Logistics transportation intelligent prediction scheduling method and system based on digital twinning

Abstract

The invention discloses a logistics transportation intelligent prediction scheduling method and system based on digital twinning. The method comprises the steps that transportation carrier data and cargo data are acquired in real time through Internet of Things sensing equipment; constructing a logistics transportation digital twinborn model, wherein the logistics transportation digital twinborn model comprises a physical entity layer, a data fusion layer, a virtual model layer and a decision service layer; simulating a logistics transportation process, and deducing and predicting resource demands in a future time window; the method comprises the following steps: dynamically adjusting marshalling configuration and loading and unloading operation time sequence of transportation carriers by constructing a digital twin-driven elastic resource scheduling mechanism; and training and optimizing the logistics transportation digital twin model based on feedback data and scheduling execution deviation data in the logistics transportation process to form a closed-loop optimization system. The method has the advantages that logistics transportation data are simulated through the digital twin technology, logistics carrier scheduling is predicted, resource demand prediction and elastic scheduling are achieved, and the transportation efficiency is optimized. And the logistics process is ensured to efficiently and flexibly cope with uncertain factors.

Description

Technical Field

[0001] The present invention relates to dynamic scheduling technology, and in particular to a logistics transportation intelligent prediction and scheduling method and system based on digital twins. Background Art

[0002] With the continuous advancement of globalization, cross-border trade and international supply chains are becoming increasingly complex. Logistics efficiency and accuracy have become core competitive advantages for companies in the global market. Furthermore, the booming e-commerce industry has driven consumer demand for fast delivery and precise service, further intensifying the demand for efficient and accurate logistics. To meet this challenge, companies must optimize their logistics networks, improve the efficiency of transport resource allocation, and enhance visibility and control over the transportation process.

[0003] Most logistics and transportation scheduling methods currently available on the market often rely on static rules and simple algorithms, lacking real-time dynamic monitoring and intelligent forecasting capabilities. These methods often rely on manual intervention or fixed scheduling schemes, and are unable to effectively respond to sudden transportation issues such as traffic congestion, weather changes, or equipment failures, resulting in inefficient transportation and wasted resources. Furthermore, traditional scheduling methods have weak predictive capabilities and struggle to accurately predict future resource demands, making it difficult to achieve optimal resource allocation. Overallocation or resource shortages are common. Due to the lack of real-time feedback and optimization mechanisms, these methods are unable to automatically adjust and optimize during the actual transportation process, resulting in execution deviations that cannot be corrected in a timely manner and poor scheduling results. However, scheduling methods based on digital twins, by integrating real-time data, simulation predictions, and feedback optimization, can make more accurate decisions in dynamically changing environments, significantly improving logistics efficiency and resource utilization. Summary of the Invention

[0004] To improve existing logistics and transportation scheduling methods, this paper proposes a digital twin-based intelligent forecasting and scheduling method and system for logistics and transportation. This method uses digital twin technology to acquire and simulate logistics and transportation data in real time and predict the scheduling of logistics vehicles. This method enables resource demand forecasting and flexible scheduling, optimizing transportation efficiency. By combining the Internet of Things, simulation, and feedback optimization, it can dynamically adjust transportation resource allocation, ensuring efficient logistics processes and flexible response to uncertainties.

[0005] In order to achieve the above objects, the technical solution adopted by the present invention is:

[0006] The intelligent prediction and scheduling method for logistics transportation based on digital twins includes:

[0007] Use IoT sensing devices to obtain real-time positioning data, working condition data, and environmental sensor data of transport vehicles, and combine it with order data to obtain cargo flow status data;

[0008] Build a digital twin model of logistics and transportation based on the acquired transport vehicle data and cargo flow data, including the physical entity layer, data fusion layer, virtual model layer and decision service layer;

[0009] Based on the digital twin model of logistics and transportation, the logistics and transportation process is simulated to deduce and predict resource demand within the future time window;

[0010] Based on the predicted resource demand within the future time window, a flexible resource scheduling mechanism driven by digital twins is built to dynamically adjust the grouping configuration of transport vehicles and the timing of loading and unloading operations;

[0011] Based on the feedback data and scheduling execution deviation data during the logistics and transportation process, the logistics and transportation digital twin model is trained and optimized to form a closed-loop optimization system.

[0012] Preferably, the logistics and transportation digital twin model is constructed based on the acquired transport vehicle data and cargo flow data, including a physical entity layer, a data fusion layer, a virtual model layer and a decision service layer, specifically including:

[0013] Connect various sensors of the transport vehicle with the onboard ECU control unit, and synchronize the operating status of the transport vehicle's physical equipment to the digital twin to build the physical entity layer;

[0014] Map the order data to the timeline of the simulation scenario based on the data fusion layer and align time and space;

[0015] Based on static data, road network topology, vehicle distribution, and storage nodes are loaded to build a three-dimensional geographic information sandbox. Based on external dynamic data sources, simulation time windows are set, weather forecasts, and traffic control plans are synchronized to build a virtual model layer.

[0016] A mixed integer programming solver is used to make online optimization decisions for numerous process problems. The trigger weights of abnormal events are dynamically adjusted using an importance sampling algorithm, randomly mapped to the timeline of the simulation scenario, and a decision-making service layer is constructed.

[0017] Based on the above four modules, a digital twin model of logistics and transportation is constructed.

[0018] Preferably, the simulation of the logistics and transportation process based on the logistics and transportation digital twin model and the prediction of resource demand in the future time window specifically include:

[0019] Based on the constructed digital twin model of logistics and transportation, the process was simulated using colored Petri nets, and the resource occupancy time series data of each path was recorded through Monte Carlo parallel simulation.

[0020] Based on the interaction mechanism between vehicles, cargo, and the environment, three types of intelligent agent models are constructed, and real-time state synchronization between intelligent agents is achieved;

[0021] Based on the time series data obtained through simulation, a resource demand forecasting model is constructed to predict the demand for key resources, including transportation capacity, energy, manpower, and facilities.

[0022] Preferably, the dynamically adjusting the grouping configuration and loading and unloading operation sequence of transport vehicles by building a flexible resource scheduling mechanism driven by digital twins based on the predicted resource demand in the future time window specifically includes:

[0023] Based on the prediction results of the resource demand prediction model, set constraints on the elastic resource scheduling mechanism, including resource constraints, time constraints, and environmental constraints;

[0024] The specific configuration of transport vehicles includes:

[0025] The vehicle-task allocation relationship and group combination coding are defined through binary matrices, and a load balance and spatiotemporal coupling constraint model is established;

[0026] Based on the double-chain chromosome encoding structure, a multi-objective weighted fitness function is designed, and adaptive crossover probability and tabu search are introduced to assist grouping;

[0027] Verify the optimality of the solution through the digital twin interface and establish elastic adjustment triggering rules for vehicle failures and traffic congestion;

[0028] The loading and unloading operation sequence specifically includes:

[0029] Based on the key resource data predicted by the resource demand forecasting model, a four-dimensional space-time cube is constructed to represent the loading and unloading resource status;

[0030] A time window penalty model is established through mixed integer programming, and a column generation algorithm is used to minimize the total loading and unloading cost and accelerate the generation of feasible solutions.

[0031] Based on cargo data, the priority of cargo is dynamically adjusted.

[0032] Preferably, the training and optimization of the logistics and transportation digital twin model based on the feedback data and scheduling execution deviation data in the logistics and transportation process to form a closed-loop optimization system specifically includes:

[0033] Based on real-time collected data and resource demand forecasting model forecast data, extract multi-dimensional deviation data and extract deviation data features;

[0034] Retrain the digital twin model through the collected feedback data and deviation data features to continuously optimize the model parameters;

[0035] Based on the trained digital twin model, key indicators in the logistics transportation process are predicted in real time. The model automatically feedbacks the deviations generated during the scheduling execution process and further adjusts the model and scheduling strategy.

[0036] Furthermore, a digital twin-based intelligent prediction and scheduling system for logistics and transportation is proposed, including:

[0037] IoT sensing module: This module is mainly used to collect real-time data on the location, working conditions, environment, and cargo status of transport vehicles, and achieve unified access of heterogeneous devices through a multi-protocol adapter;

[0038] Digital Twin Module: This module is primarily used to perform spatiotemporal alignment and feature-level fusion of multi-source heterogeneous data, construct a four-dimensional data matrix and extract key eigenvectors. It also builds a high-precision three-dimensional virtual model based on BIM and Petri nets, and integrates a physics engine to simulate vehicle dynamic characteristics.

[0039] Simulation deduction module: The simulation deduction module is mainly used to generate future resource requirements by using Monte Carlo parallel simulation and multi-agent interactive deduction;

[0040] Flexible scheduling module: The flexible scheduling module is mainly used to dynamically optimize the transport group configuration and loading and unloading sequence through mixed integer programming and reinforcement learning;

[0041] Online optimization module: The online optimization module is mainly used to retrain the digital twin model based on feedback data and optimize model parameters;

[0042] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0043] Compared with the prior art, the advantages of the present invention are:

[0044] By acquiring real-time data on transport vehicle positioning, operating conditions, and environmental conditions, combined with order information, the company can comprehensively monitor and understand the flow of goods, ensuring data accuracy and real-time availability. The constructed digital twin model integrates the physical entity layer, data fusion layer, virtual model layer, and decision-making service layer to efficiently simulate and analyze the entire logistics and transportation process. This model accurately predicts resource demand within future time windows and proactively identifies potential transportation bottlenecks or resource shortages. Furthermore, a flexible resource scheduling mechanism driven by the digital twin dynamically adjusts the vehicle grouping and operation sequence based on the predicted results, optimizing transportation efficiency and reducing costs. Through continuous feedback and optimization, a closed-loop self-learning and adjustment mechanism is formed, enabling continuous optimization and refinement of the transportation process, thereby improving the responsiveness, adaptability, and resource utilization of the overall logistics and transportation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of the method proposed in the present invention;

[0046] Figure 2 Schematic diagram of the logistics and transportation digital twin model proposed in this invention;

[0047] Figure 3 This is a schematic diagram of resource demand prediction proposed by the present invention;

[0048] Figure 4 This is a schematic diagram of the elastic resource scheduling proposed by the present invention;

[0049] Figure 5 This is a schematic diagram of the feedback optimization proposed by the present invention;

[0050] Figure 6 This is a diagram of the architecture of the electronic equipment in this solution;

[0051] Figure 7 This is a schematic diagram of the computer-readable storage medium structure in this solution. DETAILED DESCRIPTION

[0052] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are merely examples, and those skilled in the art may conceive of other obvious variations.

[0053] The intelligent prediction and scheduling system for logistics and transportation based on digital twins includes:

[0054] IoT sensing module: This module is mainly used to collect real-time data on the location, working conditions, environment, and cargo status of transport vehicles, and achieve unified access of heterogeneous devices through a multi-protocol adapter;

[0055] Digital Twin Module: This module is primarily used to perform spatiotemporal alignment and feature-level fusion of multi-source heterogeneous data, construct a four-dimensional data matrix and extract key eigenvectors. It also builds a high-precision three-dimensional virtual model based on BIM and Petri nets, and integrates a physics engine to simulate vehicle dynamic characteristics.

[0056] Simulation deduction module: The simulation deduction module is mainly used to generate future resource requirements by using Monte Carlo parallel simulation and multi-agent interactive deduction;

[0057] Flexible scheduling module: The flexible scheduling module is mainly used to dynamically optimize the transport group configuration and loading and unloading sequence through mixed integer programming and reinforcement learning;

[0058] Online optimization module: The online optimization module is mainly used to retrain the digital twin model based on feedback data and optimize model parameters;

[0059] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0060] See Figure 1 As shown in the figure, the intelligent prediction and scheduling method for logistics transportation based on digital twins includes:

[0061] Step 1: Use IoT sensing devices to obtain real-time positioning data, operating condition data, and environmental sensor data of transport vehicles, and combine this with order data to obtain cargo flow status data;

[0062] Step 2: Build a digital twin model of logistics and transportation based on the acquired transport vehicle data and cargo flow data, including the physical entity layer, data fusion layer, virtual model layer, and decision service layer;

[0063] Step 3: Based on the logistics and transportation digital twin model, simulate the logistics and transportation process and deduce and predict resource demand within the future time window;

[0064] Step 4: Based on the predicted resource demand in the future time window, a flexible resource scheduling mechanism driven by digital twins is built to dynamically adjust the transport vehicle grouping configuration and loading and unloading operation sequence:

[0065] Step 5: Based on the feedback data and scheduling execution deviation data during the logistics and transportation process, train and optimize the logistics and transportation digital twin model to form a closed-loop optimization system.

[0066] See Figure 2 As shown in the figure, a logistics and transportation digital twin model is constructed based on the acquired transport vehicle data and cargo flow data. It includes the physical entity layer, data fusion layer, virtual model layer and decision service layer. Specifically, it includes:

[0067] Connect various sensors of the transport vehicle with the onboard ECU control unit, and synchronize the operating status of the transport vehicle's physical equipment to the digital twin to build the physical entity layer;

[0068] Map the order data to the timeline of the simulation scenario based on the data fusion layer and align time and space;

[0069] Based on static data, road network topology, vehicle distribution, and storage nodes are loaded to build a three-dimensional geographic information sandbox. Based on external dynamic data sources, simulation time windows are set, weather forecasts, and traffic control plans are synchronized to build a virtual model layer.

[0070] A mixed integer programming solver is used to make online optimization decisions for numerous process problems. The trigger weights of abnormal events are dynamically adjusted using an importance sampling algorithm, randomly mapped to the timeline of the simulation scenario, and a decision-making service layer is constructed.

[0071] Based on the above four modules, a digital twin model of logistics and transportation is constructed.

[0072] Specifically, in the physical layer, various sensors (such as GPS, temperature and humidity sensors, speed sensors, etc.) and the vehicle ECU control unit are connected through communication protocols (such as CAN bus, Ethernet, etc.) to collect vehicle operating status information in real time. The formula for the MQTT / OPC UA protocol transmission data packet structure is:

[0073] P={t,ID,(s1,...,s n ),CRC}

[0074] Among them, t is the timestamp, s n is the normalized state parameter, CRC is the cyclic redundancy check;

[0075] Build a virtual model in the digital twin and synchronize the vehicle's sensor data in real time to ensure the consistency between the digital twin and the physical world.

[0076] In the data fusion layer, the UTM coordinate system is used to convert GPS data. The formula is:

[0077] (x,y)=f UTM (λ,φ,zone)

[0078] Among them, λ is longitude and φ is latitude;

[0079] When the time axis is dynamically mapped, define the simulation time axis T sim =αT real +β, aligning cargo events through dynamic time warping, the formula is:

[0080]

[0081] In the virtual model layer, static scene modeling builds a topology map based on road network data, and the node weight setting formula is:

[0082]

[0083] Among them, d k is the demand point, t teavel For travel time;

[0084] Dynamic data injection into the model through meteorological impact data.

[0085] See Figure 3 As shown in the figure, based on the digital twin model of logistics and transportation, the logistics and transportation process is simulated, and the resource demand in the future time window is predicted, including:

[0086] Based on the constructed digital twin model of logistics and transportation, the process was simulated using colored Petri nets, and the resource occupancy time series data of each path was recorded through Monte Carlo parallel simulation.

[0087] Based on the interaction mechanism between vehicles, cargo, and the environment, three types of intelligent agent models are constructed, and real-time state synchronization between intelligent agents is achieved;

[0088] Based on the time series data obtained through simulation, a resource demand forecasting model is constructed to predict the demand for key resources, including transportation capacity, energy, manpower, and facilities.

[0089] Specifically, the structure is defined based on the colored Petri net, and the place set is defined as follows: P = {p1,...,p m} represents the resource status (such as charging pile occupancy, warehouse capacity), and defines the transition set: T = {t1,...,t n} represents logistics events (loading and unloading, route selection);

[0090] Generate N parallel simulation paths based on Monte Carlo parallel simulation. The formula is:

[0091]

[0092] Where M(t) is the marker vector, ε~N(0,1);

[0093] Based on the interaction mechanism between vehicles, cargo and environment, three types of intelligent agent models are constructed:

[0094] The vehicle agent is responsible for performing transportation tasks in the logistics system. Its status includes location, speed, and status (empty, in transit, unloading, etc.);

[0095] The cargo agent is responsible for tracking the order and cargo status during transportation and participating in the decision-making process (such as priority, time requirements, etc.);

[0096] Environmental agents represent external environmental factors, such as traffic conditions and weather changes, which affect the status of vehicles and cargo;

[0097] When predicting multi-dimensional resource demand, we predict transportation demand based on LSTM+Attention, energy demand based on the Prophet model, manpower demand through the spatiotemporal graph convolutional network, and build a resource demand prediction model through multi-objective optimization.

[0098] See Figure 4 As shown in the figure, based on the predicted resource demand in the future time window, a flexible resource scheduling mechanism driven by digital twins is built to dynamically adjust the grouping configuration of transport vehicles and the timing of loading and unloading operations. Specifically, the following are involved:

[0099] Based on the prediction results of the resource demand prediction model, set constraints on the elastic resource scheduling mechanism, including resource constraints, time constraints, and environmental constraints;

[0100] The specific configuration of transport vehicles includes:

[0101] The vehicle-task allocation relationship and group combination coding are defined through binary matrices, and a load balance and spatiotemporal coupling constraint model is established;

[0102] Based on the double-chain chromosome encoding structure, a multi-objective weighted fitness function is designed, and adaptive crossover probability and tabu search are introduced to assist grouping;

[0103] Verify the optimality of the solution through the digital twin interface and establish elastic adjustment triggering rules for vehicle failures and traffic congestion;

[0104] The loading and unloading operation sequence specifically includes:

[0105] Based on the key resource data predicted by the resource demand forecasting model, a four-dimensional space-time cube is constructed to represent the loading and unloading resource status;

[0106] A time window penalty model is established through mixed integer programming, and a column generation algorithm is used to minimize the total loading and unloading cost and accelerate the generation of feasible solutions.

[0107] Based on cargo data, the priority of cargo is dynamically adjusted.

[0108] Specifically, the resource constraint formula is:

[0109]

[0110] Among them, x ijk Assign variables to tasks, is the resource consumption coefficient, δ r is the elastic redundancy rate;

[0111] The time constraint formula is:

[0112] max(t arrive ,TW start )≤t depart ≤TW end +η·|(violation)

[0113] Where η is the time window relaxation penalty coefficient;

[0114] The environmental constraint formula is:

[0115]

[0116] Among them, I wis the weather intensity index;

[0117] During the optimization process of transport vehicle grouping, the allocation relationship between vehicles and tasks is defined through a binary matrix.

[0118]

[0119] In order to ensure the load balance and space-time coupling constraints of the vehicle, a multi-objective optimization model can be designed to minimize the total transportation cost, time and load imbalance. The load balance constraint can be expressed as:

[0120]

[0121] Among them, W vehicle is the maximum load of the vehicle, W task The load requirements for each mission;

[0122] In the optimization of loading and unloading operation timing, a mixed integer programming is used to establish a time window penalty model to optimize the loading and unloading schedule. The objective function is to minimize the total loading and unloading cost. The constraints include that each task must be completed within the time window:

[0123]

[0124] t start (j)≥t ready (j),t end (j)≤t deadline (j)

[0125] Column generation algorithms can help speed up the generation of feasible solutions, especially for large-scale problems, and can effectively reduce the amount of computation;

[0126] Dynamically adjust the loading and unloading order based on the priority of the goods. Priority can be set based on factors such as delivery time, value, or customer requirements. Priority adjustment can be performed using the following rules:

[0127] P task (j)=α·Urgency(j)+β·Value(j)

[0128] Among them, P task (j) is the priority of task j, α and β are weight coefficients, Urgency(j) and Value(j) are the urgency and value of the task, respectively.

[0129] See Figure 5 As shown in the figure, based on the feedback data and scheduling execution deviation data during the logistics and transportation process, the logistics and transportation digital twin model is trained and optimized to form a closed-loop optimization system, which specifically includes:

[0130] Based on real-time collected data and resource demand forecasting model forecast data, extract multi-dimensional deviation data and extract deviation data features;

[0131] Retrain the digital twin model through the collected feedback data and deviation data features to continuously optimize the model parameters;

[0132] Based on the trained digital twin model, key indicators in the logistics transportation process are predicted in real time. The model automatically feedbacks the deviations generated during the scheduling execution process and further adjusts the model and scheduling strategy.

[0133] Specifically, since the deviation data may come from multiple dimensions (such as transportation time, transportation capacity, energy consumption, etc.), we need to extract deviation features from different angles. Let the different dimensions of the deviation data be ΔD1(t), ΔD2(t),…, ΔD k (t), where k represents the number of dimensions of the deviation data. The features of each dimension are extracted using statistical methods (such as mean, standard deviation, maximum value, etc.):

[0134] Mean:

[0135] Standard Deviation:

[0136] Maximum and minimum values:

[0137] Feedback data includes various deviations that actually occur during system operation and the effects after scheduling execution. Deviation data features and feedback data are used as input to retrain the digital twin model.

[0138] Through the back-propagation algorithm or other optimization methods (such as gradient descent), the model parameters are adjusted so that the model output is more consistent with the real-time feedback data and prediction target. The optimization goal is to minimize the loss function:

[0139]

[0140] Among them, θ is the model parameter and T is the training time;

[0141] Through the above real-time prediction and feedback mechanism, the model and scheduling strategy are continuously adjusted and optimized. This process is a closed-loop feedback process. In each cycle, the model's prediction results, execution deviation, and resource demand forecast will in turn influence the model retraining, further improving prediction accuracy and scheduling efficiency.

[0142] Furthermore, the method according to the embodiment of the present application can also be used with the aid of Figure 6 The electronic device architecture shown in FIG. Figure 6As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store the intelligent prediction and scheduling method and system for logistics transportation based on digital twins provided in this application. The electronic device 500 may also include a terminal interface 508. Of course, Figure 6 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 6 One or more components of an electronic device are shown.

[0143] Figure 7 This is a schematic diagram of the computer-readable storage medium structure provided by an embodiment of the present application. Figure 7 As shown, a computer-readable storage medium 600 according to an embodiment of the present application is shown. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, the method and system for intelligent prediction and scheduling of logistics transportation based on digital twins according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0144] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0145] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0146] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The intelligent prediction and scheduling method for logistics transportation based on digital twins is characterized by: include: Use IoT sensing devices to obtain real-time positioning data, working condition data, and environmental sensor data of transport vehicles, and combine it with order data to obtain cargo flow status data; Build a digital twin model of logistics and transportation based on the acquired transport vehicle data and cargo flow data, including the physical entity layer, data fusion layer, virtual model layer and decision service layer; Based on the digital twin model of logistics and transportation, the logistics and transportation process is simulated to deduce and predict resource demand within the future time window; Based on the predicted resource demand within the future time window, a flexible resource scheduling mechanism driven by digital twins is built to dynamically adjust the grouping configuration of transport vehicles and the timing of loading and unloading operations; Based on the feedback data and scheduling execution deviation data during the logistics and transportation process, the logistics and transportation digital twin model is trained and optimized to form a closed-loop optimization system.

2. The method for intelligent prediction and scheduling of logistics transportation based on digital twins according to claim 1 is characterized in that: The logistics and transportation digital twin model is constructed based on the acquired transport vehicle data and cargo flow data, including the physical entity layer, data fusion layer, virtual model layer and decision service layer. Specifically, it includes: Connect various sensors of the transport vehicle with the onboard ECU control unit, and synchronize the operating status of the transport vehicle's physical equipment to the digital twin to build the physical entity layer; Map the order data to the timeline of the simulation scenario based on the data fusion layer and align time and space; Based on static data, road network topology, vehicle distribution, and storage nodes are loaded to build a three-dimensional geographic information sandbox. Based on external dynamic data sources, simulation time windows are set, weather forecasts, and traffic control plans are synchronized to build a virtual model layer. A mixed integer programming solver is used to make online optimization decisions for numerous process problems. The trigger weights of abnormal events are dynamically adjusted using an importance sampling algorithm, randomly mapped to the timeline of the simulation scenario, and a decision-making service layer is constructed. Based on the above four modules, a digital twin model of logistics and transportation is constructed.

3. The method for intelligent prediction and scheduling of logistics transportation based on digital twins according to claim 1 is characterized in that: The logistics and transportation digital twin model is used to simulate the logistics and transportation process and predict resource demand in the future time window, specifically including: Based on the completed digital twin model of logistics and transportation, the process was simulated using colored Petri nets, and the resource occupancy time series data of each path was recorded through Monte Carlo parallel simulation. Based on the interaction mechanism between vehicles, cargo, and the environment, three types of intelligent agent models are constructed, and real-time state synchronization between intelligent agents is achieved; Based on the time series data obtained through simulation, a resource demand forecasting model is constructed to predict the demand for key resources, including transportation capacity, energy, manpower, and facilities.

4. The method for intelligent prediction and scheduling of logistics transportation based on digital twins according to claim 1 is characterized in that: The above-mentioned method of dynamically adjusting the transportation vehicle grouping configuration and loading and unloading operation sequence by building a flexible resource scheduling mechanism driven by digital twins based on the predicted resource demand in the future time window specifically includes: Based on the prediction results of the resource demand prediction model, set constraints on the elastic resource scheduling mechanism, including resource constraints, time constraints, and environmental constraints; The specific configuration of transport vehicles includes: The vehicle-task allocation relationship and group combination coding are defined through binary matrices, and a load balance and spatiotemporal coupling constraint model is established; Based on the double-chain chromosome encoding structure, a multi-objective weighted fitness function is designed, and adaptive crossover probability and tabu search are introduced to assist grouping; Verify the optimality of the solution through the digital twin interface and establish elastic adjustment triggering rules for vehicle failures and traffic congestion; The loading and unloading operation sequence specifically includes: Based on the key resource data predicted by the resource demand forecasting model, a four-dimensional space-time cube is constructed to represent the loading and unloading resource status; A time window penalty model is established through mixed integer programming, and a column generation algorithm is used to minimize the total loading and unloading cost and accelerate the generation of feasible solutions. Based on cargo data, the priority of cargo is dynamically adjusted.

5. The method for intelligent prediction and scheduling of logistics transportation based on digital twins according to claim 1 is characterized in that: The training and optimization of the logistics and transportation digital twin model based on the feedback data and scheduling execution deviation data during the logistics and transportation process to form a closed-loop optimization system specifically includes: Based on real-time collected data and resource demand forecasting model forecast data, extract multi-dimensional deviation data and extract deviation data features; Retrain the digital twin model through the collected feedback data and deviation data features to continuously optimize the model parameters; Based on the trained digital twin model, key indicators in the logistics transportation process are predicted in real time. The model automatically feedbacks the deviations generated during the scheduling execution process and further adjusts the model and scheduling strategy.

6. Combined with the digital twin-based logistics and transportation intelligent prediction and scheduling method, it is used to implement the digital twin-based logistics and transportation intelligent prediction and scheduling system as described in any one of claims 1 to 5, characterized in that: include: IoT sensing module: This module is mainly used to collect real-time data on the location, working conditions, environment, and cargo status of transport vehicles, and achieve unified access of heterogeneous devices through a multi-protocol adapter; Digital Twin Module: This module is primarily used to perform spatiotemporal alignment and feature-level fusion of multi-source heterogeneous data, construct a four-dimensional data matrix and extract key eigenvectors. It also builds a high-precision three-dimensional virtual model based on BIM and Petri nets, and integrates a physics engine to simulate vehicle dynamic characteristics. Simulation deduction module: The simulation deduction module is mainly used to generate future resource requirements by using Monte Carlo parallel simulation and multi-agent interactive deduction; Flexible scheduling module: The flexible scheduling module is mainly used to dynamically optimize the transport group configuration and loading and unloading sequence through mixed integer programming and reinforcement learning; Online optimization module: The online optimization module is mainly used to retrain the digital twin model based on feedback data and optimize model parameters; Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

7. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the intelligent prediction and scheduling method for logistics transportation based on digital twins as described in any one of claims 1-5.

8. A computer-readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, the digital twin-based intelligent prediction and scheduling method for logistics transportation described in any one of claims 1 to 5 is implemented.

Images