.), the historical logistics facility data (such as warehouse location, loading and unloading efficiency, etc.), the historical transportation tool data (such as vehicle type, load, fuel consumption, etc.) and the historical transportation route data, use them as the second training data set (standardize and encode the geographic location data in the historical transportation route data, construct the path and network topology structure, and extract the path features related to path planning);

[0123] Constructing a first modeling algorithm for path planning (such as an ant colony algorithm, a genetic algorithm, a machine learning algorithm, a deep learning algorithm);

[0124] Splitting the second training data set into a second training set, a second validation set, and a second test set;

[0125] Training the first modeling algorithm on the second training set to obtain a basic path planning model;

[0126] Adjusting hyperparameters on the second validation set to optimize the structure of the basic path planning model;

[0127] The optimized basic path planning model is evaluated on the second test set, and the basic path planning model is adjusted according to the evaluation result to obtain the logistics route planning model.

[0128] This embodiment also includes: deploying the model to the logistics scheduling decision system; introducing online learning, using new route data for continuous optimization, regular retraining, and updating and deploying new models; using machine learning to predict influencing factors such as traffic flow and weather; integrating map data and road network constraints; and designing rules and scoring functions that conform to reality.

[0129] In this embodiment, the model output is the optimal transportation route solution under given orders and resource conditions. At the same time, it is also necessary to solve complex problems such as route constraints and multi-order loading, which can be solved through model integration, combined optimization and other means. The ultimate goal is to improve the intelligence level of route planning and reduce transportation costs and time.

[0130] In some possible implementations of the present invention, the steps of constructing a logistics demand forecasting model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0131] Determine the first optimization goal of logistics transportation (such as minimizing transportation costs, maximizing vehicle utilization, etc.);

[0132] Construct the first constraint conditions (such as time window constraints, vehicle load constraints, service level constraints, etc.);

[0133] According to the first optimization objective and the first constraint condition, construct a first optimization model (such as integer programming model: suitable for small-scale scheduling problems; continuous optimization model: large-scale, approximate optimization; constraint programming model: processing complex rules and restrictions; heuristic algorithm: such as population algorithm, simulated annealing, etc.);

[0134] Generate a third training data set and a third test data set based on the historical logistics order data (including cargo information, time window, service level requirements, etc.), the historical transportation tool data (including vehicle type, load, fuel consumption, location, etc.), the historical logistics facility data (such as warehouse location, unloading efficiency, etc.), and the historical route data (including road conditions, weather, traffic conditions, etc.);

[0135] Using the third training data set to train the first optimization model (for precise models such as integer programming, using an optimization solver to solve; for approximate models such as heuristic algorithms, using a big data cluster to train), to obtain a basic capacity resource scheduling model;

[0136] The basic transportation resource scheduling model is evaluated on the third test data set, and the basic transportation resource scheduling model is adjusted according to the evaluation result to obtain the transportation resource scheduling model.

[0137] This embodiment also includes: deploying the model to the scheduling decision system to provide real-time scheduling solutions; introducing an online learning mechanism to continuously optimize using new orders and vehicle condition data; combining real-time traffic, weather and other data for dynamic scheduling; building a capacity visualization system to monitor vehicle status in real time; applying machine learning algorithms to predict future order demand and traffic congestion; and designing an intelligent scheduling rule engine to support rapid policy changes.

[0138] In this embodiment, the goal of the model is to achieve intelligent and optimal vehicle scheduling under multiple constraints, maximize transportation efficiency and reduce operating costs. At the same time, it is necessary to solve complex optimization problems with high dynamics, multiple variables and multiple objectives, and it requires the support of new technologies such as big data and artificial intelligence.

[0139] In some possible implementations of the present invention, the steps of constructing a logistics demand forecasting model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training respectively include:

[0140] Extracting historical status data in the historical transportation process (such as cargo or vehicle abnormality data, vehicle failure data, disaster data, etc.) from the historical logistics data;

[0141] Performing data cleaning, format conversion, missing value processing, and relationship analysis on the historical logistics order data, the historical goods data, the historical logistics facility data, the historical transportation tool data, the historical transportation route data, and the historical status data to construct a fourth data set;

[0142] Annotating known abnormal cases (such as delays, damages, thefts, etc.) in the fourth data set as supervised learning training samples;

[0143] Extracting abnormality-related feature data (such as time window, geographical location, environmental parameters, logistics status parameters, etc.) related to the abnormal case from the fourth data set;

[0144] Build basic anomaly detection and warning models (supervised learning model: based on labeled data, use logistic regression, decision tree, support vector machine, etc. to train binary anomaly classifiers; unsupervised learning model: use clustering algorithms (K-Means, etc.) to model normal patterns and detect deviated anomalies; semi-supervised model: combine a small amount of labeled data with a large amount of unlabeled data to jointly train anomaly detectors; generative adversarial model: use variational autoencoders, GAN, etc. to find anomalies from normal data distribution);

[0145] The basic anomaly detection and early warning model is trained and tested using the data of the abnormal cases and the abnormality-related feature data to obtain the anomaly detection and early warning model.

[0146] In this embodiment, it also includes: introducing transfer learning technology to improve detection capabilities using data from other fields; deploying the model to the monitoring platform of the production environment to support real-time anomaly detection; configuring an early warning rule engine to trigger different response strategies for different levels of anomalies; establishing a model scoring system, collecting manual feedback, and calculating evaluation indicators; regularly retraining the model using newly collected data to continuously optimize and improve detection capabilities; integrating the anomaly detection model with root cause analysis, intelligent decision-making and other functions.

[0147] In this embodiment, through the above steps, using machine learning and deep learning algorithms, abnormal patterns can be mined from massive historical logistics data, and an efficient anomaly detection and early warning model can be built to timely discover various abnormal situations in the logistics system, avoid the expansion of losses, and improve the level of operational intelligence.

[0148] In some possible implementations of the present invention, the step of generating a logistics plan for a logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, in combination with the logistics demand forecasting model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, includes:

[0149] Extracting key order information from the logistics order data to be processed;

[0150] Using natural language processing technology to structure the key information of the order;

[0151] Combining the key order information with the current logistics dynamic data, and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed;

[0152] Generate alternative route plans using the logistics route planning model according to the starting and ending locations, time windows, etc.;

[0153] Conduct multi-dimensional evaluation of each alternative route plan in terms of timeliness, cost, and abnormal risk;

[0154] In this step, the transportation time of each alternative route is evaluated to see whether it can meet the timeliness requirements of the order (such as next-day delivery, 48-hour delivery, etc.). The evaluation factors include route distance, road conditions, possible congestion, etc. At the same time, the estimated transportation cost of each route plan (including fuel consumption, tolls, labor costs, etc.) is also calculated, and the evaluation results serve as an important basis for cost control. In addition, the risk of abnormal situations occurring in each route plan is evaluated using anomaly detection and early warning models (such as risks of cargo damage, delays, theft, vehicle failure, etc.); according to actual needs, other evaluation dimensions can be added, such as service quality evaluation: evaluating the degree of satisfaction with customer service commitments; environmental impact assessment: evaluating the impact of route plans on the environment; safety assessment: evaluating driving safety risks; comfort assessment: evaluating the impact on the driver's comfort. Through multi-dimensional comprehensive evaluation, the differences between the alternative route plans can be comprehensively compared, and scores can be given according to the set weights to provide data support for the subsequent selection of the optimal plan.

[0155] Input the alternative route plan and the transportation demand into the transportation resource scheduling model to generate a vehicle scheduling plan (including vehicle model, quantity, scheduling time, etc.) that meets the transportation demand;

[0156] In this step, the alternative route plan obtained by the route planning model and the transportation demand predicted by the demand forecasting model are passed as input data to the trained transportation resource scheduling model. Based on these input data, the transportation resource scheduling model will output a vehicle scheduling plan so that the vehicle capacity allocated according to the plan can meet the given transportation demand. Specifically, the vehicle scheduling plan includes the following key contents: vehicle model (specify the required vehicle model for this transportation task according to factors such as cargo volume and weight, such as trucks, semi-trailers, etc.), quantity (determine how many vehicles are needed to meet the predicted transportation demand), scheduling time (specify a specific scheduling time for each vehicle to ensure that the transportation demand is met in time and the delivery task is completed), and others (may also include specific vehicle numbers, driver arrangements and other details). The generation of the vehicle scheduling plan requires the transportation scheduling model to give the best deployment plan through the optimization algorithm under the premise of meeting multiple constraints such as time window, load capacity, fuel consumption, etc., to maximize the use of existing transportation resources, improve efficiency and reduce operating costs.

[0157] Based on the current logistics dynamic data, the abnormal risk of alternative plans is analyzed using the abnormal detection and early warning model to obtain abnormal analysis results, and the logistics plan for the logistics order to be processed is determined according to the abnormal analysis results (high-risk plans are avoided and low-risk alternative plans are given priority).

[0158] In this embodiment, the route, capacity, and anomaly analysis results are integrated to form a complete logistics execution plan. A simulation model or evaluation function can also be used to comprehensively evaluate and score the logistics plan. The plan with the highest comprehensive score is selected as the final recommended plan. If there is no satisfactory plan, the parameter constraints can be adjusted and regenerated. For complex scenarios, manual decision makers can intervene to perform further optimization. Explainable machine learning is used to assist manual decision-making.

[0159] In this embodiment, through the above process, the real-time logistics data collected by the Internet of Things can be fully utilized, combined with a variety of intelligent models, to generate an intelligent and efficient logistics execution plan for each order. At the same time, it has the ability to control abnormal risks to ensure the reliability of the plan.

[0160] In some possible implementations of the present invention, the step of combining the key order information with the current logistics dynamic data and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed includes:

[0161] Collect and extract current inventory data, current transportation capacity data, current weather data, and current traffic data from the current logistics dynamic data;

[0162] The current inventory data, the current transportation capacity data, the current weather data, the current traffic data and the key information of the order (key features affecting transportation demand, such as cargo type, weight, volume, start and end locations, time window, service level, timeliness requirements, cost constraints, etc.) are integrated to construct a complete input feature vector;

[0163] The input feature vector is input into the logistics demand forecasting model to obtain the corresponding transportation demand forecast value (such as total freight volume, one-way freight volume, staged freight volume, etc.), which is the transportation demand corresponding to the logistics order to be processed.

[0164] This embodiment integrates order data and logistics dynamic data, uses machine learning models to mine patterns in the data, and more accurately predicts the transportation demand corresponding to specific orders, providing a basis for resource allocation.

[0165] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0166] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0167] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0168] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0170] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

[0171] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0172] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0173] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions without departing from the spirit and scope of the present invention, and can make various changes and modifications, including the combination of the above-mentioned different functions and implementation steps, including software and hardware implementation methods, all of which are within the scope of protection of the present invention.

Claims

1. A logistics solution intelligent design system, characterized in that: include: A cloud server, an IoT edge server communicatively connected to the cloud server, a first intelligent sensing terminal group disposed in a logistics facility, a second intelligent sensing terminal group disposed in a transportation vehicle, and a third intelligent sensing terminal group disposed in a cargo, wherein the IoT edge server is communicatively connected to the first intelligent sensing terminal group, the second intelligent sensing terminal group, and the third intelligent sensing terminal group; The cloud server is configured as follows: The first intelligent sensing terminal group acquires first logistics facility data in real time, the second intelligent sensing terminal group acquires first transportation tool data in real time, and the third intelligent sensing terminal group acquires first cargo data in real time; Using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data; Based on historical logistics data and big data training, we build logistics demand forecasting models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models. Generate a logistics plan for the logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model; The method for setting the first intelligent sensing terminal group, the second intelligent sensing terminal group, and the third intelligent sensing terminal group includes: Acquire first attribute data, first status data and first environment data of the logistics facility; Acquiring second attribute data, second state data, and second environment data of the transportation tool; Acquiring third attribute data, third state data, and third environment data of the goods; installing the first intelligent sensing terminal group in a first key area of ​​the logistics facility according to the first attribute data, the first status data, and the first environmental data; installing the second intelligent sensing terminal group in a second key area of ​​the transportation vehicle according to the second attribute data, the second state data and the second environment data; Installing the third intelligent sensing terminal group in a third key area of ​​the cargo according to the third attribute data, the third state data and the third environment data; The step of using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data and the first cargo data to obtain current logistics dynamic data includes: The cloud server uses data mining algorithms to analyze historical logistics facility data, historical transportation tool data, and historical cargo data, and constructs descriptive analysis models, diagnostic analysis models, and predictive analysis models; The cloud server uses a big data processing framework to clean, deduplicate, and convert the first logistics facility data, the first transportation tool data, and the first cargo data, and converts the three types of heterogeneous data into input data that conforms to the input format of the descriptive analysis model, the diagnostic analysis model, and the predictive analysis model through data fusion technology; Inputting the input data into the descriptive analysis model, the diagnostic analysis model and the predictive analysis model, integrating the analysis results to obtain current logistics dynamic data; The steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training include: Extracting historical logistics order data, historical cargo data, historical logistics facility data, and historical transportation tool data from the historical logistics data as a first basic data set; Cleaning, format conversion, missing value processing, and outlier processing of the data in the first basic data set; Fusing and associating the data in the first basic data set to construct a unified first training data set; Extracting features of the dimensions of time, location, cargo attributes, and environmental factors from the first training data set as first features; Adopting the time series forecasting model, an order forecasting model is trained based on the historical order quantity sequence to predict future order demand; Using a machine learning model, combined with the first feature, to train a transportation demand prediction model for predicting transportation demand; Constructing a multi-model integration framework, fusing the order prediction model and the transportation demand prediction model to obtain a first fusion model; Splitting the first training data set into a first training set, a first validation set, and a first test set; Training the first fusion model on the first training set, adjusting hyperparameters on the first validation set, and optimizing the structure of the first fusion model; Evaluating the optimized first fusion model on the first test set, and adjusting the first fusion model according to the evaluation result to obtain the logistics demand prediction model; The step of generating a logistics plan for the logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, in combination with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, includes: Extracting key order information from the logistics order data to be processed; Using natural language processing technology to structure the key information of the order; Combining the key order information with the current logistics dynamic data, and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed; Generate alternative route plans using the logistics route planning model according to the starting and ending locations and time windows; Conduct multi-dimensional evaluation of each alternative route plan in terms of timeliness, cost, and abnormal risk; Inputting the alternative route plan and the transportation demand into the transportation resource scheduling model to generate a vehicle scheduling plan that meets the transportation demand; Based on the current logistics dynamic data, the abnormal risk of the alternative plan is analyzed using the abnormal detection and early warning model to obtain an abnormal analysis result, and the logistics plan for the logistics order to be processed is determined according to the abnormal analysis result.

2. A method for intelligent design of logistics solutions, characterized in that: include: The first intelligent sensing terminal group set at the logistics facility obtains the first logistics facility data in real time, the second intelligent sensing terminal group set at the transportation tool obtains the first transportation tool data in real time, and the third intelligent sensing terminal group set at the cargo obtains the first cargo data in real time; The cloud server uses big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data, and the first cargo data to obtain current logistics dynamic data; Based on historical logistics data and big data training, we build logistics demand forecasting models, logistics route planning models, transportation resource scheduling models, and anomaly detection and early warning models. Generate a logistics plan for the logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, combined with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model; The step of obtaining first logistics facility data in real time by using a first intelligent sensing terminal group arranged at a logistics facility, obtaining first transportation tool data in real time by using a second intelligent sensing terminal group arranged at a transportation tool, and obtaining first cargo data in real time by using a third intelligent sensing terminal group arranged at the cargo includes: Acquire first attribute data, first status data and first environment data of the logistics facility; Acquiring second attribute data, second state data, and second environment data of the transportation tool; Acquiring third attribute data, third state data, and third environment data of the goods; installing the first intelligent sensing terminal group in a first key area of ​​the logistics facility according to the first attribute data, the first status data, and the first environmental data; installing the second intelligent sensing terminal group in a second key area of ​​the transportation vehicle according to the second attribute data, the second state data and the second environment data; Installing the third intelligent sensing terminal group in a third key area of ​​the cargo according to the third attribute data, the third state data and the third environment data; The step of the cloud server using big data analysis technology to perform correlation analysis on the first logistics facility data, the first transportation tool data and the first cargo data to obtain current logistics dynamic data includes: The cloud server uses data mining algorithms to analyze historical logistics facility data, historical transportation tool data, and historical cargo data, and constructs descriptive analysis models, diagnostic analysis models, and predictive analysis models; The cloud server uses a big data processing framework to clean, deduplicate, and convert the first logistics facility data, the first transportation tool data, and the first cargo data, and converts the three types of heterogeneous data into input data that conforms to the input format of the descriptive analysis model, the diagnostic analysis model, and the predictive analysis model through data fusion technology; Inputting the input data into the descriptive analysis model, the diagnostic analysis model and the predictive analysis model, integrating the analysis results to obtain current logistics dynamic data; The steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training include: Extracting historical logistics order data, historical cargo data, historical logistics facility data, and historical transportation tool data from the historical logistics data as a first basic data set; Cleaning, format conversion, missing value processing, and outlier processing of the data in the first basic data set; Fusing and associating the data in the first basic data set to construct a unified first training data set; Extracting features of the dimensions of time, location, cargo attributes, and environmental factors from the first training data set as first features; Adopting the time series forecasting model, an order forecasting model is trained based on the historical order quantity sequence to predict future order demand; Using a machine learning model, combined with the first feature, to train a transportation demand prediction model for predicting transportation demand; Constructing a multi-model integration framework, fusing the order prediction model and the transportation demand prediction model to obtain a first fusion model; Splitting the first training data set into a first training set, a first validation set, and a first test set; Training the first fusion model on the first training set, adjusting hyperparameters on the first validation set, and optimizing the structure of the first fusion model; Evaluating the optimized first fusion model on the first test set, and adjusting the first fusion model according to the evaluation result to obtain the logistics demand prediction model; The step of generating a logistics plan for the logistics order to be processed based on the current logistics dynamic data and the logistics order data to be processed, in combination with the logistics demand prediction model, the logistics route planning model, the transportation resource scheduling model, and the anomaly detection and early warning model, includes: Extracting key order information from the logistics order data to be processed; Using natural language processing technology to structure the key information of the order; Combining the key order information with the current logistics dynamic data, and using the logistics demand forecasting model to forecast the transportation demand corresponding to the logistics order to be processed; Generate alternative route plans using the logistics route planning model according to the starting and ending locations and time windows; Conduct multi-dimensional evaluation of each alternative route plan in terms of timeliness, cost, and abnormal risk; Inputting the alternative route plan and the transportation demand into the transportation resource scheduling model to generate a vehicle scheduling plan that meets the transportation demand; Based on the current logistics dynamic data, the abnormal risk of the alternative plan is analyzed using the abnormal detection and early warning model to obtain an abnormal analysis result, and the logistics plan for the logistics order to be processed is determined according to the abnormal analysis result.

3. The intelligent design method for logistics solutions according to claim 2 is characterized in that: The steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training include: extracting historical transportation route data from the historical logistics data; After establishing a first association relationship between the historical logistics order data, the historical goods data, the historical logistics facility data, the historical transportation tool data and the historical transportation route data, the historical logistics order data are used as a second training data set; constructing a first modeling algorithm for performing path planning; Splitting the second training data set into a second training set, a second validation set, and a second test set; Training the first modeling algorithm on the second training set to obtain a basic path planning model; Adjusting hyperparameters on the second validation set to optimize the structure of the basic path planning model; The optimized basic path planning model is evaluated on the second test set, and the basic path planning model is adjusted according to the evaluation result to obtain the logistics route planning model.

4. The intelligent design method for logistics solutions according to claim 3 is characterized in that: The steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training include: Determine the first optimization goal of logistics and transportation; Construct the first constraint; Constructing a first optimization model according to the first optimization objective and the first constraint condition; Generate a third training data set and a third test data set based on the historical logistics order data, the historical transportation tool data, the historical logistics facility data, and the historical route data; Using the third training data set to train the first optimization model to obtain a basic transportation resource scheduling model; The basic transportation resource scheduling model is evaluated on the third test data set, and the basic transportation resource scheduling model is adjusted according to the evaluation result to obtain the transportation resource scheduling model.

5. The intelligent design method for logistics solutions according to claim 4 is characterized in that: The steps of constructing a logistics demand prediction model, a logistics route planning model, a transportation resource scheduling model, and an anomaly detection and early warning model based on historical logistics data and big data training include: Extracting historical status data in the historical transportation process from the historical logistics data; Performing data cleaning, format conversion, missing value processing, and associating the historical logistics order data, the historical cargo data, the historical logistics facility data, the historical transportation tool data, the historical transportation route data, and the historical status data to construct a fourth data set; Marking known abnormal cases in the fourth data set as supervised learning training samples; extracting abnormality-related feature data related to the abnormal case from the fourth data set; Build basic anomaly detection and early warning models; The basic anomaly detection and early warning model is trained and tested using the data of the abnormal cases and the abnormality-related feature data to obtain the anomaly detection and early warning model.

6. The intelligent design method for logistics solutions according to claim 5 is characterized in that: The step of combining the key order information with the current logistics dynamic data and using the logistics demand prediction model to predict the transportation demand corresponding to the logistics order to be processed includes: Collect and extract current inventory data, current transportation capacity data, current weather data, and current traffic data from the current logistics dynamic data; The current inventory data, the current transportation capacity data, the current weather data, the current traffic data and the key order information are integrated to construct a complete input feature vector; The input feature vector is input into the logistics demand prediction model to obtain the corresponding transportation demand prediction value, which is the transportation demand corresponding to the logistics order to be processed.

Images