Logistics distribution transfer node dynamic site selection method and device and storage medium

Abstract

The invention provides a logistics distribution transfer node dynamic site selection method and device and a storage medium, and the method comprises the steps: carrying out the multi-dimensional real-time collection and fusion processing of demand change characteristics and supply capacity redundancy in an urban agglomeration multi-level logistics network, and forming a comprehensive data set; performing cross-level collaborative analysis and potential node layout evaluation based on the data set to generate a scenarized deployment report; then, allocating the alternative dynamic site selection scheme in real time by utilizing conflict detection and adaptive calculation of an allocation sequence to obtain a layout list; and finally, performing iterative optimization on the selected node layout scheme through rolling type application and comprehensive verification. According to the method, flexible node site selection adjustment of the multi-level logistics network can be realized in the face of differentiated management and control and cross-platform transport capacity input change of different administrative regions, and the risk of resource allocation imbalance caused by seasonal demand fluctuation or temporary peak is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics site selection, and particularly to a method, device and storage medium for dynamically selecting transfer nodes for logistics distribution. Background Art

[0002] In the current environment where e-commerce and physical retail are integrated with each other, the logistics network within an urban agglomeration often presents a multi-level structure, including central warehouses, regional distribution centers, and end delivery stations facing terminal consumption scenarios. Due to the wide distribution areas of each node and multiple restrictive factors such as a large number of real-time order transfers, transportation capacity calls, and traffic control in the cross-platform collaboration mode, how to dynamically select transfer nodes in this complex environment has become an important concern for many enterprises. Existing site selection methods usually rely on static data or cost analysis from a single dimension, and it is difficult to cover the different policies of different administrative regions within the urban agglomeration and the real-time changes in the transportation capacity of third-party logistics platforms. At the same time, during peak order periods brought about by seasonal demand fluctuations or temporary major events, if there is no cross-level collaborative management mechanism between nodes, it may lead to unbalanced or lagged allocation of network resources. Under this background, relying solely on traditional static warehousing site selection models or cost-driven logistics planning schemes often fails to take into account multiple uncertain factors. Once a distribution center or end node bears an excessive load during market fluctuations, other nodes cannot share or supplement the transportation capacity in a timely manner, and ultimately it is difficult to meet the requirements for the flexibility of node layout and multiple constraint conditions in complex scenarios. Summary of the Invention

[0003] The main purpose of the present invention is to solve the technical problem that when dynamically selecting transfer nodes in a multi-level logistics network of an urban agglomeration, it is difficult to take into account multiple restrictive factors, resulting in lagged resource allocation and lack of flexibility in node layout; In the first aspect of the present invention, a method for dynamically selecting transfer nodes for logistics distribution is provided. The method for dynamically selecting transfer nodes for logistics distribution includes: Performing multi-dimensional real-time collection and fusion processing on the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set including demand change characteristic items and supply capacity redundancy; Performing cross-level collaborative analysis and potential node layout evaluation processing on the basis of the comprehensive data set to obtain a scenario-based deployment report including multi-dimensional indicators; Based on the scenario-based deployment report, performing real-time conflict detection and adaptive calculus processing of the allocation sequence on each alternative dynamic site selection plan to obtain a layout list including node layout suggestions and executable strategy combinations; According to the layout list, performing comprehensive verification and rolling application processing on the selected node layout to obtain a dynamically updated site selection plan and a collaborative strategy between nodes.

[0004] Optionally, in the first implementation manner of the first aspect of the present invention, the multi-dimensional real-time collection and fusion processing of the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set including demand change feature items and supply capacity redundancy includes: Real-time collection of order volume, inventory balance, distribution timeliness, local city control information, and dynamic information related to cross-platform cooperation of the central warehouse, regional distribution center, and terminal delivery station to obtain the original operation data; Using remote monitoring devices, blockchain traceability nodes, and API docking with the supplier system to perform multi-source integration on the original operation data to obtain multi-dimensional data streams; Vectorize and extract features from the multi-dimensional data streams to obtain preliminary demand change feature items and supply capacity redundancy; Mark the preliminary demand change feature items and supply capacity redundancy with scenario-specific factor tags to obtain feature data with advantage and limitation condition tags; Fuse the feature data with advantage and limitation condition tags in the time and space dimensions to obtain a comprehensive data set including demand change feature items and supply capacity redundancy.

[0005] Optionally, in the second implementation manner of the first aspect of the present invention, the cross-level collaborative analysis and potential node layout evaluation processing based on the comprehensive data set to obtain a scenario-based deployment report including multi-dimensional indicators includes: Based on the comprehensive data set, perform a preliminary screening of the resource allocation of each level node to obtain an initial resource allocation plan; According to the initial resource allocation plan, combined with the limitation condition tags of each level node, identify potential resource allocation conflicts to obtain a conflict warning list; Based on the conflict warning list, perform an evaluation of temporary replacement or addition of each level node to obtain an alternative adjustment strategy set; Perform an analysis of the upstream supply chain production capacity fluctuation characteristics on the alternative adjustment strategy set to obtain a resource allocation strategy with a flexibility score; According to the resource allocation strategy with a flexibility score, generate a cross-level resource allocation plan and perform a multi-dimensional comprehensive evaluation to obtain a scenario-based deployment report.

[0006] Optionally, in the third implementation manner of the first aspect of the present invention, the generation of a cross-level resource allocation plan and multi-dimensional comprehensive evaluation based on the resource allocation strategy with a flexibility score to obtain a scenario-based deployment report includes: Perform a hierarchical decomposition of the resource allocation strategy with a flexibility score to obtain a resource allocation sub-strategy set for each level node; Based on the resource allocation sub-strategy set, perform simulation analysis on the resource flow among nodes at each level to obtain a resource flow trend map, and calculate the resource utilization rate and load balancing degree of nodes at each level according to the resource flow trend map to obtain a node performance index matrix; Use a multi-objective optimization algorithm to perform weight allocation and comprehensive calculation on the node performance index matrix to obtain a cross-level resource allocation plan; Perform multi-dimensional sensitivity analysis on the cross-level resource allocation plan to evaluate its robustness in different scenarios and obtain a scenario-based allocation report.

[0007] Optionally, in the fourth implementation manner of the first aspect of the present invention, based on the scenario-based allocation report, perform real-time conflict detection and adaptive calculus processing of the allocation order on each alternative dynamic site selection plan to obtain a layout list including node layout suggestions and executable policy combinations, including: Extract the resource urgency rankings of nodes at each level from the scenario-based allocation report, and combine traffic flow and control period data to generate a preliminary cross-level resource sharing sequence; Use a collaborative game and scheduling engine to iteratively solve the contradictions and mutual exclusion conditions among nodes at each level in the cross-level resource sharing sequence to obtain a conflict adjustment plan; Based on the conflict adjustment plan, identify data synchronization delays or incomplete information phenomena in cross-platform cooperation and generate a redundant allocation plan; Perform multiple iterations and self-learning processing on the redundant allocation plan to obtain node layout suggestions, and configure executable policy combinations for each level node according to the node layout suggestions to obtain a layout list.

[0008] Optionally, in the fifth implementation manner of the first aspect of the present invention, the performing multiple iterations and self-learning processing on the redundant allocation plan to obtain node layout suggestions, and configuring executable policy combinations for each level node according to the node layout suggestions to obtain a layout list includes: Perform Monte Carlo simulation on the redundant allocation plan to generate a large number of random scenarios, evaluate the allocation effects under each scenario to obtain an allocation effect distribution map; Use a clustering algorithm to classify and extract common characteristics based on the allocation effect distribution map to obtain an optimized allocation mode set; Input the optimized allocation mode set into a reinforcement learning model, and through repeated training and policy optimization, obtain an adaptive node layout strategy; According to the adaptive node layout strategy, perform dynamic scoring and ranking on nodes at each level, select the optimal node combination to obtain node layout suggestions; Based on the node layout suggestions, combined with the historical performance and predicted requirements of nodes at each level, customize resource allocation and collaboration mechanisms for each node to obtain a layout list that includes a node layout plan and a hierarchical strategy combination.

[0009] Optionally, in the sixth implementation manner of the first aspect of the present invention, the comprehensive verification and rolling application processing of the selected node layout according to the layout list to obtain a dynamically updated site selection plan and inter-node collaboration strategy includes: Import the node layout suggestions and executable strategy combinations in the layout list into a visual sand table system, and conduct a comparison simulation with the actual logistics operation process to obtain simulation result data; Based on the simulation result data, identify nodes with insufficient resources in peak order scenarios and generate a temporary resource addition or scheduling plan; According to the cross-platform cooperation mode constraints and advantage information of nodes at each level, evaluate the matching degree of the temporary resource addition or scheduling plan to obtain an optimized resource supplement and personnel allocation mechanism; Compare and analyze the optimized resource supplement and personnel allocation mechanism with the real-time data feedback results to obtain node layout update suggestions; Based on the node layout update suggestions, combined with the potential change trends during the prediction period, dynamically adjust the node layout suggestions to obtain a dynamically updated site selection plan and inter-node collaboration strategy.

[0010] The second aspect of the present invention provides a dynamic site selection device for logistics distribution transfer nodes, and the dynamic site selection device for logistics distribution transfer nodes includes: A data integration module for multi-dimensional real-time collection and fusion processing of the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set including demand change feature items and supply capacity redundancy; A collaborative analysis module for cross-level collaborative analysis and potential node layout evaluation processing according to the comprehensive data set to obtain a scenario-based deployment report including multi-dimensional indicators; A conflict detection module for real-time conflict detection and adaptive calculation of the allocation order of each alternative dynamic site selection plan based on the scenario-based deployment report to obtain a layout list including node layout suggestions and executable strategy combinations; A rolling verification module for comprehensive verification and rolling application processing of the selected node layout according to the layout list to obtain a dynamically updated site selection plan and inter-node collaboration strategy.

[0011] The third aspect of the present invention provides a computer-readable storage medium storing instructions which, when run on a computer, cause the computer to execute the steps of the above-mentioned dynamic site selection method for logistics distribution transfer nodes.

[0012] The above-mentioned dynamic site selection method, device and storage medium for logistics distribution transfer nodes collect and fuse the demand change characteristics and supply capacity redundancy in the multi-level logistics network of the urban agglomeration in real time through multiple dimensions to form a comprehensive data set; then based on this data set, cross-level collaborative analysis and potential node layout evaluation are carried out to generate a scenario-based deployment report; subsequently, real-time deployment of the alternative dynamic site selection scheme is carried out by using conflict detection and adaptive calculation of the allocation sequence to obtain a layout list; finally, through rolling application and comprehensive verification, the selected node layout scheme is iteratively optimized. The present invention can realize flexible node site selection adjustment for the multi-level logistics network in the face of different administrative region's differential control and cross-platform transport capacity input changes, and reduce the risk of resource allocation imbalance caused by seasonal demand fluctuations or temporary peaks.

[0013] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification, or be understood by implementing the present invention. The objectives and other advantages of the present invention are realized and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0014] To make the above objectives, features and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and described in detail as follows. Description of the Drawings

[0015] Figure 1 It is a schematic diagram of an embodiment of the dynamic site selection method for logistics distribution transfer nodes in an embodiment of the present invention; Figure 2 It is a schematic diagram of an embodiment of the dynamic site selection device for logistics distribution transfer nodes in an embodiment of the present invention. Detailed Embodiments

[0016] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] As used in the embodiments of the present invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include other unlisted steps or units, or may optionally further include other steps or units inherent to these processes, methods, products, or devices.

[0018] For ease of understanding of this embodiment, first, a dynamic site selection method for logistics distribution transfer nodes disclosed in the embodiments of the present invention will be introduced in detail. As Figure 1 shown, this method includes the following steps: 101. Multidimensionally and real-time collect and fuse the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set containing demand change characteristic items and supply capacity redundancy; In an embodiment of the present invention, the multidimensional real-time collection and fusion processing of the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set containing demand change characteristic items and supply capacity redundancy includes: real-time collecting the order volume, inventory balance, distribution timeliness, local city control information, and dynamic information related to cross-platform cooperation of the central warehouse, regional distribution center, and end delivery stations to obtain the original operation data; using remote monitoring devices, blockchain traceability nodes, and API docking with the supplier system to perform multi-source integration on the original operation data to obtain a multi-dimensional data stream; performing vectorization and feature extraction on the multi-dimensional data stream to obtain preliminary demand change characteristic items and supply capacity redundancy; performing scenario-specific factor marking on the preliminary demand change characteristic items and supply capacity redundancy to obtain feature data with advantage and limitation condition labels; and performing spatio-temporal dimension fusion on the feature data with advantage and limitation condition labels to obtain a comprehensive data set containing demand change characteristic items and supply capacity redundancy.

[0019] Specifically, real-time collection of order volume, inventory balance, delivery timeliness, local city control information, and dynamic information related to cross-platform cooperation for the central warehouse, regional distribution centers, and end-delivery stations is carried out. The specific process includes deploying embedded collection devices that can communicate with the cloud at each node and sending operation parameters to the management platform through fixed network or wireless network channels. Order volume data can be pushed in real time by the merchant ERP interface or the front-end order placement system. Inventory balance information can be provided by the warehouse management system or RFID scan records. Delivery timeliness can be extracted through vehicle trajectory positioning or the timing records of the dispatching platform. For local city control information, the management platform can access the government traffic information release port and synchronize data according to the specified update cycle. Cross-platform cooperation involves capacity sharing, vehicle allocation, and delivery time slot agreements. It is necessary to define the data format and transmission frequency in the cross-platform information exchange protocol and write the relevant updates into the shared database immediately. This process will set up automatic reconnection and caching strategies based on the data flow status between nodes to ensure the integrity of the collection process. By deploying environmental perception devices at the distribution center or end-delivery station, the on-site congestion level and site utilization can also be captured, so as to uniformly summarize the original data in each dimension and form a set of original operation data that can reflect the operation status of multi-level nodes; Using remote monitoring devices, blockchain traceability nodes, and API docking with the supplier system, multi-source integration of the original operation data is carried out. It is necessary to set up a conversion engine in the data access layer that can be compatible with multiple transmission protocols. For the information uploaded by remote monitoring devices, sensor parameters are received through a predefined encrypted channel and stored in the preset field format after parsing. For the transaction records generated by blockchain traceability nodes, smart contract calls need to be made before integration to obtain the logistics traceability code and timestamp corresponding to each record, and associate these traceability information with the order number. For the API connection of the supplier system, a specific authentication process needs to be followed. After the key management module completes the identity verification, parameters such as inventory balance and delivery progress are pulled to the integration platform. Data duplication and dirty data during the integration process can be corrected through field mapping and verification mechanisms. Duplicate records can be filtered by the duplicate detection algorithm, and dirty data can be removed or corrected using the outlier identification algorithm. After these processes are completed, multi-dimensional data streams will store the order, delivery, inventory, transportation, and traceability information from different sources in chunks according to time series and node distribution, providing multi-source and complete input for the subsequent feature extraction link; When vectorizing and extracting features from the multi-dimensional data stream, the data items can be segmented according to the same or comparable time scales in the preprocessing stage. If the collection intervals of some data sources are inconsistent, alignment can be completed through interpolation or zero-padding. After the data structures are unified, normalization processing and feature point marking are performed. The characteristic items of demand changes can be weighted and calculated by indicators such as the increase or decrease rate of the order volume within several consecutive time slices, and the overlapping degree between the delivery timeliness and the urban control period. The supply capacity redundancy can be numerically represented by the comparison relationship between the inventory surplus and the historical shipment volume, and in the case of having blockchain traceability records, the actual supply capacity can be determined according to the matching degree between the traceability code and the delivery process. In the feature extraction stage, a vectorization scheme is adopted to integrate the above indicators into a two-dimensional or three-dimensional numerical vector for direct comparison or aggregation in subsequent links. Through a unified vector representation method, the demand fluctuations and redundancy spaces of each node at different time periods can obtain quantization results, and these results are the preliminary characteristic items of demand changes and the supply capacity redundancy. When marking the scenario-specific factors for the preliminary characteristic items of demand changes and the supply capacity redundancy, it is necessary to introduce differentiated information such as the unique geographical location, administrative control policies, business characteristics of cooperation partners, and seasonal peaks of nodes within the urban agglomeration. First, map the traffic control scope or restricted driving period of the node's location to the corresponding time slice of the characteristic item. If the control measures overlap with the order volume increase period, set a control risk label for this node. Secondly, integrate elements such as the depth of cross-platform cooperation or special business requirements into the characteristic indicators. If there is a capacity sharing agreement between the node and a third-party logistics platform, mark it with a capacity cooperation advantage label. For seasonal order outbreaks or festival activities, a demand peak label can be added to the characteristic data, and it is linked with the supply capacity redundancy to judge whether there are additional inventory or spare vehicle resources once a peak appears, thereby assigning a spare resource advantage label or a resource shortage label to this node. Through the above methods, characteristic data with advantage and restriction condition labels is obtained to distinguish the attribute differences of nodes in the same network. When performing spatio-temporal dimension fusion on the characteristic data with advantage and restriction condition labels, first, taking the time dimension as the main line, horizontally merge the characteristic labels of each node at the same time scale, and then perform spatial clustering according to the node geographical location or administrative control area, dividing the nodes with similar label combinations or within the same risk range into the same group. After clustering, perform weighted or aggregation operations on the characteristic items of demand changes and the supply capacity redundancy of these groups. If a group simultaneously appears with a high order peak label and an inventory tension label, set a higher priority level in the fusion stage. Finally, store the labels and corresponding vector data of each node and group into a unified database together to form a comprehensive data set containing the characteristic items of demand changes and the supply capacity redundancy. This data set presents the operation characteristics and resource capabilities of multi-level nodes with spatio-temporal coordinates and label information.

[0020] 102. Conduct cross-level collaborative analysis and potential node layout evaluation based on the comprehensive data set to obtain a scenario-based deployment report containing multi-dimensional indicators; In one embodiment of the present invention, the cross-level collaborative analysis and potential node layout evaluation processing based on the comprehensive data set to obtain a scenario-based allocation report containing multi-dimensional indicators includes: preliminary screening of resource allocation of nodes at each level based on the comprehensive data set to obtain an initial resource allocation plan; based on the initial resource allocation plan, combined with the restriction condition labels of nodes at each level, identifying potential resource allocation conflicts to obtain a conflict warning list; based on the conflict warning list, temporarily replacing or adding nodes at each level to obtain a set of alternative adjustment strategies; analyzing the upstream supply chain capacity fluctuation characteristics of the alternative adjustment strategy set to obtain a resource allocation strategy with a flexibility score; based on the resource allocation strategy with a flexibility score, generating a cross-level resource allocation plan and performing a multi-dimensional comprehensive evaluation to obtain a scenario-based allocation report.

[0021] Specifically, based on the comprehensive data set, the resource allocation of nodes at all levels is preliminarily screened. When the initial resource allocation plan is obtained, the indicators such as order volume, inventory balance, transportation capacity distribution and urban control information are classified and summarized, and the available resources between nodes are quantitatively compared in combination with the geographical location of the nodes and the cross-platform cooperation terms; after weighted calculation of the order fluctuation curve within the time period, the demand peak is matched with the node inventory redundancy to form the initial resource matching structure, and the real-time data of all distribution centers and terminal delivery stations are integrated, and the transportation distance and traffic restriction period are calculated in combination with the geographical coordinates. The relationship with the dispatching mode; the initial resource allocation plan includes the number of vehicles available at each node in a specific period of time, available storage space and terminal coverage capacity, and includes the trunk transportation capacity from the central warehouse to the regional distribution center and the branch transportation capacity from the regional distribution center to the terminal delivery station; during the evaluation process, the historical load records of each node will be retrieved to identify the load limit during peak hours, so as to mark the potential load risk points in the allocation plan and maintain consistent measurement standards when investing in cross-platform transportation capacity; the plan gives specific numerical results in terms of allocation ratios between nodes, allocation timeliness and external resource calls; Specifically, when identifying potential resource allocation conflicts and obtaining a conflict warning list in combination with the node limit condition tags at each level according to the initial resource allocation plan, it is necessary to compare the transport capacity usage, available vehicle types, and geographical control constraints of each node within different time windows; if there is a node whose transport capacity demand exceeds the permitted range under the restriction policy during the urban peak period, it is marked as a time period conflict; if multiple nodes request the same third-party vehicle type during the same time period and all have high-priority demands, it is recorded as a resource preemption conflict; in terms of inventory surplus and order volume, by checking the matching of the weighted load index and the safety inventory threshold, it is checked whether there is a situation of over-limit resource allocation; if the over-limit allocation of one node causes a similar gap in other nodes, it is regarded as a resource sharing conflict; after the system evaluates the above situations one by one, a conflict warning list is generated, which lists the specific conflict types, corresponding nodes, and the reasons for the conflicts, and attaches the time series and external association information of the conflicts to the database; Specifically, when conducting a temporary replacement or addition assessment of nodes at each level based on the conflict warning list and obtaining an alternative adjustment strategy set, the substitution relationship between nodes will be comprehensively considered from dimensions such as urban control restrictions, inventory shortage levels, and cross-platform cooperation agreements; if the conflict warning indicates that a node cannot complete the distribution task within the scheduled time period, idle vehicles or remaining storage space will be searched for among adjacent nodes to construct replacement candidates; in scenarios where third-party vehicle conflicts are relatively serious, the feasibility of adding small front warehouses or enabling local distribution teams will be evaluated; when comparing the geographical locations, operating costs, and service coverage of multiple nodes, a hierarchical scoring model will be used to rank the nodes that meet the conflict resolution requirements in terms of priority; if multiple nodes can undertake the same replacement function, the order timeliness requirements and cross-platform priorities will be compared by weights to determine the most suitable allocation path; finally, an alternative adjustment strategy set is formed, including plans such as replacing nodes, increasing the storage scale, or adding transport capacity channels, and the adaptation conditions and implementation difficulties are noted in each plan; Specifically, when analyzing the characteristics of upstream supply chain capacity fluctuations for the alternative adjustment strategy set to obtain a resource allocation strategy with a flexibility score, the historical delivery cycle, production plan, and inventory turnover speed of the main suppliers will be extracted from the comprehensive dataset, and the supply curve will be examined by comparing the peak demand of the mapped node to check whether the supply end will be interrupted in case of peak demand; if the alternative strategy includes a plan to add warehouses or replace nodes in the region, it is necessary to check whether the relevant suppliers have additional production capacity and transportation capacity; if the historical delivery record shows that a certain supplier has long-term fluctuations, a lower flexibility score will be assigned to the corresponding strategy to reflect the supply risk; when quantifying the fluctuation characteristics, the discrete production capacity data will be superimposed on the existing order allocation curve through function fitting. If the superimposed result shows a high degree of coincidence, indicating that the supply capacity of the supplier fluctuates greatly during critical periods, the corresponding weight will be reduced in the flexibility score, and a supply uncertainty label will be added; if the fluctuation amplitude is relatively stable, the score of this strategy will be increased, thus forming a resource allocation strategy with a flexibility score. Specifically, when generating a cross-level resource allocation plan according to the resource allocation strategy with a flexibility score, it is necessary to synchronously load the priority order of the nodes, transportation routes, and supply chain capacity restrictions into the allocation process; when implementing the strategies with higher scores in the ranking, the allocation period will be split into several windows, and vehicles and warehouse locations will be allocated according to supply stability and urban control conditions in each window; if it is found that high-risk nodes and unstable suppliers overlap in the same period, the allocation order in this period will be adjusted backward, and third-party vehicles or cold chain equipment will be appropriately reserved; when connecting trunk transportation and last-mile delivery, the final route combination suitable for urban control conditions will be selected by parallel comparison of the coverage and cost consumption of each allocation plan, and the transportation distance, timeliness, and alternative plans will be recorded in the system; when there are multiple transportation paths with similar scores, they will be further screened according to the loading efficiency and scalability, and the selected paths will be stored with labels; by comparing with the previous conflict warning list and alternative adjustment strategy set, the node combinations that have not been lifted or are still in a high-risk state will be excluded, and finally, a cross-level resource allocation plan including node division of labor, vehicle allocation, and inventory allocation arrangements will be generated, and the corresponding elements will be presented in a structured document.

[0022] Further, according to the resource allocation strategy with flexibility scores, a cross - level resource allocation plan is generated and comprehensively evaluated from multiple dimensions to obtain a scenario - based allocation report, including: decomposing the resource allocation strategy with flexibility scores hierarchically to obtain a set of resource allocation sub - strategies for each level node; based on the set of resource allocation sub - strategies, simulating and analyzing the resource flow between each level node to obtain a resource flow trend map, and calculating the resource utilization rate and load balance degree of each level node according to the resource flow trend map to obtain a node performance index matrix; using a multi - objective optimization algorithm to perform weight allocation and comprehensive calculation on the node performance index matrix to obtain a cross - level resource allocation plan; performing multi - dimensional sensitivity analysis on the cross - level resource allocation plan to evaluate its robustness under different scenarios to obtain a scenario - based allocation report.

[0023] Specifically, when decomposing the resource allocation strategy with flexibility scores hierarchically to obtain a set of resource allocation sub - strategies for each level node, it is necessary to gradually break down and map the global resource allocation idea to nodes of different natures. First, screen applicable allocation terms according to the levels of nodes in the urban agglomeration logistics network (such as central warehouses, regional distribution centers, and terminal delivery stations), and identify during which periods external transportation capacity can be called or within what range transfer storage areas can be set by retrieving cross - platform cooperation terms and administrative control information collected in the early stage. Subsequently, give priority to allocating strategies with higher flexibility scores to nodes that have stable distribution capabilities and appropriate redundant warehousing in the past period of time to reduce the probability of load overload; if a node has traffic restrictions or environmental constraints during peak hours, select sub - strategies applicable to night or off - peak allocation from the overall strategy. The significance of this hierarchical decomposition process is that each level node can obtain its exclusive set of allocation instructions, including the number of transportation capacities that can be called, the specific configuration of temporarily adding cold - chain vehicles or manpower, etc., and store them in a unified data structure as a set of resource allocation sub - strategies; Specifically, when, based on the set of resource allocation sub - strategies, simulating and analyzing the resource flow between each level node to obtain a resource flow trend map, and calculating the resource utilization rate and load balance degree of each level node according to the resource flow trend map to obtain a node performance index matrix, it is necessary to incorporate the resource transfer relationship between nodes into a dynamic model for hourly simulation. To describe the variation law of resources in the time - space dimension, the following flow - balance constraints are introduced: ; Among them, represents the quantity of resources flowing from node to node at time , is the quantity of resources of node at time The net demand or net supply value. Through iterative calculation, the input and output resources of each node in a given period can be used to obtain a real-time resource flow diagram. The system uses a directional topological structure to represent these flows at the visual level, and superimposes restrictions on vehicle flow or inventory transfer during control periods. Next, when calculating the resource utilization of nodes at each level, the indicator , the node The ratio of the actual amount of resources carried to its upper limit is measured; the load balancing degree can be measured based on the coefficient of deviation or the coefficient of variation of the traffic distribution of each node. Finally, the relevant parameters are summarized to obtain the node performance indicator matrix, in which each row corresponds to the multi-dimensional indicator information of a node, including resource utilization, peak period occupancy time and other labeled attributes; using the multi-objective optimization algorithm, the node performance indicator matrix is ​​weighted and comprehensively calculated to obtain a cross-level resource allocation plan. It is necessary to find the global optimal solution based on multiple objectives such as distribution efficiency, risk control and overall scheduling cost. A multi-objective function with a weighted summation form can be defined: ; in, Indicated in The values ​​of various indicators under the alternative strategies (such as transportation costs, delivery time, inventory overrun risk, etc.), is the target weight, It is the applicable coefficient for different strategies or node combinations. Through genetic algorithms, particle swarm algorithms or other hybrid iterative algorithms, search and compare the resource scheduling of all nodes in different time periods, and set hard or soft constraints for environmental factors such as control periods, seasonal order peaks, and supplier capacity fluctuations. Finally, it converges to a set of Pareto optimal solutions, and selects the solutions with outstanding comprehensive performance in terms of transportation time, capacity allocation, and cost expenditure. This solution defines the vehicle allocation, sub-warehouse activation timing, and inventory turnover methods between cross-level nodes in the form of multiple time windows for subsequent analysis and execution reference; Specifically, when performing multi-dimensional sensitivity analysis on the cross-level resource allocation plan, evaluating its robustness in different scenarios, and obtaining a scenario-based allocation report, it is necessary to change the core parameters or external environment definitions to test the adaptation range of the allocation plan. Multiple groups of parameters can be selected in dimensions such as order fluctuations, capacity reduction ratios, and supply chain production capacity limitations to establish simulation scenarios, and the aforementioned resource flow simulation and multi-objective optimization processes are run in each scenario to observe the variation ranges of key performance indicators such as average transportation distance, inventory occupancy, and vehicle utilization rate in different scenarios. If the scheduling process can maintain a relatively stable indicator distribution under multiple environmental assumptions, it indicates that the plan has good robustness; if there is an obvious imbalance in delivery timeliness or inventory utilization rate in a certain scenario, record its triggering conditions and alternative strategies that can be corrected in the scenario-based allocation report. The report will explain the capacity utilization and resource allocation of each node from different dimensions (such as time periods, geographical spaces, and cross-platform cooperation levels), list bottleneck nodes and response plans, and present them synchronously in documents and visualizations to facilitate more flexible scheduling decisions in multi-level logistics networks.

[0024] 103. Based on the scenario-based allocation report, perform real-time conflict detection and adaptive calculus processing of the allocation order for each alternative dynamic site selection plan to obtain a layout list including node layout suggestions and executable strategy combinations; In an embodiment of the present invention, the performing real-time conflict detection and adaptive calculus processing of the allocation order for each alternative dynamic site selection plan based on the scenario-based allocation report to obtain a layout list including node layout suggestions and executable strategy combinations includes: extracting the resource urgency rankings of nodes at each level from the scenario-based allocation report, combining traffic flow and control period data to generate a preliminary cross-level resource sharing sequence; using a cooperative game and scheduling engine to iteratively solve the contradictions and mutual exclusion conditions between nodes at each level in the cross-level resource sharing sequence to obtain a conflict adjustment plan; based on the conflict adjustment plan, identifying data synchronization delays or incomplete information phenomena in cross-platform cooperation to generate a redundant allocation plan; performing multiple iterations and self-learning processing on the redundant allocation plan to obtain node layout suggestions, and configuring executable strategy combinations for each level node according to the node layout suggestions to obtain a layout list.

[0025] Specifically, when extracting the resource urgency rankings of nodes at each level from the scenario-based deployment report and generating a preliminary cross-level resource sharing sequence in combination with traffic flow and control period data, it is necessary to first screen out indicators such as the peak demand for node capacity, inventory redundancy, and distribution capacity utilization rate in the report, and classify their urgency according to the level to which the node belongs (such as central warehouse, regional distribution center, or end delivery station). The measurement of the urgency ranking can be completed by weighting multiple factors, including the order surge rate, historical carrying limit, and current inventory turnover rate, etc., and nodes that are restricted by night-time traffic or have saturated traffic flow during specific periods are marked as high-priority processing objects according to the administrative control period. The process of generating the cross-level resource sharing sequence requires retrieving the surplus or gap of the capacity of each node during the feasible period and combining it with the urgency ranking. If a regional distribution center has surplus vehicle scheduling resources during the peak period and there are no strict restrictions on the traffic in the area where it is located, it will be listed as a priority candidate to assist the surrounding end delivery stations; if the central warehouse encounters a production capacity release period and the control period is staggered from that of the downstream nodes, more shipping channels will be allocated to the central warehouse in the preliminary resource sharing sequence. Through the above method, a preliminary cross-level resource sharing path can be set without changing the overall delivery order, and the flow possibilities of the resources of each level of nodes at different times can be integrated into the same sequence for adaptive calculation and conflict detection in subsequent steps; When using the cooperative game and scheduling engine to iteratively solve the contradictions and mutual exclusion conditions between nodes at each level in the cross-level resource sharing sequence to obtain a conflict adjustment plan, it is necessary to define the objective function and constraint set of the nodes in the engine, and input the resource requirements or capacity allocation intentions of different levels of nodes at the same time into the cooperative game model. If a regional distribution center has a strong dependence on a certain type of vehicle, and another end delivery station also requests the same type of vehicle at the same time, it is regarded as an exclusive condition for resource allocation. The engine will evaluate the difference in benefits and losses of different decisions made by each node through multiple rounds of iterative games based on the urgency ranking, the available number of vehicles, and the terms of priority scheduling in the cross-platform cooperation agreement, and check whether constraints such as traffic control or inventory safety thresholds are violated during this process. If it is detected that a node exceeds the limit of using vehicles during a certain period or attempts to make a delivery during the control period, the scheduling engine will trigger a punitive correction, move the scheduling order of the node backward or allocate alternative capacity, so as to reduce the risk of occupying other nodes. After multiple rounds of iteration, a conflict adjustment plan is generated. This plan gives corresponding adjustment measures for different types of conflicts at different levels and retains the calculation results of each round of game inside the engine for subsequent inspection; Based on the conflict resolution plan, when identifying data synchronization delays or incomplete information phenomena in cross-platform cooperation and generating a redundant deployment plan, it is necessary to find the occurrence time periods of missing data or time-delay data between the original shared sequence and the corrected conflict resolution plan. Cross-platform cooperation often involves independent supply chain management systems and vehicle scheduling systems, and some data may lag or be missing information when switching platforms or during cross-regional scheduling. After the scheduling engine reads the conflict resolution plan, if it finds that the actual execution time of a vehicle request at a certain node does not match the record of the third-party platform, it marks it as a data synchronization delay. If incomplete information causes the system to be unable to determine the specific capacity matching situation of a certain node within a certain time period, a certain safety margin is reserved for this node in the redundant deployment plan or a temporary cooperation plan is added to ensure that necessary operations can still be maintained after normal data recovery. The generation process of the redundant deployment plan will divide alternative plans for each node according to the resource urgency and time period overlap degree. For example, additional vehicles of other types are dispatched or warehouse personnel work overtime at a terminal delivery station during the night when cross-platform cooperation is not smooth. Such redundant arrangements will bring additional resource consumption in the overall layout, but can avoid capacity gaps during sensitive time periods and reallocate idle resources during subsequent data recovery; Perform multiple iterations and self-learning processes on the redundant deployment plan to obtain node layout suggestions. When obtaining a layout list by configuring an executable policy combination for each hierarchical node according to the node layout suggestions, it is necessary to retain all historical versions of the redundant deployment records in the scheduling engine and check their execution results one by one to find out whether there are problems such as resource waste or excessive redundancy. Through the self-learning algorithm, inject the feedback data of each node in actual operation into the engine to update its urgency ranking and frequently conflicted time periods, thereby fine-tuning the cross-hierarchical resource sharing sequence in the next iteration, removing redundant redundant allocations and retaining effective coping strategies. If a certain node always shows an urgent need for night vehicles in multiple iterations, more night-time capacity is allocated in the latest version of the layout suggestion. When the deployment strategies of all nodes are optimized, the scheduling engine outputs node layout suggestions, and summarizes the executable strategies of each node for different time periods, different capacity types, or inventory replenishment methods in a layout list. This list contains elements such as the node layout location or the selection of transfer stations, presented together with the corresponding policy combinations, providing clear configuration instructions for implementing the plan in the subsequent actual operation link.

[0026] Further, perform multiple iterations and self-learning processing on the redundant deployment plan to obtain a node layout suggestion, and configure an executable policy combination for each hierarchical node according to the node layout suggestion to obtain a layout list, including: performing a Monte Carlo simulation on the redundant deployment plan to generate a large number of random scenarios, evaluating the deployment effect in each scenario to obtain a deployment effect distribution map; using a clustering algorithm to classify and extract common features based on the deployment effect distribution map to obtain an optimized deployment mode set; inputting the optimized deployment mode set into a reinforcement learning model, and obtaining an adaptive node layout strategy through repeated training and policy optimization; according to the adaptive node layout strategy, dynamically score and rank each hierarchical node, select the optimal node combination to obtain a node layout suggestion; based on the node layout suggestion, combined with the historical performance and predicted demand of each hierarchical node, customize a resource allocation and cooperation mechanism for each node to obtain a layout list including a node layout plan and a hierarchical policy combination.

[0027] Specifically, when performing a Monte Carlo simulation on the redundant deployment plan to generate a large number of random scenarios and evaluating the deployment effect in each scenario to obtain a deployment effect distribution map, it is necessary to first set corresponding probability distributions for various uncertainty factors within the system, including the occurrence period of the order volume peak, the proportion of capacity reduction, traffic control delay and other factors, and then generate multiple groups of input parameters within a given range by means of independent sampling or joint sampling. Each group of input parameters represents a complete random scenario. After the system reads this scenario, it will load the established resource call and scheduling rules in the redundant deployment plan, send the scenario data and node attribute information to the simulation engine for hourly calculation, and count indicators such as node inventory occupancy rate, vehicle availability rate and delivery completion rate within each hour, and use the redundant paths or additional warehousing strategies within the plan in the event of urban control or cross-platform capacity delay. After each scenario simulation ends, the overall deployment cost and order fulfillment rate will be recorded. If resource conflicts or overloading occur in some scenarios, the severity of this situation will be marked in the result table. To facilitate identifying the global characteristics of the scheduling strategy in different environments, the system summarizes the results of all scenarios and draws a deployment effect distribution map, where the horizontal axis and the vertical axis can respectively represent capacity utilization and order fulfillment. If there are highly concentrated points in some areas, it means that there are similar performances in most scenarios. Through large-scale sampling and repeated simulation, the deployment effect distribution map can show the adaptation range of the redundant deployment plan under multiple uncertainties and provide quantifiable evaluation data for subsequent mode extraction and policy optimization; When using a clustering algorithm to classify and extract common features based on the distribution map of deployment effects to obtain an optimized deployment mode set, it is necessary to aggregate the scenario simulation results in a multi-dimensional index space. The distribution map of deployment effects already contains numerical values such as the utilization rate of transportation capacity, delivery timeliness, inventory redundancy utilization rate, and vehicle type conflict rate corresponding to a large number of scenarios. Each scenario can be mapped to a high-dimensional vector , where represents the dimension of the index. To more accurately measure the difference degree between scenarios, an adaptive weighted distance function can be defined ; where is the index weight that can be dynamically adjusted. If the attention to delivery timeliness is relatively high in the current evaluation, the corresponding is taken as a larger value. After calculating the distance matrix between all scenarios, the improved hierarchical clustering or spectral clustering method can be used to divide the scenarios with similar distances into the same class, and extract the resource allocation and node scheduling features representative of these scenarios at the class center point. Due to the differences in hierarchical node layout and cross-platform cooperation, some classes may only correspond to specific traffic control periods or capacity sharing agreements, so similar redundant deployment effects will appear within the class. After extracting the common features of each class, it is necessary to construct an optimized deployment mode set according to these common marks. For example, use a higher frequency of less-than-truckload vehicle scheduling strategy in areas with strict night traffic restrictions, or reserve more safety inventory thresholds for the central warehouse when the supply is unstable. The final optimized deployment mode set is an abstract summary based on the commonality and differences of scenarios, providing a diverse and feasible solution space for the subsequent reinforcement learning model. When inputting the optimized deployment mode set into the reinforcement learning model and obtaining an adaptive node layout strategy through repeated training and policy optimization, it is necessary to define state, action, and reward functions in the reinforcement learning environment, and use the resource deployment method extracted from the mode set as the initial policy candidate. First, set a group of observable state variables for the system, such as the remaining inventory of nodes, vehicle availability, remaining time of traffic control, etc. The action set corresponds to different resource call and node swapping decisions. An iterative mechanism for Q-value update or policy gradient update can be constructed to make the reinforcement learning agent select different deployment actions in each round and obtain the corresponding return value. If a policy gradient-based model is used, the policy can be regarded as a differentiable function with parameters . After multiple rounds of training, maximize the expected return , that is ; where It includes indicators such as delivery completion rate, penalty for illegal allocation, and total cost. During the training process, the system updates the gradient according to the performance feedback of each strategy in random scenarios, and uses the optimized allocation mode set as prior knowledge, which helps to accelerate the strategy convergence speed. After multiple rounds of iteration, the model can learn the optimal layout and scheduling methods when encountering peak impacts or incomplete information at different hierarchical nodes, and finally output an adaptive node layout strategy, and record the specific strategy options available for deployment internally to cope with the changing external environment; When dynamically scoring and ranking each hierarchical node according to the adaptive node layout strategy and selecting the optimal node combination to obtain the node layout suggestion, it is necessary to interpret the foregoing reinforcement learning results from multiple dimensions. The system generates a node-level scoring matrix in each policy evaluation session to measure the transport capacity adaptability, cooperation efficiency, and remaining redundancy capacity of the node at various time periods, and compares these scores with the urban control time window, the depth of cross-platform cooperation, or the supply-side production capacity fluctuation. If a node always maintains high efficiency and stable resource flexibility during frequent allocation, its dynamic score is relatively high. By cross-comparing the rankings of multiple nodes at the same time period, the system screens out the node combination that can best meet the current or future period's delivery needs among all levels, and regards these combinations as the optimal layout suggestions. If some nodes perform well in local indicators but have too much restraint or a high conflict rate in cross-level cooperation, their overall ranking will be reduced. The final generated node layout suggestions will indicate the score weights and applicable time period boundaries of each node, laying a clear execution idea for the subsequent issuance of resource allocation and cooperation mechanisms; Based on the node layout suggestions, combined with the historical performance and predicted demand of each hierarchical node, when customizing the resource allocation and cooperation mechanism for each node to obtain a layout list including the node layout plan and hierarchical strategy combination, it is necessary to build a dynamic configuration rule library within the management platform. By retrospectively analyzing the historical operation data of the nodes, identify their common bottlenecks and adaptability intervals. If a regional distribution center shows a stable night loading level in past simulations or real operations, the night transport capacity tilt ratio for this node can be increased according to the node layout suggestions, and corresponding load limit control can be carried out during the daytime traffic congestion period to ensure that the overall traffic remains balanced during peak hours. For the terminal delivery station, if the predicted demand curve shows that it will face a large influx of orders in the next few days, increase the pre-reserved storage resources or staffing in the layout list, and set increasing or decreasing quotas according to the production capacity fluctuations of the upstream supply chain. The final layout list will be presented in the form of a table or structured data, giving executable strategy combinations for the central warehouse, regional distribution centers, and terminal delivery stations respectively, including scheduling priorities, redundant vehicle call frequencies, and data docking time periods with cooperation platforms. The node layout plan gives operation instructions from the two aspects of overall spatial distribution and time period allocation, ensuring a highly coordinated logistics network operation mode among different levels.

[0028] 104. According to the layout list, comprehensively verify and apply the selected node layout in a rolling manner to obtain a dynamically updated site selection plan and a collaborative strategy among nodes.

[0029] In an embodiment of the present invention, the step of comprehensively verifying and applying the selected node layout according to the layout list to obtain a dynamically updated site selection plan and a collaborative strategy among nodes includes: importing the node layout suggestions and executable policy combinations in the layout list into a visual sand table system, and performing a comparison simulation with the actual logistics operation process to obtain simulation result data; based on the simulation result data, identifying nodes with insufficient resources in peak order scenarios and generating a temporary resource addition or scheduling plan; according to the cross-platform cooperation mode constraints and advantage information of each level of nodes, evaluating the matching degree of the temporary resource addition or scheduling plan to obtain an optimized resource supplement and personnel allocation mechanism; comparing and analyzing the optimized resource supplement and personnel allocation mechanism with the real-time data feedback result to obtain node layout update suggestions; based on the node layout update suggestions, combining with the potential change trend during the prediction period, dynamically adjusting the node layout suggestions to obtain a dynamically updated site selection plan and a collaborative strategy among nodes.

[0030] Specifically, when importing the node layout suggestions and executable policy combinations in the layout list into a visual sand table system and performing a comparison simulation with the actual logistics operation process to obtain simulation result data, it is necessary to load the node layout suggestions in chronological order on the sand table interface, and synchronize the instructions such as vehicle scheduling, inventory allocation, and distribution routes in each time period with the real operation data. After the sand table system reads information such as urban control time periods, transportation mileage, and order timeliness, it will construct a virtual logistics network based on the geographical locations and cross-platform collaboration status of nodes at different levels. Each simulation will detect the capacity coordination and order flow between nodes hour by hour, and dynamically mark possible load peaks or inventory shortage signals in the view. If a deviation between the distribution process and the existing strategy is detected in a certain time period, the system will record the time period and impact degree of the deviation, and note the corresponding nodes and resource usage in the log. The simulation result data output after a complete round of comparison includes quantitative indicators such as distribution completion rate, resource turnover rate, and vehicle utilization rate, and organizes the coincidence degree of each node with the policy instructions in different time periods into a specific format for further in-depth analysis and decision-making.

[0031] Based on the simulation result data, when identifying resource - insufficient nodes in peak - order scenarios and generating temporary resource addition or scheduling plans, it is necessary to find the order - volume surge segments and the transportation capacity or inventory information of nodes from the load - change curves output by the visual sand - table system. If the order influx volume of a regional distribution center during peak hours far exceeds its historical average carrying level, the system will screen the available vehicle quantity, personnel scheduling, and remaining warehouse capacity of this center during the corresponding period. By comparing the difference between the node margin and the order demand, the actual gap value is calculated. If another last - mile delivery station also has a similar gap during the same period, a temporary scheduling combination is searched for between the single node and related nodes, and different priority assignments are given with reference to the cross - level logistics structure. The generated temporary resource addition or scheduling plan may include dispatching outsourced vehicles to this distribution center during peak hours, or pre - allocating reserved warehouse space during non - regulated night hours to ensure smooth turnover in subsequent distribution links. Finally, the plan number and its applicable period are stored for matching and evaluation with other elements in the next step.

[0032] When evaluating the matching degree of the temporary resource addition or scheduling plan according to the cross - platform cooperation mode constraints and advantage information of each level of nodes to obtain an optimized resource replenishment and personnel allocation mechanism, it is necessary to retrieve the cooperation agreements and platform docking parameters of the nodes, and compare the content of the temporary addition or scheduling with the available terms therein. If the cross - platform cooperation agreement stipulates that a regional distribution center can call a part of the third - party cold - chain vehicles at night, such corresponding relationships are preferentially recorded during the matching - degree evaluation process, and it is checked whether there is a conflict between the vehicle call and the scheduling period of the management platform. If the node itself has the advantage of in - depth community cooperation, a positive weight is added to its scoring system to enable the added human or warehouse resources to fully exert the marginal effect. After confirming that there are no violations or inapplicable situations, the system makes fine - tuning of the resource replenishment and personnel configuration. If parallel demand conflicts are detected in some nodes, iterative processing is triggered again, and finally, a personnel scheduling, vehicle call, and temporary warehouse activation mechanism for each affected node is generated and output in a multi - field record manner.

[0033] When comparing and analyzing the optimized resource replenishment and personnel allocation mechanism with the real-time data feedback results to obtain suggestions for updating the node layout, it is necessary to establish an interface between the sand table system and the existing network monitoring platform, and compare key indicators such as the recent actual order volume, vehicle location, and inventory consumption rate. If there is a large deviation between the distribution efficiency or resource utilization degree and the system-set threshold, retrieve the temporary scheduling plan and optimization results output in the previous link to confirm whether the deviation is caused by external environmental changes or node execution deviations. If there are extreme fluctuations in the external environment in a short period, such as sudden traffic control in some urban areas or a sharp reduction in supplier production capacity, then revise the original personnel allocation mechanism and run a short-term simulation again in the model to verify whether the new parameter combination can cover the new bottleneck positions. After this process, the system extracts the operation performance of each node under the revised plan and generates suggestions for updating the node layout, specifying which nodes need to be adjusted in location or supplemented with more resource configurations to maintain the stable operation of the overall logistics network.

[0034] Based on the suggestions for updating the node layout and combining the potential change trends during the prediction period, when dynamically adjusting the node layout suggestions to obtain a dynamically updated site selection plan and the collaborative strategy between nodes, it is necessary to enable the prediction module and the rolling deployment tool in the management platform to estimate and analyze the order demand trend and urban control plan for the next stage. If the prediction model shows that a large-scale commodity promotion activity will occur in a certain area within a specified period, then reconstruct the time period or slightly adjust the resource quota of the node layout update suggestions so that the node has more collaborative vehicles and safety inventory during the expected high-load period. If the cross-platform cooperation partner increases the vehicle frequency of a specific route in the new round of transportation capacity investment plan, then correspondingly increase the priority of the nodes on this route during the dynamic adjustment process and present the new coordination plan in the results. After the correlation operation, the site selection plan output by the system will clarify the interaction relationship between the node spatial location, regional control time period, and supply chain production capacity at the document and visualization levels, and use the collaborative strategy between nodes as an additional explanation to minimize resource waste and distribution delays through scheduling instructions for each time period and each node.

[0035] In this embodiment, by multi-dimensionally and real-time collecting and fusion-processing the demand change characteristics and supply capacity redundancy in the multi-level logistics network of the urban agglomeration, a comprehensive data set is formed; then, based on this data set, cross-level collaborative analysis and potential node layout evaluation are carried out to generate a scenario-based deployment report; subsequently, using the adaptive calculation of conflict detection and allocation order, real-time deployment of alternative dynamic site selection plans is carried out to obtain a layout list; finally, through rolling application and comprehensive verification, iterative optimization of the selected node layout plan is carried out. The present invention can achieve flexible node site selection adjustment for the multi-level logistics network in the face of different administrative region's differential control and cross-platform transportation capacity input changes, and reduce the risk of resource allocation imbalance caused by seasonal demand fluctuations or temporary peaks.

[0036] The above describes the dynamic location selection method for logistics distribution transfer nodes in the embodiments of the present invention. Next, the dynamic location selection device for logistics distribution transfer nodes in the embodiments of the present invention will be described. Please refer to Figure 2 One embodiment of the dynamic location selection device for logistics distribution transfer nodes in the embodiments of the present invention includes: A data integration module 201, configured to perform multi-dimensional real-time collection and fusion processing on the operation data of each node in the multi-level logistics network of the urban agglomeration, so as to obtain a comprehensive data set including demand change feature items and supply capacity redundancy; A collaborative analysis module 202, configured to perform cross-level collaborative analysis and potential node layout evaluation processing according to the comprehensive data set, so as to obtain a scenario-based deployment report including multi-dimensional indicators; A conflict detection module 203, configured to perform real-time conflict detection and adaptive calculus processing of the allocation sequence on each alternative dynamic location selection scheme based on the scenario-based deployment report, so as to obtain a layout list including node layout suggestions and executable policy combinations; A rolling verification module 204, configured to perform comprehensive verification and rolling application processing on the selected node layout according to the layout list, so as to obtain a dynamically updated location selection scheme and collaborative strategies between nodes.

[0037] In the embodiments of the present invention, the dynamic location selection device for logistics distribution transfer nodes runs the above-mentioned dynamic location selection method for logistics distribution transfer nodes. The dynamic location selection device for logistics distribution transfer nodes forms a comprehensive data set by multi-dimensional real-time collection and fusion processing of the demand change characteristics and supply capacity redundancy in the multi-level logistics network of the urban agglomeration; then performs cross-level collaborative analysis and potential node layout evaluation based on this data set to generate a scenario-based deployment report; subsequently, uses real-time allocation of alternative dynamic location selection schemes through conflict detection and adaptive calculus of the allocation sequence to obtain a layout list; finally, through rolling application and comprehensive verification, iteratively optimizes the selected node layout scheme. The present invention can realize flexible node location selection adjustment of the multi-level logistics network in the face of differential control in different administrative regions and changes in cross-platform transport capacity input, and reduce the risk of resource allocation imbalance caused by seasonal demand fluctuations or temporary peaks.

[0038] The present invention also provides a computer-readable storage medium. The computer-readable storage medium can be a non-volatile computer-readable storage medium, and the computer-readable storage medium can also be a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is made to execute the steps of the dynamic location selection method for logistics distribution transfer nodes.

[0039] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, or units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0040] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0041] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; 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 recorded in the foregoing embodiments, or perform equivalent replacements for 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 various embodiments of the present invention.

Claims

1. A method for dynamic site selection of logistics distribution transfer nodes, characterized in that: The dynamic site selection method for the logistics distribution transfer node includes: The operation data of each node in the multi-level logistics network of the urban agglomeration is collected and integrated in real time in multiple dimensions to obtain a comprehensive data set containing demand change characteristic items and supply capacity redundancy; Perform cross-level collaborative analysis and potential node layout evaluation processing on the comprehensive data set to obtain a scenario-based deployment report containing multi-dimensional indicators; Based on the scenario-based deployment report, real-time conflict detection and adaptive calculation of allocation order are performed on each alternative dynamic site selection scheme to obtain a layout list including node layout suggestions and executable strategy combinations; According to the layout list, the selected node layout is comprehensively verified and rolled-up applied to obtain a dynamically updated site selection plan and inter-node coordination strategy.

2. The method for dynamic location selection of logistics distribution transfer nodes according to claim 1 is characterized in that: The multi-dimensional real-time collection and fusion processing of the operation data of each node in the multi-level logistics network of the urban agglomeration to obtain a comprehensive data set containing demand change feature items and supply capacity redundancy includes: Collect the order quantity, inventory balance, delivery time, local city control information and dynamic information related to cross-platform cooperation of the central warehouse, regional distribution center and terminal delivery station in real time to obtain the original operation data; Utilize remote monitoring equipment, blockchain traceability nodes, and API docking with supplier systems to integrate the original operating data from multiple sources to obtain a multi-dimensional data stream; Vectorizing and extracting features from the multi-dimensional data stream to obtain preliminary demand change feature items and supply capacity redundancy; Marking the preliminary demand change feature items and supply capacity redundancy with scenario-specific factors to obtain feature data with advantage and restriction labels; The feature data with advantage and constraint condition labels are fused in time and space dimensions to obtain a comprehensive data set including demand change feature items and supply capacity redundancy.

3. The method for dynamic location selection of logistics distribution transfer nodes according to claim 1 is characterized in that: The cross-level collaborative analysis and potential node layout evaluation processing are performed on the comprehensive data set to obtain a scenario-based deployment report containing multi-dimensional indicators, including: Based on the comprehensive data set, preliminary screening is performed on resource allocation of nodes at each level to obtain an initial resource allocation plan; According to the initial resource allocation plan, combined with the node restriction labels at each level, potential resource allocation conflicts are identified to obtain a conflict warning list; Based on the conflict warning list, temporarily replace or add nodes at each level to evaluate and obtain a set of alternative adjustment strategies; Analyzing the upstream supply chain capacity fluctuation characteristics of the candidate adjustment strategy set to obtain a resource allocation strategy with a flexibility score; Based on the resource allocation strategy with flexibility scoring, a cross-level resource allocation plan is generated and a multi-dimensional comprehensive evaluation is performed to obtain a scenario-based allocation report.

4. The method for dynamic location selection of logistics distribution transfer nodes according to claim 3 is characterized in that: According to the resource allocation strategy with flexibility score, a cross-level resource allocation plan is generated and a multi-dimensional comprehensive evaluation is performed to obtain a scenario-based allocation report including: Decomposing the resource allocation strategy with flexibility score into different levels to obtain resource allocation sub-strategy sets of nodes at each level; Based on the resource allocation sub-strategy set, the resource flow between nodes at each level is simulated and analyzed to obtain a resource flow situation diagram, and the resource utilization and load balancing degree of nodes at each level are calculated according to the resource flow situation diagram to obtain a node performance indicator matrix; Using a multi-objective optimization algorithm, weight allocation and comprehensive calculation are performed on the node performance indicator matrix to obtain a cross-level resource allocation plan; Conduct a multi-dimensional sensitivity analysis on the cross-level resource allocation scheme to evaluate its robustness in different scenarios and obtain a scenario-based allocation report.

5. The method for dynamic location selection of logistics distribution transfer nodes according to claim 1 is characterized in that: Based on the scenario-based deployment report, each candidate dynamic site selection scheme is subjected to real-time conflict detection and adaptive calculation processing of the allocation order to obtain a layout list including node layout suggestions and executable strategy combinations, including: Extract the resource urgency ranking of nodes at each level from the scenario-based deployment report, combine traffic flow and control period data, and generate a preliminary cross-level resource sharing sequence; By using the collaborative game and scheduling engine, the contradictions and mutually exclusive conditions between nodes at each level in the cross-level resource sharing sequence are iteratively solved to obtain a conflict adjustment solution; Based on the conflict resolution scheme, identify data synchronization delays or incomplete information in cross-platform collaboration and generate a redundant deployment plan; The redundant deployment plan is iterated and self-learned for multiple times to obtain node layout suggestions, and according to the node layout suggestions, an executable strategy combination is configured for each level node to obtain a layout list.

6. The method for dynamic location selection of logistics distribution transfer nodes according to claim 5 is characterized in that: The redundant deployment plan is iterated multiple times and self-learned to obtain node layout suggestions, and according to the node layout suggestions, an executable strategy combination is configured for each level node to obtain a layout list including: Performing Monte Carlo simulation on the redundant deployment plan to generate a large number of random scenarios, evaluating the deployment effect in each scenario, and obtaining a deployment effect distribution map; Using a clustering algorithm to classify and extract common features based on the distribution diagram of the deployment effect, to obtain an optimized deployment pattern set; The optimized deployment pattern set is input into a reinforcement learning model, and an adaptive node layout strategy is obtained through repeated training and strategy optimization; According to the adaptive node layout strategy, nodes at each level are dynamically scored and sorted, the optimal node combination is selected, and node layout suggestions are obtained; Based on the node layout suggestions, combined with the historical performance and predicted needs of nodes at each level, resource allocation and collaboration mechanisms are customized for each node, and a layout list including node layout solutions and hierarchical strategy combinations is obtained.

7. The method for dynamic location selection of logistics distribution transfer nodes according to claim 1 is characterized in that: The step of performing comprehensive verification and rolling application processing on the selected node layout according to the layout list to obtain a dynamically updated site selection plan and inter-node coordination strategy includes: The node layout suggestions and executable strategies in the layout list are combined and imported into the visual sandbox system, and compared and simulated with the actual logistics operation process to obtain simulation result data; Based on the simulation result data, identify resource-deficient nodes under peak order scenarios and generate temporary resource addition or scheduling plans; According to the cross-platform cooperation mode constraints and advantage information of nodes at each level, the matching degree of the temporary resource addition or scheduling plan is evaluated to obtain an optimized resource supplement and personnel deployment mechanism; Compare and analyze the optimized resource replenishment and personnel deployment mechanism with the real-time data feedback results to obtain node layout update suggestions; Based on the node layout update suggestion and combined with the potential change trend in the forecast period, the node layout suggestion is dynamically adjusted to obtain a dynamically updated site selection plan and inter-node coordination strategy.

8. A dynamic site selection device for logistics distribution transfer nodes, characterized in that: The logistics distribution transfer node dynamic site selection device comprises: The data integration module is used to collect and integrate the operation data of each node in the multi-level logistics network of the urban agglomeration in real time in multiple dimensions to obtain a comprehensive data set containing demand change feature items and supply capacity redundancy; A collaborative analysis module, used to perform cross-level collaborative analysis and potential node layout evaluation processing based on the comprehensive data set to obtain a scenario-based deployment report containing multi-dimensional indicators; A conflict detection module, for performing real-time conflict detection and adaptive calculation of allocation order on each candidate dynamic site selection scheme based on the scenario-based deployment report, and obtaining a layout list including node layout suggestions and executable strategy combinations; The rolling verification module is used to perform comprehensive verification and rolling application processing on the selected node layout according to the layout list to obtain a dynamically updated site selection plan and inter-node coordination strategy.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the method for dynamic site selection of logistics distribution transit nodes as described in any one of claims 1-7 are implemented.

Images