Intelligent allocation method and system for logistics resources based on operation data analysis

Abstract

The application discloses a logistics resource intelligent deployment method and system based on operation data analysis, and relates to the technical field of resource management.The method comprises the following steps: performing network slicing processing on the computing network infrastructure to determine the computing network resource architecture; performing dynamic space-time folding of the logistics resource end-computing network resource architecture, building a lightweight deployment module, injecting disturbance factors for supervised training, and embedding the lightweight deployment module in the logistics management platform; performing resource tag mapping on the logistics resource end; obtaining a resource deployment task, triggering the lightweight deployment module to perform resource folding processing, obtaining a task association tag, performing resource deployment decision and deployment compensation under random disturbance, determining a resource deployment strategy, visualizing the platform interface, and performing terminal issuing management.The technical problems of low logistics resource deployment efficiency, delayed response, and low resource utilization in the prior art are solved, and the technical effects of improving the intelligent level of resource deployment and realizing efficient resource utilization are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of resource management, in particular to a logistics resource intelligent deployment method and system based on operation data analysis. BACKGROUND

[0002] With the expansion of the operation scale of enterprises and large organizations, the logistics resource management is facing problems such as diverse resource types, dynamic demand changes, low deployment efficiency, and fragmented operation data. Traditional logistics management systems mostly use static configuration and manual scheduling, lack intelligent analysis capability of real-time operation data, and are difficult to respond to sudden demand and optimize resource utilization. At the same time, there are complex logical dependencies and space-time correlations between logistics resources, and traditional methods cannot realize precise matching and intelligent collaboration of resources, resulting in resource redundancy, deployment conflicts, and increased overall operation cost. SUMMARY

[0003] The present application provides a logistics resource intelligent deployment method and system based on operation data analysis, which solves the technical problems of low logistics resource deployment efficiency, delayed response, and low resource utilization in the prior art.

[0004] In a first aspect, the present application provides a logistics resource intelligent deployment method based on operation data analysis, which comprises:

[0005] For the logistics management platform, the algorithm network infrastructure is subjected to network slicing processing, and the algorithm network resource architecture is determined, wherein each logistics resource end corresponds to a logical network; after the dynamic space-time folding of the logistics resource end-algorithm network resource architecture, a lightweight deployment module is built and injected with disturbance factors for supervised training, and is embedded in the logistics management platform; a virtual label layer is introduced to map the resource labels of the logistics resource end; the logistics management platform obtains a resource deployment task, triggers the lightweight deployment module to perform task-oriented resource folding processing, obtains task-associated labels through a cascaded virtual label layer, performs resource deployment decision and deployment compensation under random disturbance, determines a resource deployment strategy, visualizes it on the platform interface, and performs terminal issuance management.

[0006] In a second aspect, the present application provides a logistics resource intelligent deployment system based on operation data analysis, which comprises:

[0007] The slice processing unit: for the logistics management platform, the network slice processing is performed on the computing network infrastructure, and the computing network resource architecture is determined, wherein each logistics resource end corresponds to a logical network; the training unit: after the dynamic space-time folding of the logistics resource end-computing network resource architecture, a lightweight deployment module is built and a disturbance factor is injected for supervised training, and the embedded deployment is performed on the logistics management platform; the label mapping unit: a virtual label layer is introduced, and resource label mapping is performed on the logistics resource end; the management unit: the logistics management platform obtains a resource deployment task, triggers the lightweight deployment module to perform task-oriented resource folding processing, obtains a task associated label through a cascaded virtual label layer, performs resource deployment decision and deployment compensation under random disturbance, determines a resource deployment strategy, visualizes on a platform interface, and performs terminal issuing management.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] The slice processing unit: for the logistics management platform, the network slice processing is performed on the computing network infrastructure, and the computing network resource architecture is determined, wherein each logistics resource end corresponds to a logical network; the training unit: after the dynamic space-time folding of the logistics resource end-computing network resource architecture, a lightweight deployment module is built and a disturbance factor is injected for supervised training, and the embedded deployment is performed on the logistics management platform; the label mapping unit: a virtual label layer is introduced, and resource label mapping is performed on the logistics resource end; the management unit: the logistics management platform obtains a resource deployment task, triggers the lightweight deployment module to perform task-oriented resource folding processing, obtains a task associated label through a cascaded virtual label layer, performs resource deployment decision and deployment compensation under random disturbance, determines a resource deployment strategy, visualizes on a platform interface, and performs terminal issuing management. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0011] Figure 1 The flowchart of the logistics resource intelligent deployment method based on operation data analysis provided by the embodiments of the present application is shown.

[0012] Figure 2 The structural diagram of the logistics resource intelligent deployment system based on operation data analysis provided by the embodiments of the present application is shown.

[0013] Explanation of reference signs: slice processing unit 11, training unit 12, label mapping unit 13, management unit 14. DETAILED DESCRIPTION

[0014] The present application provides a logistics resource intelligent allocation method and system based on operation data analysis, which solves the technical problems of low efficiency, lagging response and low resource utilization in the prior art.

[0015] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0016] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to those clearly listed steps or units, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] Embodiment one, as shown in the present application provides a logistics resource intelligent allocation method based on operation data analysis, wherein the method comprises: Figure 1

[0018] For the logistics management platform, the network slicing processing is performed on the computing power network infrastructure, and the computing network resource architecture is determined, wherein each logistics resource end corresponds to a logical network.

[0019] In the embodiments of the present application, for the running environment and management demand of the logistics management platform, the network slicing processing is performed on the accessed computing power network infrastructure to realize the fine management and on-demand allocation of logistics resources. Specifically, network slicing refers to dividing the underlying computing power network into multiple isolated and customizable logical networks through virtualization technology, and each logical network independently carries specific logistics resource end data transmission and allocation tasks.

[0020] Based on the logistics management platform, a plurality of logistics resource ends are determined, for example, including warehouse terminals, transportation vehicles, repair stations, catering sites and other entity resources, and according to their functional roles and business attributes, they are mapped to corresponding logical management units. Then, combined with the access mode, data flow characteristics, service priority and computing power demand of each logistics resource end, the computing power network is sliced and divided to construct multiple heterogeneous but cooperative logical networks.

[0021] ​Furthermore, network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, including:

[0022] A first logistics resource terminal is determined, wherein the first logistics resource terminal is any one of the logistics resource terminals; for the first logistics resource terminal, an intrinsic allocation logic point is determined; the first logistics resource terminal and each associated logistics resource terminal are combined to determine associated allocation logic points, wherein the associated allocation logic points correspond to each associated logistics resource terminal; the intrinsic allocation logic points and the associated allocation logic points are integrated to perform network resource reconstruction and computing power configuration, which serves as the first logical network of the first logistics resource terminal.

[0023] From multiple logistics resource terminals connected to the logistics management platform, one logistics resource terminal to be allocated is selected as the first logistics resource terminal. The first logistics resource terminal can be a specific functional node, such as a warehousing node, distribution terminal, or maintenance station, serving as the starting point for constructing the allocation logic. For the first logistics resource terminal, its core role and resource dependencies within the allocation system are identified and abstracted, determining its intrinsic allocation logic points. These intrinsic allocation logic points refer to the functional logic directly manifested by the resource terminal in the allocation network, including its basic computing power requirements, resource types, bandwidth usage, and service interaction frequency, used to define its native role in the network slicing structure. Based on business collaboration relationships and operational data analysis, several related logistics resource terminals that interact with or share resources with the first logistics resource terminal are identified. These related logistics resource terminals may have spatial and temporal connections with the first resource terminal in terms of transportation path association, task concurrency, or resource complementarity, necessitating joint allocation. For these related logistics resources, corresponding related allocation logic points are constructed, with each logic point representing its resource behavior model and network allocation characteristics. These related allocation logic points can be obtained by modeling multi-dimensional data such as historical allocation data, communication logs, and resource request frequencies, forming embeddable network abstract components.

[0024] After extracting the intrinsic allocation logic points and multiple associated allocation logic points, the system integrates these associated allocation logic points, including network resource reconstruction and computing resource allocation. This involves mapping these allocation logic points to virtual nodes in the physical network and optimizing the deployment of the logical network topology, inter-node communication paths, and computing node distribution based on real-time computing load, bandwidth capacity, and task priorities. Ultimately, a first logical network corresponding to the first logistics resource end is formed. As part of a network slice, the first logical network encapsulates the complete network structure originating from this resource end and covering its associated allocation domains, serving as the basic unit in the computing network resource architecture.

[0025] Based on the dynamic spatiotemporal folding of the logistics resource-computing network resource architecture, a lightweight allocation module is built and perturbation factors are injected for supervised training, which is then embedded and deployed on the logistics management platform.

[0026] After the computing network resource architecture is determined, the system implements dynamic spatiotemporal folding processing based on the relationship between the logistics resource end and the constructed logical network to further improve resource allocation efficiency and control granularity. Spatiotemporal folding refers to constructing a lightweight allocation model with compressed representation by integrating the temporal evolution characteristics of tasks and the spatial distribution patterns of resources, enabling the system to achieve high-frequency dynamic scheduling under low computational load.

[0027] Specifically, multi-dimensional data is collected from the logistics resource end, including task request flow, resource occupancy status, interaction frequency, geographical location information, and historical task trajectories. Combined with the established computing network resource logic network, this data is structured and modeled according to the correspondence between the "logistics resource end - computing network resource architecture." Then, graph embedding and feature compression methods are used to perform resource status folding operations, thereby generating a task-oriented lightweight allocation module structure. The lightweight allocation module is essentially an end-to-end resource allocation sub-network, characterized by low-latency inference and high adaptability.

[0028] By introducing perturbation factors during the training phase, the robustness of the lightweight allocation module under complex and uncertain scenarios is enhanced. These perturbation factors include sudden task disturbances (such as the temporary insertion of high-priority tasks), abnormal resource states (such as partial node failures or congestion), and communication delay disturbances, simulating dynamic perturbation conditions in actual operation. Specifically, a supervised training set is constructed, and historical allocation samples with perturbation labels are used for training, enabling the model to possess the ability to "perturb - adjust strategy - compensate for allocation".

[0029] After training, the lightweight allocation module is embedded in the logistics management platform and runs in the control logic layer of the resource scheduling hub. When it receives a task allocation request, the lightweight allocation module can quickly complete resource folding determination, tag matching, and allocation strategy output, enabling flexible control of multi-source heterogeneous logistics resources.

[0030] Furthermore, a lightweight allocation module is built and perturbation factors are injected for supervised training, including:

[0031] Using the aforementioned computing network resource architecture as the task interpretation benchmark, a first folded node is deployed; using the selective retrieval of virtual tags as a decision constraint, a second decision node based on folded state decision is deployed; introducing task-oriented random perturbations, a third compensation node is deployed; the first folded node, the second decision node, and the third compensation node are cascaded, and the lightweight allocation module is built through supervised training until convergence.

[0032] First, using the computing network resource architecture as the task interpretation benchmark, the first fold node is deployed. Specifically, the system maps the currently pending logistics tasks to the computing network resource architecture, constructing a matching relationship between tasks and the logical network by parsing task elements (including task type, target resource category, allocation time limit, location requirements, etc.). The first fold node compresses and represents the spatially dispersed and temporally staggered resource states, extracting the key allocation dimensions that have a decisive impact on task execution. Through graph compression, feature embedding, and other methods, the first fold node generates a low-dimensional but highly expressive folded representation, significantly reducing the computational complexity of the model.

[0033] Secondly, a second decision node is deployed, using the selective retrieval of virtual tags as a decision constraint. In this second decision node, the system retrieves tag information highly relevant to the task from the virtual tag layer (including resource flow direction, allocability weight, historical usage patterns, etc.) as boundary conditions for the allocation process. Based on the folded state and tag constraints, the second decision node performs resource optimization and allocation path planning, generating a preliminary allocation strategy.

[0034] Next, a third compensation node is deployed by introducing task-oriented random disturbances. In this third compensation node, several disturbance modes (such as sudden resource unavailability, allocation path interruption, and task objective change) are designed to simulate uncertainties in actual operation, and these disturbance signals are introduced as inputs into the allocation process. The third compensation node is responsible for analyzing the disturbance response of the preliminary strategy output by the second decision node and outputting an optimized allocation compensation strategy to ensure the system has good anti-interference capability and strategy recovery capability.

[0035] Finally, the first folding node, the second decision node, and the third compensation node are cascaded to form a complete lightweight allocation neural network structure. Supervised training is then performed on a dataset with perturbation labels until the model reaches convergence on the validation set. During supervised training, the system continuously iterates and optimizes parameter weights, ensuring that the allocation module maintains a balance between allocation efficiency and resource utilization under various task constraints and perturbation conditions. After training, the lightweight allocation module will be deployed embedded in the logistics management platform to enable real-time invocation and allocation strategy output, meeting the needs of high-frequency, low-latency, and dynamically evolving logistics scenarios.

[0036] A virtual tag layer is introduced to perform resource tag mapping on the logistics resource end.

[0037] Furthermore, the virtual tag layer is a feature layer of resource flow, and the virtual tag layer includes a first tag area based on physical flow and a second tag area based on flow rules.

[0038] The virtual tag layer is a characteristic layer for resource mobility, providing a flexible and dynamic tag mapping mechanism for logistics resources. The virtual tag layer includes a first tag area based on physical flow and a second tag area based on flow rules. The first tag area maps the flow characteristics of logistics resources in physical space, such as geographical location, movement path, and storage location. These tags can be updated based on real-time operational data to reflect the latest physical state of the resources. The second tag area defines and maps the rules and constraints for logistics resource flow, such as resource allocation time windows, priority rules, and resource usage restrictions, providing logical constraints for resource allocation. By conducting a comprehensive analysis and evaluation of each logistics resource, its physical attributes and flow rules are determined. The physical attributes are mapped to the first tag area to form tags based on physical flow, such as the current location, destination, and estimated arrival time of transport vehicles. At the same time, the flow rules are mapped to the second tag area to form tags based on flow rules, such as specific allocation time periods for certain types of materials and priority sequences followed by certain resources. Finally, a complete virtual tag layer is constructed by cascading the first and second tag areas. This layer can comprehensively reflect the physical status and logical constraints of the logistics resource and is tightly integrated with the logistics management platform, enabling it to receive real-time updates of platform operation data.

[0039] The logistics management platform acquires resource allocation tasks, triggers the lightweight allocation module to perform task-oriented resource folding processing, acquires task-related tags through cascading virtual tag layers, performs resource allocation decisions and allocation compensation under random disturbances, determines resource allocation strategies, visualizes them on the platform interface, and executes terminal-based management.

[0040] Once the logistics management platform receives a resource allocation task, it immediately triggers the lightweight allocation module to start working. This module, task-oriented, performs resource folding processing on the logistics resource-computing network resource architecture. Specifically, based on the task's spatial location, timing requirements, and allocation constraints, it selects resource nodes highly relevant to the task from the existing logistics resource-computing network resource architecture and dynamically compresses (i.e., "folds") their resource status and network topology to form a minimum feasible resource set. After folding, the system obtains task-related tags through cascaded virtual tag layers. Based on task elements, the system extracts sets of tags strongly related to the task from the first tag area (physical tags) and the second tag area (rule tags), forming a "task-related tag vector" to guide allocation decisions. Subsequently, the platform executes resource allocation decisions and allocation compensation under random disturbances. The second decision node generates a preliminary allocation strategy based on the task-related tags, while the third compensation node introduces simulated disturbances, including resource anomalies, path interruptions, and priority conflicts, to evaluate strategy robustness and optimize anti-interference allocation, thereby determining the final resource allocation strategy. Finally, the system visualizes the resource allocation strategy on the platform interface, using a directed graph centered on the allocation task, along with dynamic resource flow diagrams, task-resource matching scores, and other interface modules for dispatchers to refer to and confirm. Simultaneously, the system automatically distributes the strategy to terminal execution nodes, completing a closed loop of resource instruction push, status monitoring, and feedback collection, thus achieving an intelligent, task-driven, and disturbance-resistant logistics resource allocation management process.

[0041] Furthermore, determining resource allocation strategies includes:

[0042] Upon receiving the resource allocation task, the first folding node interprets the task to determine the task elements; based on the task elements, it performs spatiotemporal folding processing on the logistics resource end-computing network resource architecture to determine a lightweight architecture; for the lightweight architecture, it triggers the second decision node to execute resource allocation decisions under label constraints to determine an initialization strategy; based on the task elements, it introduces random disturbances and triggers the third compensation layer to execute anti-interference compensation of the initialization strategy to determine the resource allocation strategy.

[0043] When the logistics management platform receives a resource allocation task, the lightweight allocation module initiates the first folding node to interpret the task, extracting key information such as allocation type, required resource type and quantity, target service area, time limit, and resource call priority. Based on this, the first folding node further performs dynamic spatiotemporal folding processing on the already constructed logistics resource end and computing network resource architecture according to these task elements. It filters resource paths and node sets highly relevant to the current task from the entire network, eliminating redundant calculations and allocation paths to generate a streamlined, fast-responding, and highly computable lightweight allocation architecture. Subsequently, the system automatically triggers the second decision node to execute a tag-constrained resource allocation decision operation. Specifically, the system accesses the first and second tag areas in the virtual tag layer, which are composed of resource mobility characteristics, and obtains task-related tags by combining them with task elements. These tags are then used as decision constraints to optimize the combination and matching of resource nodes and paths in the folded architecture, thereby generating an initial allocation strategy. To enhance the adaptability and robustness of the initial allocation strategy in actual deployment environments, the system further introduces simulated disturbance factors based on task context information, including but not limited to resource node failure, path blocking, allocation time deviation, or insertion of high-priority tasks. By activating the third compensation node, dynamic compensation processing is performed on the basis of the initial strategy to adjust the allocation path or replace resource nodes, and the reachability and allocation efficiency of the strategy are re-evaluated, thereby obtaining a set of final resource allocation strategies with anti-interference capabilities and optimal allocation performance.

[0044] Furthermore, the resource allocation task includes subjective allocation tasks and objective allocation tasks; wherein, the subjective allocation task is the allocation instruction input into the logistics management platform, and the objective allocation task is a dynamic allocation guide generated based on real-time operational data.

[0045] Resource allocation tasks can be categorized into subjective and objective allocation tasks based on their generation methods and execution logic. Subjective allocation tasks are explicit allocation instructions actively input into the logistics management platform by manual or pre-set system operators based on experience or scheduling intentions. These tasks typically have clear resource allocation targets, operation paths, or time requirements, and are based on human subjective intent. Objective allocation tasks, on the other hand, are driven by real-time operational data monitored by the platform. They are allocation requests or task guidance automatically generated by the system through embedded algorithms based on information such as the current operating status of logistics resources, demand change trends, and bottleneck node identification results. They exhibit a high degree of dynamism and automation. Objective allocation tasks can combine real-time indicators such as resource utilization, task congestion levels, and regional service load to generate allocation demands through intelligent analysis and prediction models. This drives the system to trigger a lightweight allocation module to perform dynamic resource reconfiguration and strategy optimization, enabling allocation behavior to adapt to real-time changes in the logistics system's operation and improving overall operational efficiency and service responsiveness.

[0046] Furthermore, the lightweight allocation module is connected to the virtual tag layer; the first tag area is updated based on real-time operational data; for the task elements, a type of tag based on the first tag area and a type of tag based on the second tag area are located as the task-related tags; resource allocation decision processing is performed with the task-related tags as constraints.

[0047] A data-driven dynamic connection is established between the lightweight allocation module and the virtual tag layer to achieve efficient mapping and real-time linkage between resource attributes and allocation strategies.

[0048] The virtual tag layer consists of two types of tag areas: a first tag area reflecting the physical resource mobility status and a second tag area describing the logical characteristics of resource allocation rules and policy constraints. The first tag area is continuously and dynamically updated based on real-time operational data from the logistics system, including real-time indicators such as current resource location, availability, operating load, and response latency, instantly reflecting the physical schedulability of resources. The second tag area stores the decision-making logic for resource allocation rules, priorities, usage strategies, and fault tolerance limits under specific conditions, used to constrain the legality of allocation behavior and the formation of optimal paths. After the lightweight allocation module receives task elements, it automatically invokes the virtual tag layer. Through matching calculations between task content and system status, it locates the tag type from the first tag area that best matches the current task's physical conditions and further combines it with the second tag type from the second tag area that satisfies the scheduling rules, forming a task-related tag with contextual integrity. Task association tags, as the core input of policy decision constraints, are introduced into the resource allocation reasoning process of the second decision node to guide key steps such as resource selection, path planning, and policy generation. This ensures that the resulting allocation scheme meets real-time schedulability requirements while also conforming to established logical rules and global optimal goals, thereby improving resource utilization efficiency and task execution reliability.

[0049] Furthermore, the resource allocation strategy is determined and visualized on the platform interface, including:

[0050] In the display interface of the logistics management platform, a allocation identifier is set; based on the allocation identifier, the resource allocation strategy is transformed into a directed block diagram to determine the resource allocation map; the resource allocation map is visualized on the platform interface.

[0051] In the display interface of the logistics management platform, a allocation identifier corresponding to the current resource allocation task is defined. This identifier serves as a unique index for the allocation instance, used to associate the generation, execution, and subsequent monitoring of allocation strategies. Based on this allocation identifier, the logistics management platform calls the generated resource allocation strategy content and transforms it into a structured directed graph representation. This graph presents the key nodes of the allocation process (such as resource providers, allocation paths, and receiving units) and their scheduling relationships and guiding logic in a graph-like manner, thus forming a clear resource allocation graph. The resource allocation graph not only reflects the flow path of resources in the computing network architecture but also intuitively displays the sequential dependencies and strategy rules between each step. The system visualizes the resource allocation graph in the platform display interface using a graphical user interface (GUI), enabling schedulers to intuitively understand resource distribution, allocation paths, and strategy logic.

[0052] Furthermore, after the terminal management is deployed, it includes:

[0053] As the strategy terminal issues and executes the strategy, it transmits back the execution operation data; updates the first tag area based on the execution operation data, verifies the operation rules based on the second tag area, and manages the feedback and allocation of logistics resources.

[0054] After the management is distributed to the execution terminals, the system enters a policy closed-loop feedback and adaptive optimization phase to achieve dynamic iteration and continuous optimization of the resource allocation process. Specifically, as the resource allocation strategy is distributed and actually executed at each logistics resource terminal, the system will simultaneously activate the execution monitoring mechanism to collect and transmit execution operation data from each terminal in real time. Operation data includes, but is not limited to, allocation task completion status, resource call response time, path execution delay, resource consumption, and abnormal information occurring during the allocation process. Based on the transmitted data, the original allocation process is dynamically analyzed, and the first tag area in the virtual tag layer is updated in real time accordingly, ensuring that the tag information accurately reflects the current physical state and flow changes of resources. Simultaneously, using the preset allocation rules, policy constraints, and performance standards in the second tag area, the actual execution is validated to assess the compliance and effectiveness of the current policy execution. When it is found that the strategy execution effect does not meet the expected goal, there is a deviation in resource allocation, or there is a significant change in the system status, the feedback allocation management process will be automatically triggered to reassess the task requirements and resource distribution. If necessary, the allocation strategy will be regenerated through the lightweight allocation module to achieve adaptive secondary allocation of logistics resources. This will build an intelligent closed-loop system of "task-strategy-execution-feedback-reallocation", which will significantly improve the flexibility, robustness and response efficiency of the logistics resource allocation system.

[0055] In summary, the embodiments of this application have at least the following technical effects:

[0056] For the logistics management platform, network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, where each logistics resource terminal corresponds to a logical network. A lightweight allocation module is built based on the dynamic spatiotemporal folding of the logistics resource terminal-computing network resource architecture, and supervised training is performed by injecting perturbation factors. This module is then embedded in the logistics management platform. A virtual tag layer is introduced to map resource tags to the logistics resource terminals. The logistics management platform receives resource allocation tasks, triggering the lightweight allocation module to perform task-oriented resource folding processing. Task-related tags are obtained through cascaded virtual tag layers, resource allocation decisions and allocation compensation under random perturbations are executed, and the resource allocation strategy is determined. This strategy is visualized on the platform interface and managed via terminal distribution. This approach solves the technical problems of low efficiency, slow response, and low resource utilization in existing logistics resource allocation technologies, achieving the technical effects of improving the intelligence level of resource allocation and realizing efficient resource utilization.

[0057] Example 2, based on the same inventive concept as the intelligent allocation method for logistics resources based on operational data analysis in the foregoing examples, such as... Figure 2 As shown, this application provides an intelligent logistics resource allocation system based on operational data analysis, wherein the system includes:

[0058] Slicing Unit 11: For the logistics management platform, network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, wherein each logistics resource end corresponds to a logical network; Training Unit 12: Based on the dynamic spatiotemporal folding of the logistics resource end-computing network resource architecture, a lightweight allocation module is built and perturbation factors are injected for supervised training, which is embedded in the logistics management platform; Tag Mapping Unit 13: A virtual tag layer is introduced to perform resource tag mapping on the logistics resource end; Management Unit 14: The logistics management platform obtains resource allocation tasks, triggers the lightweight allocation module to perform task-oriented resource folding processing, obtains task-related tags through the cascaded virtual tag layer, performs resource allocation decisions and allocation compensation under random perturbations, determines the resource allocation strategy, visualizes it on the platform interface, and executes terminal-based management.

[0059] Furthermore, the slicing processing unit 11 is used to perform the following method:

[0060] A first logistics resource terminal is determined, wherein the first logistics resource terminal is any one of the logistics resource terminals; for the first logistics resource terminal, an intrinsic allocation logic point is determined; the first logistics resource terminal and each associated logistics resource terminal are combined to determine associated allocation logic points, wherein the associated allocation logic points correspond to each associated logistics resource terminal; the intrinsic allocation logic points and the associated allocation logic points are integrated to perform network resource reconstruction and computing power configuration, which serves as the first logical network of the first logistics resource terminal.

[0061] Furthermore, the training unit 12 is used to perform the following methods:

[0062] Using the aforementioned computing network resource architecture as the task interpretation benchmark, a first folded node is deployed; using the selective retrieval of virtual tags as a decision constraint, a second decision node based on folded state decision is deployed; introducing task-oriented random perturbations, a third compensation node is deployed; the first folded node, the second decision node, and the third compensation node are cascaded, and the lightweight allocation module is built through supervised training until convergence.

[0063] Furthermore, the label mapping unit 13 is used to perform the following method:

[0064] The virtual tag layer is a feature layer for resource flow, and it includes a first tag area based on physical flow and a second tag area based on flow rules.

[0065] Furthermore, the management unit 14 is used to perform the following methods:

[0066] Upon receiving the resource allocation task, the first folding node interprets the task to determine the task elements; based on the task elements, it performs spatiotemporal folding processing on the logistics resource end-computing network resource architecture to determine a lightweight architecture; for the lightweight architecture, it triggers the second decision node to execute resource allocation decisions under label constraints to determine an initialization strategy; based on the task elements, it introduces random disturbances and triggers the third compensation layer to execute anti-interference compensation of the initialization strategy to determine the resource allocation strategy.

[0067] Furthermore, the management unit 14 is used to perform the following methods:

[0068] The resource allocation task includes subjective allocation tasks and objective allocation tasks; wherein, the subjective allocation task is the allocation instruction input into the logistics management platform, and the objective allocation task is a dynamic allocation guide generated based on real-time operational data.

[0069] Furthermore, the management unit 14 is used to perform the following methods:

[0070] The lightweight allocation module is connected to the virtual tag layer; the first tag area is updated based on real-time operational data; for the task elements, a type of tag based on the first tag area and a type of tag based on the second tag area are located as the task-related tags; resource allocation decision processing is performed with the task-related tags as constraints.

[0071] Furthermore, the management unit 14 is used to perform the following methods:

[0072] In the display interface of the logistics management platform, a allocation identifier is set; based on the allocation identifier, the resource allocation strategy is transformed into a directed block diagram to determine the resource allocation map; the resource allocation map is visualized on the platform interface.

[0073] Furthermore, the management unit 14 is used to perform the following methods:

[0074] As the strategy terminal issues and executes the strategy, it transmits back the execution operation data; updates the first tag area based on the execution operation data, verifies the operation rules based on the second tag area, and manages the feedback and allocation of logistics resources.

[0075] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0076] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0077] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for intelligent allocation of logistics resources based on operational data analysis, characterized in that the method... include: For the logistics management platform, network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, in which each logistics resource terminal corresponds to a logical network; Based on the dynamic spatiotemporal folding of the logistics resource end-computing network resource architecture, a lightweight allocation module is built and perturbation factors are injected for supervised training, which is then embedded in the logistics management platform. Specifically, this includes: using the computing network resource architecture as the task interpretation benchmark, deploying a first folding node; the system maps the currently pending logistics tasks to the computing network resource architecture; and constructing a matching relationship between tasks and the logical network by parsing task elements; the first folding node is used to compress and represent the spatially dispersed and temporally staggered resource states; deploying a second decision node using the selective retrieval of virtual tags as a decision constraint; in the second decision node, the system calls tag information with high task relevance from the virtual tag layer as decision boundary conditions in the allocation process; introducing random disturbances with task orientation, and deploying a third compensation node; the third compensation node is responsible for performing disturbance response analysis on the preliminary strategy output by the second decision node, and outputting an optimized allocation compensation strategy; cascading the first folding node, the second decision node, and the third compensation node, and building the lightweight allocation module through supervised training until convergence; A virtual tag layer is introduced to perform resource tag mapping on the logistics resource end; The logistics management platform acquires resource allocation tasks, triggers the lightweight allocation module to perform task-oriented resource folding processing, acquires task-related tags through cascading virtual tag layers, performs resource allocation decisions and allocation compensation under random disturbances, determines resource allocation strategies, visualizes them on the platform interface, and executes terminal-based management.

2. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 1, characterized in that, Network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, including: Determine a first logistics resource terminal, wherein the first logistics resource terminal is any one of the logistics resource terminals; For the first logistics resource end, determine the intrinsic allocation logic point; Combine the first logistics resource terminal with each associated logistics resource terminal to determine the associated allocation logic point, wherein the associated allocation logic point corresponds to each associated logistics resource terminal; Integrate the intrinsic allocation logic points and the associated allocation logic points to perform network resource reconstruction and computing power configuration, which serves as the first logical network of the first logistics resource end.

3. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 1, characterized in that, The virtual tag layer is a feature layer for resource flow, and it includes a first tag area based on physical flow and a second tag area based on flow rules.

4. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 3, characterized in that, Determine resource allocation strategies, including: Upon receiving the resource allocation task, the first folded node performs task interpretation to determine task elements. Based on the task elements, the logistics resource terminal-computing network resource architecture is subjected to spatiotemporal folding processing; Trigger the second decision node to execute resource allocation decisions under label constraints and determine the initialization strategy; Based on the task elements, random disturbances are introduced, triggering the third compensation node to execute the anti-interference compensation of the initialization strategy, and determining the resource allocation strategy.

5. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 4, characterized in that, The resource allocation task includes subjective allocation tasks and objective allocation tasks; The subjective allocation task is the allocation instruction input into the logistics management platform, while the objective allocation task is a dynamic allocation generated based on real-time operational data.

6. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 4, characterized in that, The lightweight allocation module is connected to the virtual tag layer; The first tag area is updated based on real-time operational data; For the task elements, a type of tag based on the first tag area and a type of tag based on the second tag area are identified as the task-related tags; Resource allocation decisions are made based on the task-related tags.

7. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 1, characterized in that, Determine resource allocation strategies and visualize them on the platform interface, including: Set the allocation identifier in the display interface of the logistics management platform; Based on the allocation identifier, the resource allocation strategy is transformed into a directed block diagram to determine the resource allocation map; The resource allocation map is visualized on the platform interface.

8. The intelligent allocation method for logistics resources based on operational data analysis as described in claim 3, characterized in that, After the terminal management is executed, it includes: As resource allocation strategies are distributed to terminals and executed, operational data is transmitted back. The first tag area is updated based on the execution operation data, the operation rules are verified based on the second tag area, and the feedback and allocation management of logistics resources are executed.

9. A logistics resource intelligent allocation system based on operational data analysis, characterized in that, The system is used to implement the intelligent allocation method for logistics resources based on operational data analysis as described in any one of claims 1-8, the system comprising: Slicing unit: For the logistics management platform, network slicing is performed on the computing network infrastructure to determine the computing network resource architecture, where each logistics resource corresponds to a logical network; Training Unit: Based on the dynamic spatiotemporal folding of the logistics resource end-computing network resource architecture, a lightweight allocation module is built and perturbation factors are injected for supervised training, which is then embedded in the logistics management platform. Tag mapping unit: Introduces a virtual tag layer to perform resource tag mapping on the logistics resource end; Management Unit: The logistics management platform acquires resource allocation tasks, triggers the lightweight allocation module to perform task-oriented resource folding processing, acquires task-related tags through cascading virtual tag layers, performs resource allocation decisions and allocation compensation under random disturbances, determines resource allocation strategies, visualizes them on the platform interface, and executes terminal-based management.

Images