Logistics order management optimization method and device, equipment and storage medium

Abstract

The invention belongs to the technical field of logistics management, and discloses a logistics order management optimization method and device, equipment and a storage medium. The method comprises the following steps: collecting multi-source logistics data in real time through a multi-source heterogeneous sensor; performing space-time dimension feature extraction and correlation analysis on the multi-source logistics data based on a space-time diagram neural network to generate a feature tensor; performing multi-objective optimization by adopting an improved hybrid swarm intelligence optimization algorithm in combination with the feature tensor to generate a plurality of initial optimization schemes for logistics order management; screening out a target optimization scheme from the initial optimization schemes by using a virtual logistics network constructed by a digital twinning technology; in the delivery process, the order execution state is monitored in real time through the Internet of Things equipment, and when an abnormal condition is detected, secondary optimization is performed on the target optimization scheme, so that heterogeneous data streams can be synchronously processed, the optimization time consumption is reduced, secondary optimization can be performed on the optimization scheme for an emergent abnormal condition, and the optimization efficiency is improved. And the logistics order management efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics management, and particularly to an optimization method, device, equipment and storage medium for logistics order management. Background Art

[0002] Modern logistics refers to a new type of integrated management that combines logistics activities such as information, transportation, warehousing, inventory, handling, and packaging. Its task is to minimize the total cost of logistics as much as possible and provide the best service for customers. The e-commerce of logistics provides relevant information of the logistics industry in the form of the Internet, including freight information, air freight information, land freight information, sea freight information, as well as logistics industry news, logistics knowledge, laws and regulations, etc. It also provides an enterprise database of the logistics industry for suppliers to search, and suppliers can also publish supply information through the logistics network for logistics enterprises to cooperate. Among them, the management of logistics orders plays a crucial role in modern logistics and can significantly improve the efficiency of modern logistics.

[0003] However, the current logistics order management cannot synchronously process heterogeneous data streams such as order attributes, traffic conditions, and resource status. When orders surge, the optimization of order allocation takes a long time, and there is no secondary optimization for sudden abnormal situations.

[0004] The above content is only used to assist in understanding the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of the present invention is to provide an optimization method, device, equipment and storage medium for logistics order management, aiming to solve the technical problems that the current logistics order management cannot synchronously process heterogeneous data streams such as order attributes, traffic conditions, and resource status, takes a long time to optimize order allocation when orders surge, and there is no secondary optimization for sudden abnormal situations.

[0006] To achieve the above purpose, the present invention provides an optimization method for logistics order management, and the optimization method for logistics order management includes the following steps:

[0007] Real-time collect multi-source logistics data through multi-source heterogeneous sensors, where the multi-source data is composed of logistics order data and dynamic data associated with logistics. The logistics order data includes order attributes, delivery addresses, volume and weight of goods, and time limit requirements specified by users. The dynamic data associated with logistics includes real-time traffic conditions, inventory capacity of warehousing nodes, and dynamic data of delivery vehicles / staff;

[0008] Based on a spatio-temporal graph neural network, perform spatio-temporal dimensional feature extraction and correlation analysis on the multi-source logistics data to generate a feature tensor;

[0009] An improved hybrid swarm intelligence optimization algorithm is used to perform multi-objective optimization in combination with the feature tensor to generate several initial optimization schemes for logistics order management, and the optimization schemes include order allocation, distribution route planning, and cost optimization;

[0010] A virtual logistics network is constructed based on digital twin technology, and the virtual logistics network is used to dynamically simulate and predict each initial optimization scheme to screen out the target optimization scheme from several initial optimization schemes;

[0011] During the distribution process, the order execution status is monitored in real time through Internet of Things devices. When traffic anomalies, order changes, or resource failures are detected, a dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization scheme.

[0012] In some embodiments, the spatio-temporal dimension feature extraction and correlation analysis of the multi-source logistics data by the spatio-temporal graph neural network to generate a feature tensor includes:

[0013] The logistics network is abstracted into a spatio-temporal graph structure, and the nodes of the spatio-temporal graph structure include warehouses, distribution centers, and transportation tools, and the edge weights include real-time distance, travel time, and resource occupancy cost;

[0014] The dual-channel attention mechanism is used to obtain the periodic features in the time dimension and the topological correlation features in the space dimension from the spatio-temporal graph structure respectively;

[0015] A dynamic graph convolutional layer is used to fuse the periodic features in the time dimension and the topological correlation features in the space dimension and output a dimension-compressed feature tensor.

[0016] In some embodiments, the dual-channel attention mechanism uses an LSTM network in the time channel to extract time series features and introduces periodic attention weights, and the spatial channel models the topological relationship between nodes through a graph attention network to dynamically adjust the spatial correlation degree weights, and integrates the spatio-temporal dual-channel features through a gated fusion module.

[0017] In some embodiments, the improved hybrid swarm intelligence optimization algorithm fuses the ant colony algorithm and the particle swarm optimization algorithm. The use of the improved hybrid swarm intelligence optimization algorithm in combination with the feature tensor for multi-objective optimization includes:

[0018] The global distribution route planning is carried out through the ant colony algorithm in combination with the feature tensor, and the order allocation is processed by using the particle swarm optimization algorithm in combination with the feature tensor,

[0019] A dynamic inertia weight mechanism is used to adjust the cost weight of the feature tensor;

[0020] Based on a multi-objective evaluation function, order allocation, distribution route planning, and optimization of costs are synchronously carried out in combination with global distribution route planning, order allocation, and adjusted cost weights.

[0021] In some embodiments, the dynamic inertia weight mechanism is W(t) = W1 - (W1 - W s ) × (t / T max ), where t is the current iteration number and T max is the maximum number of iterations.

[0022] In some embodiments, a virtual logistics network is constructed based on digital twin technology, and the virtual logistics network is used to dynamically simulate and predict each initial optimization plan to screen out a target optimization plan from several initial optimization plans, including:

[0023] Construct a virtual logistics network based on digital twin technology, and the virtual logistics network integrates real-time traffic flow data and weather environment parameters;

[0024] Dynamically simulate and predict the potential risk probabilities of different initial optimization plans through Monte Carlo simulation;

[0025] Take the initial optimization plan with the lowest potential risk probability as the target optimization plan.

[0026] In some embodiments, the dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization plan, including:

[0027] Perform local adjustment on the target optimization plan at the edge nodes of the virtual logistics network;

[0028] If the local adjustment cannot meet the constraint conditions, an incremental optimization algorithm is used to recalculate the affected sub-virtual logistics network to perform secondary optimization on the target optimization plan, where the affected sub-virtual logistics network is a circular range with an abnormal node as the center and a preset radius.

[0029] In addition, to achieve the above object, the present invention also proposes a logistics order management optimization device, and the logistics order management optimization device includes:

[0030] An acquisition module, configured to collect multi-source logistics data in real time through multi-source heterogeneous sensors, where the multi-source data is composed of logistics order data and logistics-related dynamic data, the logistics order data includes order attributes, delivery addresses, cargo volume and weight, and time limit requirements specified by users, and the logistics-related dynamic data includes real-time traffic conditions, inventory capacity of warehouse nodes, and dynamic data of delivery vehicles / staff;

[0031] An extraction module, configured to perform spatio-temporal dimensional feature extraction and correlation analysis on the multi-source logistics data based on a spatio-temporal graph neural network, and generate a feature tensor;

[0032] A generation module, configured to perform multi-objective optimization by using an improved hybrid swarm intelligence optimization algorithm in combination with the feature tensor to generate a number of initial optimization schemes for logistics order management, where the optimization schemes include order allocation, distribution path planning, and cost optimization;

[0033] A screening module, configured to construct a virtual logistics network based on digital twin technology, and perform dynamic simulation prediction on each initial optimization scheme by using the virtual logistics network, so as to screen out a target optimization scheme from a number of initial optimization schemes;

[0034] An optimization module, configured to, during the distribution process, monitor the order execution status in real time through Internet of Things devices, and when traffic anomalies, order changes, or resource failures are detected, trigger a dynamic re-optimization mechanism to perform secondary optimization on the target optimization scheme.

[0035] In addition, to achieve the above object, the present invention further provides a logistics order management optimization device, where the logistics order management optimization device includes: a memory, a processor, and a logistics order management optimization program stored on the memory and executable on the processor, and the logistics order management optimization program is configured to implement the steps of the logistics order management optimization method as described above.

[0036] In addition, to achieve the above object, the present invention further provides a storage medium, where a logistics order management optimization program is stored on the storage medium, and when the logistics order management optimization program is executed by a processor, the steps of the logistics order management optimization method as described above are implemented.

[0037] The present invention collects multi-source logistics data in real time through multi-source heterogeneous sensors; performs spatio-temporal dimensional feature extraction and correlation analysis on the multi-source logistics data based on a spatio-temporal graph neural network to generate a feature tensor; performs multi-objective optimization by using an improved hybrid swarm intelligence optimization algorithm in combination with the feature tensor to generate a number of initial optimization schemes for logistics order management; uses the virtual logistics network constructed based on digital twin technology to screen out a target optimization scheme from the initial optimization schemes; during the distribution process, monitors the order execution status in real time through Internet of Things devices, and when anomalies are detected, performs secondary optimization on the target optimization scheme, which can synchronously process heterogeneous data streams, reduce the optimization time, and can perform secondary optimization on the optimization scheme for sudden anomalies, improving the efficiency of logistics order management. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic flowchart of the first embodiment of the logistics order management optimization method of the present invention;

[0039] Figure 2 It is a structural block diagram of the first embodiment of the logistics order management optimization device of the present invention.

[0040] The realization of the purpose, functional characteristics and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Specific embodiments

[0041] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0042] The embodiment of the present invention provides a logistics order management optimization method, with reference to Figure 1 , Figure 1 It is a flowchart of the first embodiment of a logistics order management optimization method of the present invention.

[0043] In this embodiment, the logistics order management optimization method includes the following steps:

[0044] Step S10: Real-time collect multi-source logistics data through multi-source heterogeneous sensors.

[0045] In this embodiment, the execution subject of this embodiment is a logistics order management optimization device. Among them, the logistics order management optimization device has functions such as data processing, data communication and program operation. The logistics order management optimization device can be a computer terminal device or other network devices. Of course, it can also be other devices with similar functions. This embodiment does not limit this.

[0046] It should be noted that the current logistics order management cannot synchronously process heterogeneous data streams such as order attributes, traffic conditions, and resource status. When the number of orders surges, the optimization of order allocation takes a long time, and there is no secondary optimization for sudden abnormal situations.

[0047] To solve the above technical problems, in this embodiment, multi-source heterogeneous sensors are used to collect multi-source logistics data in real time; based on the spatio-temporal graph neural network, spatio-temporal dimensional feature extraction and correlation analysis are performed on the multi-source logistics data to generate feature tensors; an improved hybrid swarm intelligence optimization algorithm is used to combine the feature tensors for multi-objective optimization to generate several initial optimization schemes for logistics order management; the virtual logistics network constructed by digital twin technology is used to screen out the target optimization scheme from the initial optimization schemes; during the distribution process, the order execution status is monitored in real time through Internet of Things devices. When an abnormal situation is detected, the target optimization scheme is secondarily optimized, which can synchronously process heterogeneous data streams, reduce the optimization time, and can secondarily optimize the optimization scheme for sudden abnormal situations, improving the efficiency of logistics order management. Specifically, it can be realized in the following manner.

[0048] In a specific implementation, in this embodiment, multi-source logistics data is collected in real time through multi-source heterogeneous sensors. Among them, the multi-source data consists of logistics order data and dynamically related logistics data. The logistics order data includes order attributes, delivery addresses, cargo volume and weight, and the time limit requirements specified by users. The dynamically related logistics data includes real-time traffic conditions, inventory capacity of warehouse nodes, and dynamic data of delivery vehicles / staff. The multi-source heterogeneous sensors in this embodiment include video monitoring sensors, radio frequency identification (RFID) sensors, lidar (LiDAR), global positioning system (GPS), inertial measurement unit (IMU), and ultrasonic sensors, etc.

[0049] Step S20: Based on the spatio-temporal graph neural network, perform spatio-temporal dimensional feature extraction and correlation analysis on the multi-source logistics data to generate a feature tensor.

[0050] In a specific implementation, the logistics network can be abstracted as a spatio-temporal graph structure. The nodes of the spatio-temporal graph structure include warehouses, distribution centers, and transportation tools. The edge weights include real-time distance, travel time, and resource occupancy cost. Then, the dual-channel attention mechanism is used to obtain the periodic features in the time dimension and the topological correlation features in the space dimension from the spatio-temporal graph structure respectively; the dynamic graph convolutional layer is used to fuse the periodic features in the time dimension and the topological correlation features in the space dimension, and output a feature tensor with compressed dimensions.

[0051] It should be noted that the dual-channel attention mechanism uses an LSTM network in the time channel to extract time series features and introduces periodic attention weights. In the space channel, the topological relationship between nodes is modeled through a graph attention network, and the spatial correlation degree weights are dynamically adjusted. The spatio-temporal dual-channel features are integrated through a gated fusion module.

[0052] Furthermore, the construction process of the dynamic graph convolutional layer in this embodiment is to define the update rule of the dynamic adjacency matrix, update the graph structure every hour according to the real-time traffic data, adopt the residual connection and batch normalization techniques to prevent gradient disappearance, and compress the dimension of the output feature tensor to 15%-20% of the original data space.

[0053] Specifically, a temporal convolutional layer (such as TCN) is used to process time series data. TCN ensures the sequentiality of time series through causal convolution and uses dilated convolution to increase the receptive field and capture long-range temporal dependencies. A graph convolutional layer (such as GCN) is used to process spatial topology data. GCN updates the feature representation of the current node by aggregating information from neighboring nodes. The features output by the temporal convolutional layer and the graph convolutional layer are fused. This can be achieved in the following ways: Concatenation: Concatenate the temporal features and spatial features in the dimension; Element-wise Addition: Directly add the temporal features and spatial features; Attention Mechanism: Use the attention mechanism to assign different weights to the temporal features and spatial features and perform weighted summation. A fully connected layer or a pooling layer is used to compress the dimensions of the fused features, reduce the number of parameters, and extract the most important features. After dimension compression, a feature tensor with a lower dimension is output, and this tensor will be used for subsequent tasks such as classification, regression, or prediction.

[0054] Step S30: An improved hybrid swarm intelligence optimization algorithm is adopted to combine with the feature tensor for multi-objective optimization to generate several initial optimization schemes for logistics order management.

[0055] In this embodiment, the hybrid swarm intelligence optimization algorithm is a fusion algorithm combining the ant colony algorithm and the particle swarm optimization algorithm. The global delivery path planning is carried out by combining the ant colony algorithm with the feature tensor, and the order allocation is processed by combining the particle swarm optimization algorithm with the feature tensor. The dynamic inertia weight mechanism is used to adjust the cost weight of the feature tensor; based on the multi-objective evaluation function, the order allocation, delivery path planning, and cost optimization are synchronized by combining the global delivery path planning, order allocation, and the adjusted cost weight.

[0056] It should be noted that the dynamic inertia weight mechanism is W(t) = W1 - (W1 - W s ) × (t / T max ), where t is the current iteration number, T max is the maximum iteration number, W1 is the initial weight value, and W s is the final weight value.

[0057] After the cost weight iteration, optimization is carried out using a multi-objective evaluation function. The multi-objective evaluation function is a multi-objective evaluation function based on the Pareto front. In multi-objective optimization, the goal is to find a set of solutions (called the Pareto optimal solution set), and these solutions cannot be improved simultaneously (that is, improving in one objective will not get worse in other objectives). Order allocation corresponds to the service quality objective, distribution path planning corresponds to the time objective, and cost optimization corresponds to the cost objective. Specifically, the following process can be referred to:

[0058] # Define the multi-objective evaluation function def evaluation_function(solution, cost_weights):

[0059] total_cost = 0

[0060] total_time = 0

[0061] total_satisfaction = 0

[0062] # Calculate the value of each objective

[0063] for route in solution.routes:

[0064] total_cost += calculate_cost(route) * cost_weights['cost']

[0065] total_time += calculate_time(route) * cost_weights['time']

[0066] total_satisfaction += calculate_satisfaction(route) * cost_weights['satisfaction']

[0067] # Return the multi-objective evaluation result

[0068] return total_cost, total_time, total_satisfaction

[0069] # Initialize the parameters

[0070] cost_weights = {'cost': 0.5, 'time': 0.3,'satisfaction': 0.2}

[0071] initial_solution = generate_initial_solution(orders, vehicles)

[0072] # Optimization process

[0073] for iteration in range(max_iterations):

[0074] # Order assignment

[0075] orders_assignment = order_assignment(orders, vehicles)

[0076] # Delivery route planning

[0077] routes_planning = route_planning(orders_assignment, vehicles)

[0078] # Cost weight adjustment

[0079] cost_weights = adjust_cost_weights(cost_weights, current_performance)

[0080] # Evaluate the current solution

[0081] current_cost, current_time, current_satisfaction = evaluation_function(routes_planning, cost_weights)

[0082] # Update the optimal solution

[0083] if current_cost < best_cost: best_solution = routes_planning; best_cost =

[0084] current_cost

[0085] # Conduct local search or iterative improvement

[0086] improved_solution = local_search(best_solution)

[0087] # Update the solution: best_solution = improved_solution

[0088] # Output the optimal solution: print(best_solution)

[0089] It should be noted that order allocation refers to allocating orders to deliverymen based on information such as order priority, time window, and geographical location. Delivery route planning refers to planning delivery routes for each deliveryman on the basis of order allocation, considering factors such as the shortest route and minimized waiting time. Cost weight adjustment refers to adjusting the weights of different costs according to the actual situation to achieve a balance among multiple objectives.

[0090] Step S40: Construct a virtual logistics network based on digital twin technology, and use the virtual logistics network to dynamically simulate and predict each initial optimization plan, so as to screen out the target optimization plan from several initial optimization plans.

[0091] In specific implementation, a virtual logistics network is constructed based on digital twin technology; the potential risk probability of different initial optimization plans is dynamically simulated and predicted through Monte Carlo simulation; the initial optimization plan with the lowest potential risk probability is used as the target optimization plan.

[0092] It should be noted that the virtual logistics network integrates real-time traffic flow data and weather environment parameters, and the Monte Carlo technique is used in the simulation process. Specifically, from the dimensions of three types of risk events: traffic congestion, weather mutation, and equipment failure, the probability corresponding to each event in each plan can be obtained, and finally the potential risk probability of each plan can be obtained through weighted calculation. The initial optimization plan with the lowest potential risk probability is used as the target optimization plan.

[0093] Step S50: During the delivery process, the execution status of the order is monitored in real time through Internet of Things devices. When traffic anomalies, order changes, or resource failures are detected, a dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization plan.

[0094] In specific implementation, a re-optimization mechanism is provided in this embodiment. When traffic anomalies, order changes, or resource failures occur, the dynamic re-optimization mechanism is triggered. The specific process is to locally adjust the target optimization plan at the edge nodes of the virtual logistics network; if the local adjustment cannot meet the constraint conditions, an incremental optimization algorithm is used to recalculate the affected sub-virtual logistics network to perform secondary optimization on the target optimization plan.

[0095] It should be noted that when the above abnormal situation occurs, while adjusting the order allocation, the delivery route will also be adjusted. The delivery route corresponds to each node of the virtual logistics network. To improve the re-optimization efficiency, local adjustment is performed on the target optimization plan at the edge nodes of the virtual logistics network, that is, the order allocation of some nodes is changed without affecting the overall delivery route. The process of local adjustment can also refer to the multi-objective function optimization described above. If the local adjustment cannot meet the constraint conditions, an incremental optimization algorithm is used to recalculate the affected sub-virtual logistics network to perform secondary optimization on the target optimization plan. Among them, the affected sub-virtual logistics network is a circular range formed by taking the abnormal node as the center and a preset radius. For example, nodes within a radius of 5 kilometers from the affected area are defined as the optimization scope.

[0096] In this embodiment, multi-source heterogeneous sensors are used to collect multi-source logistics data in real time; spatio-temporal dimension feature extraction and correlation analysis are performed on the multi-source logistics data based on a spatio-temporal graph neural network to generate a feature tensor; an improved hybrid swarm intelligence optimization algorithm is used to perform multi-objective optimization in combination with the feature tensor to generate several initial optimization plans for logistics order management; a virtual logistics network constructed by digital twin technology is used to screen out the target optimization plan from the initial optimization plans; during the delivery process, the order execution status is monitored in real time through Internet of Things devices. When an abnormal situation is detected, secondary optimization is performed on the target optimization plan, which can synchronously process heterogeneous data streams, reduce the optimization time, and can perform secondary optimization on the optimization plan for sudden abnormal situations, improving the efficiency of logistics order management.

[0097] In addition, an embodiment of the present invention also proposes a storage medium, on which a logistics order management optimization program is stored. When the logistics order management optimization program is executed by a processor, the steps of the logistics order management optimization method described above are implemented.

[0098] Refer to Figure 2 , Figure 2 which is the structural block diagram of the first embodiment of the logistics order management optimization device of the present invention.

[0099] As Figure 2 shown, the logistics order management optimization device proposed by the embodiment of the present invention includes:

[0100] An acquisition module 10, configured to collect multi-source logistics data in real time through multi-source heterogeneous sensors. The multi-source data consists of logistics order data and dynamic data related to logistics. The logistics order data includes order attributes, delivery addresses, cargo volume and weight, and time limit requirements specified by users. The dynamic data related to logistics includes real-time traffic conditions, inventory capacity of warehouse nodes, and dynamic data of delivery vehicles / staff;

[0101] The extraction module 20 is used to perform spatio-temporal dimensional feature extraction and correlation analysis on the multi-source logistics data based on the spatio-temporal graph neural network to generate a feature tensor;

[0102] The generation module 30 is used to perform multi-objective optimization by combining the improved hybrid swarm intelligence optimization algorithm with the feature tensor to generate several initial optimization schemes for logistics order management, and the optimization schemes include order allocation, distribution route planning and cost optimization;

[0103] The screening module 40 is used to construct a virtual logistics network based on digital twin technology, and use the virtual logistics network to perform dynamic simulation prediction on each initial optimization scheme to screen out the target optimization scheme from several initial optimization schemes;

[0104] The optimization module 50 is used to, during the distribution process, real-time monitor the order execution status through Internet of Things devices. When traffic anomalies, order changes or resource failures are detected, a dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization scheme.

[0105] In this embodiment, multi-source heterogeneous sensors are used to collect multi-source logistics data in real time; spatio-temporal dimensional feature extraction and correlation analysis are performed on the multi-source logistics data based on the spatio-temporal graph neural network to generate a feature tensor; the improved hybrid swarm intelligence optimization algorithm is used to perform multi-objective optimization in combination with the feature tensor to generate several initial optimization schemes for logistics order management; the virtual logistics network constructed by digital twin technology is used to screen out the target optimization scheme from the initial optimization schemes; during the distribution process, the order execution status is real-time monitored through Internet of Things devices. When abnormal conditions are detected, secondary optimization is performed on the target optimization scheme, which can synchronously process heterogeneous data streams, reduce the optimization time, and can perform secondary optimization on the optimization scheme for sudden abnormal conditions, improving the efficiency of logistics order management.

[0106] In some embodiments, the extraction module 20 is used to abstract the logistics network into a spatio-temporal graph structure. The nodes of the spatio-temporal graph structure include warehouses, distribution centers, and transportation tools, and the edge weights include real-time distance, travel time, and resource occupancy cost;

[0107] The dual-channel attention mechanism is used to obtain the periodic features in the time dimension and the topological correlation features in the space dimension from the spatio-temporal graph structure respectively;

[0108] The dynamic graph convolutional layer is used to fuse the periodic features in the time dimension and the topological correlation features in the space dimension and output a dimension-compressed feature tensor.

[0109] In some embodiments, the dual-channel attention mechanism uses an LSTM network in the time channel to extract time series features and introduces periodic attention weights. In the spatial channel, it models the topological relationship between nodes through a graph attention network, dynamically adjusts the spatial correlation weight, and integrates the spatio-temporal dual-channel features through a gated fusion module.

[0110] In some embodiments, the improved hybrid swarm intelligence optimization algorithm fuses the ant colony algorithm and the particle swarm optimization algorithm. The generation module 30 is used to perform global delivery path planning through the ant colony algorithm in combination with the feature tensor, and to process order allocation using the particle swarm optimization algorithm in combination with the feature tensor.

[0111] Adjust the cost weight of the feature tensor using a dynamic inertia weight mechanism.

[0112] Based on a multi-objective evaluation function, synchronously optimize order allocation, delivery path planning, and cost in combination with global delivery path planning, order allocation, and the adjusted cost weight.

[0113] In some embodiments, the dynamic inertia weight mechanism is W(t) = W1 - (W1 - W s ) × (t / T max ), where t is the current iteration number and T max is the maximum iteration number.

[0114] In some embodiments, the screening module 40 is used to construct a virtual logistics network based on digital twin technology, and the virtual logistics network integrates real-time traffic flow data and weather environment parameters.

[0115] Dynamically simulate and predict the potential risk probability of different initial optimization schemes through Monte Carlo simulation.

[0116] Take the initial optimization scheme with the lowest potential risk probability as the target optimization scheme.

[0117] In some embodiments, the optimization module 50 is used to perform local adjustment of the target optimization scheme at the edge nodes of the virtual logistics network.

[0118] If the local adjustment cannot meet the constraint conditions, an incremental optimization algorithm is used to recalculate the affected sub-virtual logistics network to perform secondary optimization on the target optimization scheme, where the affected sub-virtual logistics network is a circular range centered on the abnormal node with a preset radius.

[0119] The embodiment of the present application also provides an optimized device for logistics order management, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The memory is used to store the optimized program for logistics order management; the processor is used to implement the above-mentioned optimized method for logistics order management when executing the program stored on the memory.

[0120] The communication bus mentioned in the above-mentioned optimized device for logistics order management may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0121] The communication interface is used for communication between the above-mentioned optimized device for logistics order management and other devices.

[0122] The memory may include a Random Access Memory (RAM), or may also include a Non-Volatile Memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located far from the aforementioned processor.

[0123] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0124] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

[0125] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device that includes a series of elements includes not only those elements but also other elements that are not explicitly listed, or elements that are inherent to such process, method, article, or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article, or device that includes the element.

[0126] Each embodiment in this specification is described in a related manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and reference can be made to the corresponding parts of the method embodiments for the relevant content.

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

[0128] It should be understood that the above is only for illustration and does not constitute any limitation to the technical solutions of the present invention. In specific applications, those skilled in the art can set according to needs, and the present invention does not limit this.

[0129] It should be noted that the above-described work process is only illustrative and does not constitute a limitation to the protection scope of the present invention. In actual applications, those skilled in the art can select some or all of them according to actual needs to achieve the purpose of the solution of this embodiment, and there is no limitation here.

[0130] In addition, for the technical details not described in detail in this embodiment, reference can be made to the logistics order management optimization method provided in any embodiment of the present invention, and details will not be repeated here.

[0131] In addition, it should be noted that in this article, the terms "including", "comprising" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or system. Without further limitation, an element defined by the phrase "including a..." does not exclude the existence of additional identical elements in the process, method, article or system including that element.

[0132] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0133] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solutions of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as a read-only memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions to enable a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0134] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be similarly included in the patent protection scope of the present invention.

[0135] It can be understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention. The explanations, examples and beneficial effects of related content can refer to the corresponding parts in the above method.

Claims

1. A logistics order management optimization method, characterized in that: The logistics order management optimization method comprises: Real-time collection of multi-source logistics data through multi-source heterogeneous sensors. The multi-source data consists of logistics order data and dynamic data associated with logistics. The logistics order data includes order attributes, delivery address, cargo volume and weight, and user-specified timeliness requirements. The dynamic data associated with logistics includes real-time traffic conditions, storage node inventory capacity, and dynamic data of delivery vehicles / personnel. Based on the spatiotemporal graph neural network, the spatiotemporal dimension features of the multi-source logistics data are extracted and associated, and a feature tensor is generated; Using an improved hybrid swarm intelligent optimization algorithm combined with the feature tensor to perform multi-objective optimization to generate several initial optimization schemes for logistics order management, the optimization schemes including order allocation, delivery route planning and cost optimization; Building a virtual logistics network based on digital twin technology, and using the virtual logistics network to dynamically simulate and predict each initial optimization plan, so as to select a target optimization plan from a number of initial optimization plans; During the delivery process, the order execution status is monitored in real time through IoT devices. When traffic anomalies, order changes or resource failures are detected, the dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization plan.

2. The logistics order management optimization method according to claim 1, characterized in that: The spatiotemporal dimension feature extraction and correlation analysis of the multi-source logistics data based on the spatiotemporal graph neural network to generate a feature tensor includes: The logistics network is abstracted into a space-time graph structure, the nodes of which include warehouses, distribution centers, and transportation tools, and the edge weights include real-time distance, travel time, and resource occupation cost; A dual-channel attention mechanism is used to obtain the periodic features of the time dimension and the topological correlation features of the space dimension from the spatiotemporal graph structure respectively; A dynamic graph convolution layer is used to fuse the periodic features of the time dimension and the topological association features of the space dimension, and output a dimensionally compressed feature tensor.

3. The logistics order management optimization method according to claim 2, characterized in that: The dual-channel attention mechanism uses an LSTM network to extract time series features in the time channel and introduces periodic attention weights. The spatial channel uses a graph attention network to model the topological relationship between nodes, dynamically adjusts the spatial correlation weights, and integrates the spatiotemporal dual-channel features through a gated fusion module.

4. The logistics order management optimization method according to claim 1, characterized in that: The improved hybrid swarm intelligent optimization algorithm integrates the ant colony algorithm and the particle swarm optimization algorithm. The improved hybrid swarm intelligent optimization algorithm is combined with the feature tensor to perform multi-objective optimization, including: Performing global delivery route planning by combining the ant colony algorithm with the feature tensor, and processing order allocation by combining the particle swarm optimization algorithm with the feature tensor; adjusting the cost weight of the feature tensor using a dynamic inertia weight mechanism; Based on the multi-objective evaluation function, global delivery path planning, order allocation and adjusted cost weights are combined to simultaneously optimize order allocation, delivery path planning and cost.

5. The logistics order management optimization method according to claim 4, characterized in that: The dynamic inertia weight mechanism is W(t)=W1-(W1-W s )×(t / T max ), t is the current iteration number, T max is the maximum number of iterations.

6. The logistics order management optimization method according to claim 1, characterized in that: The virtual logistics network is constructed based on the digital twin technology, and the virtual logistics network is used to dynamically simulate and predict each initial optimization scheme to select a target optimization scheme from a number of initial optimization schemes, including: Building a virtual logistics network based on digital twin technology, the virtual logistics network integrates real-time traffic flow data and weather environment parameters; Predict the potential risk probability of different initial optimization solutions through Monte Carlo simulation dynamic simulation; The initial optimization plan with the lowest potential risk probability is selected as the target optimization plan.

7. The logistics order management optimization method according to any one of claims 1 to 6, characterized in that: The triggering of the dynamic re-optimization mechanism to perform secondary optimization on the target optimization scheme includes: Locally adjusting the target optimization scheme at the edge nodes of the virtual logistics network; If the local adjustment cannot meet the constraint conditions, an incremental optimization algorithm is used to recalculate the affected sub-virtual logistics network to perform secondary optimization on the target optimization scheme, wherein the affected sub-virtual logistics network is a circular range consisting of the abnormal node as the center point and a preset radius.

8. A logistics order management optimization device, characterized in that: The logistics order management optimization device comprises: A collection module is used to collect multi-source logistics data in real time through multi-source heterogeneous sensors. The multi-source data consists of logistics order data and dynamic data associated with logistics. The logistics order data includes order attributes, delivery address, cargo volume and weight, and timeliness requirements specified by the user. The dynamic data associated with logistics includes real-time traffic conditions, storage node inventory capacity, and dynamic data of delivery vehicles / personnel. An extraction module, used to extract and analyze the spatiotemporal dimension features of the multi-source logistics data based on a spatiotemporal graph neural network to generate a feature tensor; A generation module, used to use an improved hybrid swarm intelligent optimization algorithm combined with the feature tensor to perform multi-objective optimization to generate several initial optimization schemes for logistics order management, wherein the optimization schemes include order allocation, delivery route planning, and cost optimization; A screening module is used to build a virtual logistics network based on digital twin technology, and use the virtual logistics network to dynamically simulate and predict each initial optimization plan to screen out a target optimization plan from a number of initial optimization plans; The optimization module is used to monitor the order execution status in real time through IoT devices during the delivery process. When traffic anomalies, order changes or resource failures are detected, a dynamic re-optimization mechanism is triggered to perform secondary optimization on the target optimization plan.

9. A logistics order management optimization device, characterized in that: The logistics order management optimization device includes: a memory, a processor, and a logistics order management optimization program stored in the memory and executable on the processor, wherein the logistics order management optimization program is configured to implement the steps of the logistics order management optimization method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a logistics order management optimization program, and when the logistics order management optimization program is executed by the processor, the steps of the logistics order management optimization method according to any one of claims 1 to 7 are implemented.

Images