Data maintenance system and method for distributed material packing and coding system

Abstract

The application provides a data maintenance method based on a distributed material case packing coding system, which comprises the following steps: obtaining a single coding binding record, calculating a credibility measurement value for each coding node; constructing a coding behavior causal relationship structure; identifying a conflict path formed by multiple bindings of the same material or the same bag code in a period, and calculating a path reliability measurement value of all the conflict paths; performing abnormal path identification on the conflict paths, and screening out a conflict path with the highest path reliability measurement value as a reserved path; dynamically determining a verification strategy of the binding record in the current case packing period based on the reserved path, and generating a coding binding record that passes the verification; and performing redundant persistent storage based on the coding binding record that passes the verification and according to a corresponding connection degree in the coding behavior causal relationship structure. The application unifies the consistency of data, the efficiency of the system and the traceability in the same framework.

Description

Technical Field

[0001] This invention belongs to the field of data maintenance, and in particular relates to a data maintenance system and method based on a distributed material packing and coding system. Background Technology

[0002] As the manufacturing industry moves towards high automation and large-scale production, material packing and coding systems are gradually evolving into distributed systems with multiple workstations operating in parallel and cross-node collaborative processing. The coding process not only requires establishing a stable one-to-one correspondence between materials and bag codes, but also needs to ensure the consistency and availability of the bound data in subsequent packing verification, logistics traceability, and quality management under high concurrency and asynchronous execution conditions. Existing packing and coding systems typically use multiple coding nodes operating simultaneously and centrally write the binding results into the backend system to improve overall efficiency. However, in actual operation, different coding nodes have objective differences in network status, equipment load, and operational stability, inevitably leading to delays, duplications, and disordered ordering of bound data during generation and transmission. While these problems can be handled with simple rules when the data volume is small or the number of nodes is limited, they easily evolve into structural conflicts as the distributed scale expands, making it difficult for the system to accurately determine which binding relationships originate from actual business processes and which are merely caused by data collection conditions or concurrent execution. Meanwhile, when verification or rejection operations are triggered at key nodes such as packing completion and box sorting, existing solutions mostly rely on fixed ratios or static thresholds, lacking the ability to characterize the overall state of coding behavior. They can often only be remedied afterward through manual intervention or full backtracking, which affects production cycle and increases data maintenance costs.

[0003] Therefore, how to systematically characterize the generation process, evolution relationship and operation status of coding data in a distributed bin packing coding scenario, and thereby realize a data maintenance mechanism that enables conflict identification, controllable processing and traceable results, has become a key issue restricting the further development of existing technologies. Summary of the Invention

[0004] The purpose of this invention is to propose a data maintenance system and method based on a distributed material packing and coding system to solve the above-mentioned problems.

[0005] To achieve the above objectives, a data maintenance system based on a distributed material packing and coding system is provided in a first aspect of the present invention, the system comprising:

[0006] The coding acquisition unit is used to acquire a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time marker, and a coding node identifier. Based on the single coding binding record, a reliability metric value is calculated for each coding node.

[0007] The causal relationship construction unit is used to construct a causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier.

[0008] The conflict and anomaly analysis unit is used to identify conflict paths formed by multiple bindings of the same material or the same bag code within a cycle based on the causal relationship structure of the coding behavior, and to calculate the path reliability metric value of all conflict paths; based on the path reliability metric value, the conflict paths are identified as abnormal paths, and the conflict paths with the highest path reliability metric value are selected as retained paths, while the others are identified as abnormal paths; at the same time, the coding binding records corresponding to the retained paths are used as valid candidate binding chains within the current packing cycle;

[0009] The valid path acquisition unit is used to dynamically determine the verification strategy of the binding record in the current packing cycle based on the reserved path, and generate the verification-passed coded binding record as a valid binding record;

[0010] The coding statistics unit is used to perform redundant persistent storage based on the verified coding binding records and the corresponding connection degree in the causal relationship structure of the coding behavior, while recording its path index information.

[0011] A second aspect of the present invention provides a data maintenance method for a distributed material packing and coding system, the method comprising:

[0012] Obtain a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time stamp, and a coding node identifier. Calculate a reliability metric value for each coding node based on the single coding binding record.

[0013] Based on the credibility metric and the corresponding coding node identifier, a causal relationship structure for coding behavior is constructed.

[0014] Based on the causal relationship structure of the coding behavior, conflict paths formed by multiple bindings of the same material or the same bag code within a cycle are identified, and the path reliability metric of all conflict paths is calculated. Based on the path reliability metric, abnormal paths are identified for the conflict paths, and the conflict path with the highest path reliability metric is selected as the retained path, while the others are identified as abnormal paths. At the same time, the coding binding record corresponding to the retained path is used as a valid candidate binding chain within the current packing cycle.

[0015] Based on the reserved path, the verification strategy for the binding records in the current packing cycle is dynamically determined, and the verification-passed coded binding records are generated as valid binding records.

[0016] Based on the verified binding record, redundant persistent storage is performed according to the corresponding connection degree in the causal relationship structure of the binding behavior, and its path index information is recorded at the same time.

[0017] Furthermore, the step of calculating a trustworthiness metric for each coded node based on the single coded binding record specifically involves:

[0018] For the same coding node, the time interval distribution of coding binding records within the sliding time window is analyzed to evaluate time stability.

[0019] To assess repeatability, the system detects whether identical combinations of material identifiers and bag codes exist within a short period of time.

[0020] Obtain operational monitoring information of the coded nodes to assess their operational status;

[0021] Based on the combined evaluation results of time stability, repeatability, and node operating status, the reliability metric is calculated.

[0022] Furthermore, the construction of the causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier specifically involves:

[0023] The coded binding records are grouped according to the coded node identifier, and a directed association representing the order of chronological sequence is established within each group according to the order of the coded time markers.

[0024] If the time difference of the coding time marker falls within the preset candidate time range for the same box between coding binding records identified by different coding nodes, a candidate association representing the concurrency relationship is established.

[0025] By combining the credibility metric, node running stability flag, and time interval, a correlation weight is calculated for each directed association and candidate association.

[0026] Furthermore, the node operation stability marker corresponds to the dimension of the coded node identifier and is used to characterize the operational stability of the current coded node identifier within the current sliding time range.

[0027] Furthermore, based on the causal relationship structure of the coding behavior, conflict paths are identified due to the same material or the same bag code being bound multiple times within a period, specifically as follows:

[0028] Based on the principle that only one bag code is allowed to correspond to the same material identifier within the same packing cycle, and only one material identifier is allowed to be bound to the same bag code within the same packing cycle, similar nodes are aggregated to filter out the set of nodes with multiple binding candidates.

[0029] Based on each conflict trigger point in the node set, multiple candidate causal paths are constructed by backtracking along directed associations with the trigger point as the endpoint, thus forming a conflict path. During the backtracking process, links from the same coded node are expanded along the workstation sequence edge, and links across coded nodes are expanded along concurrent candidate edges. During traversal, the most recent record index mapping of the coded node is used, and the latest record number of each coded node is used as the backtracking starting point. Backtracking is only performed along the sequence edge of that node to a preset backtracking depth. This restricts path construction to the recent behavior chain of each node during high-concurrency periods, avoiding unbounded expansion of the entire graph.

[0030] Furthermore, the conflict path reliability metric accumulates the weights of each associated path and penalizes frequent jumps across coding nodes in the path, and is calculated based on the number of jumps in the coding node dimension of the conflict path.

[0031] Furthermore, the step of dynamically determining the verification strategy for binding records within the current packing cycle based on the reserved path, and generating verification-passed coded binding records, specifically involves:

[0032] Obtain the nodes of the abnormal path and the reserved path. The number of failed behavior nodes corresponds to the abnormal path, and the number of behavior nodes participating in packing in the current packing cycle corresponds to the reserved path.

[0033] The structural anomaly ratio is calculated based on the number of failed behavioral nodes and the total number of behavioral nodes involved in packing during the current packing cycle.

[0034] Calculate the average of the credibility metrics of all valid nodes in the current period;

[0035] The verification status index for the current packing cycle is calculated by combining the average of the structural anomaly ratio and the reliability metric.

[0036] Based on the preset verification strategy mapping table and the interval in which the verification status index is located, the final verification strategy is determined. The strategy includes at least exemption from inspection, random inspection, and full inspection.

[0037] Furthermore, the execution of redundant persistent storage specifically includes:

[0038] Obtain the number of direct predecessors and direct successors of the valid binding records in the causal relationship structure of the coding behavior;

[0039] The redundancy factor driven by the structure is calculated based on the sum of the number of direct predecessors and the number of direct successors.

[0040] The number of redundant replicas is determined based on the redundancy factor, and the binding records are written to the corresponding number of redundant storage nodes;

[0041] In the step of recording path index information:

[0042] The identification information of the direct predecessor and direct successor nodes in the causal relationship structure of the coding behavior is saved in the binding record.

[0043] Furthermore, when selecting redundant storage nodes, the principle of distribution should be followed, and storage nodes located in different physical or logical locations should be given priority.

[0044] The beneficial technical effects of the present invention are at least as follows:

[0045] This invention proposes a data maintenance system and method based on a distributed material packing and coding system. By continuously characterizing and structuring the state of coding behavior during operation, data maintenance is transformed from judging individual records to collaborative processing based on overall behavioral relationships. Starting with the coding data generation process, the system characterizes the reliability of binding records generated by different coding nodes at the operational state level. Based on this, it constructs a temporal and concurrent relationship structure between coding behaviors, thus organizing discrete binding operations into analyzable and inferable behavioral chains. During data maintenance, the system no longer relies on simple deduplication or static rules, but instead identifies paths that may lead to binding conflicts based on the behavioral relationship structure. Abnormal behaviors are stripped and marked on a path-by-path basis, allowing the remaining data to reflect an evolution result closer to the actual packing process. Furthermore, the system dynamically determines the packing verification strategy based on the overall state of the coding behavior structure, matching the verification strength with the current data consistency level, thereby ensuring data reliability while avoiding unnecessary verification overhead. For the finally confirmed and trusted binding results, the system implements differentiated redundant storage and path indexing records based on its positional characteristics in the behavioral structure. This enables the coded data to be quickly located and reconstructed in subsequent anomaly recovery, audit backtracking, or system reconstruction. Through this method, the present invention achieves end-to-end maintenance of coded data in a distributed binning environment, from generation, conflict handling, verification decisions to long-term storage, ensuring data consistency, system efficiency, and traceability are uniformly guaranteed within the same framework. Attached Figure Description

[0046] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0047] Figure 1 This is a flowchart of the data maintenance method for the distributed material packing and coding system of the present invention. Detailed Implementation

[0048] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0049] like Figure 1 As shown in the embodiment of the present invention, a data maintenance method for a distributed material packing and coding system is provided, the method comprising:

[0050] S1. Obtain a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time stamp, and a coding node identifier. Calculate a reliability metric value for each coding node based on the single coding binding record.

[0051] Specifically, in a distributed material packing and coding system, the coding operation is typically completed in parallel by multiple coding stations. After each coding station completes the binding of a material to a bag code, it sends a binding record to the central system. Due to differences in network conditions, equipment status, and processing load among the coding stations, even if the business processes are completely identical, the actual conditions experienced by different binding records during generation and transmission may differ. If this difference is not explicitly characterized before the data enters the maintenance process, it will amplify its impact in subsequent data consistency maintenance, conflict removal, and other operations, ultimately making it difficult for the system to reliably distinguish between business anomalies and data collection anomalies.

[0052] This step uses a single coded and bound record as the basic processing unit. Each record must contain at least a material identifier. Bag code label , coding time stamp and the coded node identifier In actual deployment, The coding is usually issued by the production management system or material management system before the material enters the packing station and remains unchanged throughout the entire coding process; The code is generated by the code reader at the coding station when the bag code is scanned, and uploaded along with the binding record; The coding node records the information immediately upon completing a scan confirmation and generating a binding record. This corresponds to a specific coding station or edge node number, used to distinguish coding actions occurring at different physical locations. This information originates from the normal operation of the coding system itself and does not require additional configuration of new acquisition interfaces.

[0053] In practical implementation, the system performs a specific step for each coding node at the data access layer or locally at the coding node. Maintain a set of binding records within a sliding time range. This set reflects the node's coding behavior status over a recent period. It is updated by adding a new binding record to the set whenever the node generates one, while simultaneously removing older records that are outside the preset time range. In this way, the system can continuously obtain behavioral samples of the node during its current operational phase without needing to store large amounts of historical data for extended periods.

[0054] Based on the above record set, the system first analyzes the temporal relationship between adjacent bound records on the same coding node. If, under normal operating conditions, the coding cycle of this node is relatively stable, the time intervals between adjacent records will exhibit a relatively concentrated distribution. When the time stamp corresponding to a certain record... When a record deviates significantly from its distribution range compared to its most recent record, the system considers it a potential instance of data acquisition instability and accordingly lowers the stability characterization value of that record over time. This process does not rely on a fixed threshold but rather on a reference range formed by the node's recent behavior, thus adapting to the actual cycle time differences at different workstations. Specifically:

[0055] Maintain a sliding time window for each coding node, and only retain the most recently generated set of binding records for that node;

[0056] Analyze the time intervals between adjacent records within the window to form a statistical distribution (such as mean, standard deviation, or interval range) that reflects the recent normal operating rhythm of the node.

[0057] When a new record is generated, the time interval between it and the previous record is calculated. If this interval deviates significantly from the statistical distribution range recently formed by the node, the record is considered unstable in the time dimension.

[0058] Innovation: Using the node's own dynamic historical behavior as a reference, rather than a globally fixed threshold, it can adapt to the actual operating speed differences of different workstations.

[0059] At the same time, the system performs a duplicate check on the bound record set within the same time frame. Specifically, for newly added records to the set, the system searches for records with the same material identifier. Bag code label The record has already appeared in this set. If multiple identical binding combinations are found within a short period, this usually corresponds to repeated triggering of the code reader or retransmission of communication, rather than actual repeated code assignment. Based on this, the system adjusts the characterization of the record in the repeatability dimension to reflect its abnormal characteristics during the acquisition process. Specifically:

[0060] Within the sliding time window of the same coding node, check if there are any binding records where the material identifier and bag code are exactly the same;

[0061] If the exact same binding combination appears within a very short time window (e.g., within a few seconds), it is most likely due to communication retransmission or scan jitter, rather than two real independent coding operations.

[0062] The system will give a negative evaluation on the repeatability dimension for such records that are highly repeated in a short period of time.

[0063] Innovation: By performing repeated detection within a short time window, technical noise is effectively filtered out, while reasonable duplication that may exist in business operations (assigning codes to the same material with a longer interval).

[0064] Furthermore, the system combines the operational status information of the coded nodes to characterize the overall stability of the nodes. In engineering implementation, this operational status information can be directly derived from the node's local operational monitoring module, such as whether the node has experienced record processing delays, reporting interruptions, or anomaly recovery within the current time frame. This type of information does not directly participate in business logic judgments, but it effectively reflects whether the node is in a stable operating state during this phase. Specifically:

[0065] The status information is directly obtained from the local running monitoring module of the coding node.

[0066] This information includes, but is not limited to: whether the node has reported a processing delay alert, whether there is a history of communication interruption, whether it is in an unstable period after an anomaly recovery, and whether the CPU / memory load is consistently too high.

[0067] Based on these observable engineering indicators, the operational stability of nodes during this time period is comprehensively scored or classified (e.g., stable, sub-healthy, abnormal).

[0068] Innovation: The reliability of the node itself is used as an important factor in the credibility of the data. Even if a record has no abnormalities in its time and content, if it is generated from an unstable node, its overall credibility should be reduced.

[0069] After synthesizing the results from the above three aspects, the system calculates a reliability metric for each coded binding record. Its form is as follows:

[0070] ;

[0071] in, This represents the stability profile of the record in the time series of the coding node, and its value is obtained by mapping the statistical characteristics of the recent coding time interval of the node. This indicates the repeatability profile of the record over a short time period, and its value is obtained from the repeatability detection process in the bound record set; This represents the operational stability characterization result of the coded node within the current time range, and its value is obtained by mapping the node's operational state information. To ensure comparability between characterization results from different sources, the above quantities have been converted to a uniform scale before being used in the calculation. , and Preset weights are used to reflect the relative impact of different unstable factors on the trustworthiness of binding records in the box-packing coding scenario. Through this implementation, the system can generate a trustworthiness profile for each binding record based on actual operating conditions before the coded data enters the subsequent maintenance process. After completing this step, the system output includes an accompanying trustworthiness metric. The set of assigned binding records, and the identifier of the assigned node. Stability markers for nodes formed by dimensions ,in and A one-to-one correspondence is used to characterize the stability of the coded node within the current sliding time range.

[0072] S2. Based on the credibility metric and the corresponding coding node identifier, construct the causal relationship structure of coding behavior.

[0073] Specifically, this step builds upon the output of step one. Step one has already attached a trustworthiness metric to each coded and bound record. It also provides the identifier of the assigned node. The corresponding node operation stability markers. This step organizes these discrete records, which are already "characterized by reliability," into a causal relationship structure. This allows subsequent conflict identification and anomaly removal to move beyond the level of individual records and instead perform path-level reasoning and processing along the sequence and concurrency of coding actions. The causal relationship here does not pursue abstract inference but is established around two facts of the packing and coding scene: coding events at the same workstation naturally have a sequential order; coding events at different workstations may occur concurrently in time, but under the packing cycle, they can still be confined to a "candidate time range for the same packing box," thus forming an operable association structure.

[0074] In the specific implementation, the system first identifies the coded node. Group the input records by dimension, and in each Groups are marked by the time of coding. In ascending order. For the same... For two adjacent records within a group, the system establishes a directed association to express the objective order in which "the previous coding occurs before the next coding" during actual operation at that workstation. This type of association is one of the most stable structural information in the packing and coding process, capable of abstracting the coding process of each workstation into a continuous chain of behaviors. For different Records between groups are time-stamped by the system. Constructing "concurrent candidate associations": When the time difference between two records falls within the configured candidate time range for the same box, they are considered as candidate associations that can be concurrently triggered by the same packing process. This candidate association is not directly equivalent to a business binding relationship, but rather provides a feasible search boundary across workstations for subsequent conflict path identification, so that the graph structure can both cover the reality of parallel coding and avoid forming too dense invalid connections globally.

[0075] Furthermore, after the above two types of associations are established, the system calculates the association weight for each directed association. , used to quantify "by the first The record points to the first The causal reliability of each record. This weight design incorporates both the reliability characterization from step one and the node stability characterization, giving the graph structure the inherent property that "weak edges are more easily sacrificed in subsequent conflict resolution." This explicitly maps common factors in binning and coding scenarios, such as communication retransmission, short-term jitter, and node instability, to the structural level. The association weight is calculated using the following formula:

[0076] ;

[0077] in, For the purpose of recording Pointing to record The association weight; and The output of step one is the first , No. The credibility metric for each record; and They represent the first , No. The identifier of the assigned node to which the record belongs. and Run stability flags for the corresponding nodes output in step one; This represents the normalized interval between two records over time, used to express the packing time constraint that "the closer the times are, the higher the probability of forming a valid causal relationship"; The preset attenuation coefficient is used to control the influence of time intervals on weights. The introduction of the exponential term in the formula makes the candidate associations across workstations naturally exhibit a sparsity effect in time, avoiding the formation of a large number of weak long-distance connections during high-concurrency periods, while not affecting the sequential links within the same workstation.

[0078] Furthermore, time interval measurement The result is calculated from the timestamps of the two records and normalized within the candidate time range of the same box, as follows:

[0079] ;

[0080] in, For normalized interval measurement; and The first , No. The time stamp assigned to each record; The timescale parameter corresponding to the candidate time range for the same container is determined by the on-site packing cycle time and the concurrent mode configuration of the workstation. Through this normalization process, The value of and The settings are directly bound to the packing and coding scenario, ensuring that the causal relationship modeling remains consistent across different production lines and cycle times.

[0081] To facilitate path-level retrieval along the graph structure in subsequent steps, this step, while generating the causal relationship structure, also maintains a "recent record index" for each coded node. This index is used to quickly locate the end of the node's behavioral chain and the connection points of adjacent records. The key stores the latest record number currently added to the map for this node, and it is updated as new records are added. This process does not introduce new data features, but only utilizes the data from the input. This can be achieved by matching the sorted record numbers, and can be directly stored in a memory-mapped table or key-value store.

[0082] The output control in this step consists of two items. One is the causal relationship structure of the coding behavior. It includes behavioral nodes, two types of directed associations, and the weight corresponding to each association. Secondly, there is the mapping of the most recent record index of the node, based on the node identifier assigned by the key. Locate the latest record number for this node. Subsequent conflict identification and anomaly removal steps will follow. As the main input, this index is used to quickly expand path search and locate conflicting links in multi-node concurrent scenarios, thereby transforming the distributed concurrency characteristics of the bin packing and coding system into an operable structured maintenance object.

[0083] S3. Based on the causal relationship structure of the coding behavior, identify conflict paths formed by multiple bindings of the same material or the same bag code within the cycle, and calculate the path reliability metric value of all conflict paths; based on the path reliability metric value, identify abnormal paths of the conflict paths, and select the conflict path with the highest path reliability metric value as the retained path, and the others as abnormal paths; at the same time, the coding binding record corresponding to the retained path is used as a valid candidate binding chain within the current packing cycle.

[0084] Specifically, this step is executed based on the causal relationship structure of the coding behavior formed in step two. Through the continuous processing of steps one and two, the system has transformed each coding binding record from a simple four-element information... Expanded to include credibility metrics The behavior nodes were identified, and directed associations and their weights were further established between these nodes. To form a causal relationship structure In the real-world process of packing and coding, conflicts are often not as simple as whether a certain record is duplicated, but rather manifest as the same material identifier. Different bag code identifiers were assigned along different paths in a short period of time. or the same bag code identifier Multiple material identifiers Contention. Due to the simultaneous presence of factors such as distributed workstation parallel coding, network retransmission, and short-term node instability, these conflicts are often accompanied by complex sequence and concurrency relationships. This step, focusing on this scenario characteristic, adopts a structured processing approach "based on paths as basic units," unifying conflict identification, conflict interpretation, and anomaly handling within a causal relationship structure. This process is completed in the middle, thus providing an input basis with eliminated structural contradictions for subsequent full-box verification and data retention.

[0085] The input for this step strictly comes from the output of step two, and includes two items. The first is the causal relationship structure of the coding behavior. This structure includes behavioral nodes, directed associations between nodes, and corresponding association weights. Secondly, there is the recent record index mapping for each assigned node, used to retrieve data from each assigned node. The behavior chain ends are quickly located and the structure is expanded for traversal. Because... The structure has been embedded The node stability markers and time interval decays are used in this step. Instead of recalculating at the original record level during conflict handling, these characteristics are directly used as structural weights to drive path selection and anomaly handling order.

[0086] Furthermore, in practical implementation, the system first establishes a causal relationship structure. Location conflict trigger point. Location method is based on business constraints: same material identifier. Only one bag code is allowed within the same packing cycle. Only one material identifier is allowed to be bound within the same packing cycle. The system is based on and To aggregate similar nodes within the structure, the system filters out a set of nodes with multiple binding candidates. For each conflict trigger point, the system backtracks along the directed association from the trigger point to construct multiple candidate causal paths. During the backtracking process, nodes originating from the same key node... The links are expanded along the workstation sequence edges, and the links across coding nodes are expanded along the concurrent candidate edges. During traversal, the index mapping of the most recent record of the coding node is used to map the links from each coding node. The latest record number is used as the backtracking starting point. Backtracking is only performed along the sequential edges of that node until the preset backtracking depth. This restricts path construction to the recent behavior chain of each node during high-concurrency periods, avoiding unbounded expansion of the entire graph.

[0087] To form an interpretable and executable processing decision among multiple candidate paths, the system calculates a path reliability metric for each candidate path. This metric not only accumulates the association weights on the path, but also incorporates a "path structure regularization term" for bin packing and coding scenarios to suppress accidental associations that occur on paths with frequent cross-workstation jumps under high concurrency. The path reliability metric is calculated as follows:

[0088] ;

[0089] in, A reliable metric for candidate causal paths (i.e., conflicting paths); This represents the set of directed associations on the candidate path; The correlation weights are the output of step two; This represents the number of jumps in the path along the dimension of the coding nodes, i.e., the coding node identifiers of adjacent nodes along the path. The number of times changes occurred; The preset penalty coefficient is used to reflect the instability tendency of "a single material forming a long link across multiple workstations in a short period of time" in the packing and coding process. This regularization term makes "continuous links within the workstation" that are more in line with the physical process more likely to become highly reliable paths under the same weight product condition, while multiple cross-workstation jump links formed by concurrent accidental triggering will be naturally suppressed.

[0090] Furthermore, after obtaining the reliability metric for each candidate path, the system sorts the path set under the same conflict trigger point, selects the path with the highest reliability metric as the retained path, and uses its corresponding binding relationship as the valid candidate binding chain for the current packing cycle. The remaining paths are identified as anomalous paths. The handling of anomalous paths follows the principle of "path-level stripping with minimal impact": the system does not directly delete the original record, but rather strips it from the causal relationship structure. The system marks the behavior nodes on abnormal paths as invalid, preventing them from participating in subsequent binning checks and data retention. To ensure structural consistency, after the invalidation is executed, the system also logically sets the inbound and outbound edges of the node to an unavailable state, so that subsequent traversals will not pass through the abnormal node again, thereby avoiding the repeated occurrence of conflicting paths in subsequent steps.

[0091] Another common problem in packing and coding is "competition for the same bag code," which means multiple material identifiers are used. Pointing to the same bag code within a similar time period In this case, the system also... Construct candidate paths for conflict trigger points and use the same The calculation method involves path sorting and path retention selection. Because... Having taken into account both record reliability and node stability, bag code contention conflicts will naturally manifest in the reliability metric ranking as "paths from unstable nodes or low-reliability records are more easily stripped away," thus achieving a processing effect consistent with field engineering experience.

[0092] After resolving conflict paths, the system generates conflict statistics for subsequent steps, which include the total number of behavioral nodes involved in packing during this packing cycle. and the number of behavior nodes marked as invalid. . This is obtained by counting the number of behavior nodes that satisfy the time range constraint within this packing cycle. The result is obtained by counting the behavior nodes whose failure flags are true; simultaneously, the number of affected material identifiers and bag code identifiers can be recorded as supplementary statistics. This statistical result comes directly from the failure flag operation in this step and does not require the introduction of new collected data. Meanwhile, the system establishes a causal relationship structure. The node status and associated availability status are synchronized and updated to ensure that subsequent steps are executed within the specified timeframe. At that time, you will only see a valid behavior diagram that shows "structural contradictions have been removed".

[0093] S4. Based on the reserved path, dynamically determine the verification strategy for the binding records in the current packing cycle, and generate the verification-passed coded binding records as valid binding records.

[0094] Specifically, this resulted in an updated causal relationship structure for coding actions. Each coding action node has been marked as valid or invalid, and a set of conflict statistics corresponding to the current binning cycle has also been obtained. This state means that the system is no longer dealing with raw, mixed coding records, but with a set of structured cleaned data. The core objective of this step is to transform this "structural state" into "directly executable verification decisions" on this basis, thereby providing a clear and operable binding verification strategy when binning is completed.

[0095] In the specific implementation process, the system uses the packing cycle as the processing boundary to determine the structure. A state summary is performed at the execution layer. First, the system filters the set of behavior nodes participating in the current packing cycle from the structure. This set is determined without additional judgment; it is directly based on the packing start and end times recorded by the packing control system within this cycle. Chinese coding time stamp Nodes falling within this time frame are filtered. Then, the system counts the number of valid nodes and the number of invalid nodes within this set, thus obtaining the anomaly percentage at the causal structure level in the current container. This anomaly percentage reflects how many coding behaviors are still excluded in the path-level judgment after structural cleaning has been completed, directly quantifying the structural stability.

[0096] Furthermore, after completing the structural-level statistics, the system further evaluates the credibility metrics corresponding to the valid behavioral nodes. Aggregation processing is then performed. Since abnormal paths were already removed in step three, only nodes deemed valid behaviors participate in aggregation in this step. The system then processes these nodes... The summation and averaging are used to characterize the "overall credibility level of the retained coding behavior" within the current bin. This processing method differs from the credibility judgment of a single record; it reflects whether the current bin still shows signs of acquisition instability after undergoing distributed coding, concurrent execution, and structural conflict removal.

[0097] Based on the above two types of statistical results, the system constructs a state index for driving the determination of the verification strategy. This indicator combines the proportion of structural anomalies with the credibility aggregation results in a weighted manner, allowing two different types of information to participate in decision-making on the same scale. Its calculation form is as follows:

[0098] ;

[0099] in, This indicates the verification status indicators for the current packing cycle; This indicates the number of behavior nodes marked as failed in step three during the packing cycle. This indicates the total number of nodes involved in the packing process during this period; The proportion of structural anomalies; This represents the credibility metric value corresponding to all valid action nodes. The average result; Preset weighting parameters are used to adjust the relative influence of structural anomalies and credibility factors in the verification and judgment process. This combination method allows for adjustments when there are many abnormal paths at the structural level, or when the overall credibility of the remaining valid nodes is low. This will increase accordingly, thus reflecting the potential risks in the current coding quality of the enclosure.

[0100] In obtaining Subsequently, the system does not perform complex real-time inference or dynamic model calculations. Instead, it uses the metric as an index to directly query a pre-configured verification strategy mapping table. This mapping table is configured by process engineers or system administrators during the system deployment phase based on production line cycle time, quality requirements, and manual verification costs. Its typical form consists of several consecutive intervals, each corresponding to a specific verification strategy. For example, when... When the value falls into the lowest range, the system determines that the current enclosure structure is stable and highly reliable, and directly adopts an inspection-free or extremely low-percentage sampling strategy; when When the data falls into the middle range, the system uses a proportional verification strategy to check the binding results within the container; when When the container falls into the highest interval, the system determines that the current container has a high risk and triggers a full verification strategy to verify all binding relationships one by one. Since this process only involves interval matching and strategy reading, it can quickly provide results after the packing completion signal is triggered, meeting the on-site requirements for response timeliness. After the verification strategy is determined, the system outputs the strategy result along with the set of valid behavior nodes within the current packing cycle. Upon receiving this output, the execution layer can directly execute the corresponding verification actions according to the strategy requirements without having to re-access the causal relationship structure or recalculate conflict situations. In this way, the structural judgment completed in step three is fully inherited and transformed into a decision result that can directly drive on-site operations.

[0101] S5. Based on the verified code binding record, perform redundant persistent storage according to the corresponding connection degree in the causal relationship structure of the code assignment behavior, and record its path index information at the same time.

[0102] Specifically, this step is executed after the binding verification strategy determination is completed in step four, marking the formal transition of the assigned data from "runtime processing" to "lifecycle maintenance." In the preceding steps, the system has already used a causal relationship structure... The coding process underwent structured cleaning, and the verification strategy corresponding to the current packing cycle was determined based on the structure state. Simultaneously, the set of valid behavior nodes that participated in and passed the verification was identified. At this point, the system's task is not to continue judging data correctness, but rather to solidify the confirmed trustworthy binding results in a way that supports long-term maintenance, anomaly recovery, and audit traceability. This step, centered on this goal, completes the redundant persistence of trustworthy data and the corresponding causal path index records, ensuring that data maintenance truly reaches an executable and recoverable engineering level.

[0103] In its implementation, the system first filters the set of valid behavior nodes based on the verification strategy results, forming a "persistent node set." This filtering operation does not involve new rule judgments; instead, it directly reads the pass / fail flags generated by the verification execution layer in step four. For example, after the execution layer completes its sampling inspection, it writes a verification result flag to each sampled node. This step only selects nodes marked as pass for further processing. This ensures that the persistence logic and verification logic maintain a clear boundary of responsibility within the system.

[0104] After determining the set of persistent nodes, the system performs redundant persistence operations on each behavior node. The binding record corresponding to each behavior node contains the material identifier. Bag code label , coding time stamp , coded node identifier and credibility metrics This information has already been determined in steps one and three, and no changes will be made to its content in this step. The system first writes the binding record to the main storage node, which is usually the core business database of the packing and coding system, used to support daily queries and downstream system integration.

[0105] After completing the main storage write, the system uses the behavior node in the causal relationship structure. The system determines the location of the node in the structure, the number of redundant replicas to be written, and the target node for redundancy. The key point here is that the redundancy strategy is not a uniform, fixed number of replicas, but rather directly related to the "criticality" of the encoding action within the structure. Specifically, the system has already obtained the location of the node performing the action in steps two and three. The number of direct predecessors and direct successors of a node reflects its connectivity in the coding behavior evolution chain. A higher connectivity indicates a greater impact on path backtracking and anomaly recovery should the node be lost.

[0106] Furthermore, based on the above considerations, the system calculates a structure-driven redundancy factor for each persistent behavior node. This is used to determine the number of redundant replicas that need to be written to the node, and its calculation form is as follows:

[0107] ;

[0108] in, This indicates the number of redundant replicas ultimately used by the node performing this action; The maximum redundant replica limit is pre-configured for the system to constrain the use of storage resources; This indicates that the behavior node is in the causal relationship structure. The number of direct predecessors in; This indicates that the behavior node is in The number of direct successors in; This is a scaling factor used to adjust the impact of structural connectivity on redundancy strength. Through this calculation method, nodes located at critical positions in the coding link naturally receive a higher redundancy level, while nodes at the link edge use a lower redundancy level, thus achieving a balance between reliability and storage cost.

[0109] After determining the number of redundant replicas, the system selects the corresponding number of target nodes from the pre-configured redundant storage node pool and writes the binding records corresponding to these nodes to these redundant nodes. The writing process can use the same data structure as the primary storage to ensure that no additional conversion logic is needed for recovery. To avoid multiple replicas of the same node being concentrated in the same physical or logical location, the selection of redundant nodes follows a distribution principle, that is, prioritizing storage nodes corresponding to different racks, different network partitions, or different process instances.

[0110] At the same time, the system records the causal path index information for each behavior node that has completed persistence. This index information does not copy the complete causal relationship structure, but rather saves the node's path index. The local topology information includes the set of identifiers of its direct predecessor nodes and the set of identifiers of its direct successor nodes. These identifiers can be directly obtained from the adjacency table maintained during the structure construction in step two. In subsequent anomaly tracing or binning process replay, the system can start from any archived node and recursively read the index information of adjacent nodes to gradually recover the complete coding behavior path without retracing all historical data. After completing redundant writing and path index recording, the system updates the status of the behavior node to "archived" to distinguish between data still in operation and data that has entered long-term maintenance. This status flag can be used in subsequent steps to limit repeated writing and avoid participating in conflict identification or verification.

[0111] This invention also provides a data maintenance system based on a distributed material packing and coding system, the system comprising:

[0112] The coding acquisition unit is used to acquire a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time marker, and a coding node identifier. Based on the single coding binding record, a reliability metric value is calculated for each coding node.

[0113] The causal relationship construction unit is used to construct a causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier.

[0114] The conflict and anomaly analysis unit is used to identify conflict paths formed by multiple bindings of the same material or the same bag code within a cycle based on the causal relationship structure of the coding behavior, and to calculate the path reliability metric value of all conflict paths; based on the path reliability metric value, the conflict paths are identified as abnormal paths, and the conflict paths with the highest path reliability metric value are selected as retained paths, while the others are identified as abnormal paths; at the same time, the coding binding records corresponding to the retained paths are used as valid candidate binding chains within the current packing cycle;

[0115] The valid path acquisition unit is used to dynamically determine the verification strategy of the binding record in the current packing cycle based on the reserved path, and generate the verification-passed coded binding record as a valid binding record;

[0116] The coding statistics unit is used to perform redundant persistent storage based on the verified coding binding records and the corresponding connection degree in the causal relationship structure of the coding behavior, while recording its path index information.

[0117] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0118] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0119] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0120] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A data maintenance system based on a distributed material packing and coding system, characterized in that, The system includes: The coding acquisition unit is used to acquire a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time marker, and a coding node identifier. Based on the single coding binding record, a reliability metric value is calculated for each coding node. The causal relationship construction unit is used to construct a causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier. The conflict and anomaly analysis unit is used to identify conflict paths formed by multiple bindings of the same material or the same bag code within a cycle based on the causal relationship structure of the coding behavior, and to calculate the path reliability metric value of all conflict paths; based on the path reliability metric value, the conflict paths are identified as abnormal paths, and the conflict paths with the highest path reliability metric value are selected as retained paths, while the others are identified as abnormal paths; at the same time, the coding binding records corresponding to the retained paths are used as valid candidate binding chains within the current packing cycle; The valid path acquisition unit is used to dynamically determine the verification strategy of the binding record in the current packing cycle based on the reserved path, and generate the verification-passed coded binding record as a valid binding record; The coding statistics unit is used to perform redundant persistent storage based on the coding binding records that have passed the verification, according to the corresponding connection degree in the causal relationship structure of the coding behavior, and at the same time record its path index information. Specifically, the construction of the causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier is as follows: The coded binding records are grouped according to the coded node identifier, and a directed association representing the order of chronological sequence is established within each group according to the order of the coded time markers. If the time difference of the coding time marker falls within the preset candidate time range for the same box between coding binding records identified by different coding nodes, a candidate association representing the concurrency relationship is established. By combining the credibility metric, node running stability flag, and time interval, a correlation weight is calculated for each directed association and candidate association.

2. A data maintenance method based on a distributed material packing and coding system, characterized in that, The method includes: Obtain a single coding binding record and map the coding binding record to a coding behavior node. The single coding binding record includes at least a material identifier, a bag code identifier, a coding time stamp, and a coding node identifier. Calculate a reliability metric value for each coding node based on the single coding binding record. Based on the credibility metric and the corresponding coding node identifier, a causal relationship structure for coding behavior is constructed. Based on the causal relationship structure of the coding behavior, conflict paths formed by multiple bindings of the same material or the same bag code within a cycle are identified, and the path reliability metric of all conflict paths is calculated. Based on the path reliability metric, abnormal paths are identified for the conflict paths, and the conflict path with the highest path reliability metric is selected as the retained path, while the others are identified as abnormal paths. At the same time, the coding binding record corresponding to the retained path is used as a valid candidate binding chain within the current packing cycle. Based on the reserved path, the verification strategy for the binding records in the current packing cycle is dynamically determined, and the verification-passed coded binding records are generated as valid binding records. Based on the verified binding record, redundant persistent storage is performed according to the corresponding connection degree in the causal relationship structure of the binding behavior, and its path index information is recorded at the same time. Specifically, the construction of the causal relationship structure for coding behavior based on the credibility metric and the corresponding coding node identifier is as follows: The coded binding records are grouped according to the coded node identifier, and a directed association representing the order of chronological sequence is established within each group according to the order of the coded time markers. If the time difference of the coding time marker falls within the preset candidate time range for the same box between coding binding records identified by different coding nodes, a candidate association representing the concurrency relationship is established. By combining the credibility metric, node running stability flag, and time interval, a correlation weight is calculated for each directed association and candidate association.

3. The data maintenance method for the distributed material packing and coding system according to claim 2, characterized in that, The process of calculating a trustworthiness metric for each coded node based on the single coded binding record is as follows: For the same coding node, the time interval distribution of coding binding records within the sliding time window is analyzed to evaluate time stability. To assess repeatability, the system detects whether identical combinations of material identifiers and bag codes exist within a short period of time. Obtain operational monitoring information of the coded nodes to assess their operational status; Based on the combined evaluation results of time stability, repeatability, and node operating status, the reliability metric is calculated.

4. The data maintenance method for the distributed material packing and coding system according to claim 3, characterized in that, The node operation stability flag corresponds to the dimension of the coded node identifier and is used to characterize the degree of operational stability of the current coded node identifier within the current sliding time range.

5. The data maintenance method for a distributed material packing and coding system according to claim 2, characterized in that, Based on the causal relationship structure of the coding behavior, conflict paths formed by the same material or the same bag code being bound multiple times within a period are identified, specifically: Based on the principle that only one bag code is allowed to correspond to the same material identifier within the same packing cycle, and only one material identifier is allowed to be bound to the same bag code within the same packing cycle, similar nodes are aggregated to filter out the set of nodes with multiple binding candidates. Based on each conflict trigger point in the node set, multiple candidate causal paths are constructed by backtracking along directed associations with the trigger point as the endpoint, thus forming a conflict path. During the backtracking process, links from the same coded node are expanded along the workstation sequence edge, and links across coded nodes are expanded along concurrent candidate edges. During traversal, the most recent record index mapping of the coded node is used, and the latest record number of each coded node is used as the backtracking starting point. Backtracking is only performed along the sequence edge of that node to a preset backtracking depth. This restricts path construction to the recent behavior chain of each node during high-concurrency periods, avoiding unbounded expansion of the entire graph.

6. The data maintenance method for the distributed material packing and coding system according to claim 5, characterized in that, The reliability metric for conflict paths accumulates the weights of each associated path and penalizes frequent jumps across coding nodes in the path. It is calculated and generated based on the number of jumps in the coding node dimension of the conflict path.

7. The data maintenance method for a distributed material packing and coding system according to claim 2, characterized in that, The step of dynamically determining the verification strategy for binding records within the current packing cycle based on the reserved path and generating successfully verified coded binding records is as follows: Obtain the nodes of the abnormal path and the reserved path. The number of failed behavior nodes corresponds to the abnormal path, and the number of behavior nodes participating in packing in the current packing cycle corresponds to the reserved path. The structural anomaly ratio is calculated based on the number of failed behavioral nodes and the total number of behavioral nodes involved in packing during the current packing cycle. Calculate the average of the credibility metrics of all valid nodes in the current period; The verification status index for the current packing cycle is calculated by combining the average of the structural anomaly ratio and the reliability metric. Based on the preset verification strategy mapping table and the interval in which the verification status index is located, the final verification strategy is determined. The strategy includes at least exemption from inspection, random inspection, and full inspection.

8. The data maintenance method for a distributed material packing and coding system according to claim 2, characterized in that, The execution of redundant persistent storage specifically includes: Obtain the number of direct predecessors and direct successors of the valid binding records in the causal relationship structure of the coding behavior; The redundancy factor driven by the structure is calculated based on the sum of the number of direct predecessors and the number of direct successors. The number of redundant replicas is determined based on the redundancy factor, and the binding records are written to the corresponding number of redundant storage nodes; In the step of recording path index information: The identification information of the direct predecessor and direct successor nodes in the causal relationship structure of the coding behavior is saved in the binding record.

9. The data maintenance method for a distributed material packing and coding system according to claim 8, characterized in that, When selecting redundant storage nodes, follow the principle of distribution and give priority to storage nodes located in different physical or logical locations.

Images