A parameter composite deviation anomaly identification method and system

Abstract

The application relates to the technical field of process exception identification and data processing, and discloses a parameter composite deviation exception identification method and system. The method obtains object identification reading data corresponding to a plurality of processing nodes of a target object and node processing parameters corresponding to the processing nodes, determines node object sets corresponding to the processing nodes and time window connection constraints and processing state continuation constraints between adjacent processing nodes; generates candidate mapping relationships according to common object relationships, missing object relationships and new object relationships, generates target mapping relationships through conflict resolution and feasibility screening; generates a processing topology reconstruction result according to the target mapping relationships, identifies abnormal separation, abnormal integration, abnormal backflow or abnormal jump, and outputs corresponding processing results; and the method can improve the accuracy and reliability of multi-processing node flow conversion exception identification.

Description

Technical Field

[0001] This application relates to the field of process anomaly identification and data processing technology, and more specifically, to a method and system for identifying parameter composite deviation anomalies. Background Technology

[0002] In scenarios involving continuous flow across multiple processing nodes, a target object typically needs to pass through multiple processing nodes sequentially. Different processing nodes generate object identification reading records and node processing parameters related to the current processing process. In related technologies, the flow status of the target object in the processing chain is usually determined by collecting object identifications from each processing node, verifying the number of adjacent processing nodes, or checking for anomalies based on the status records of a single processing node.

[0003] However, in practical applications, the transfer of target objects between different processing nodes does not necessarily form explicit and one-to-one corresponding transfer records. For sequentially adjacent processing nodes, the object identifier reading results of the two may simultaneously contain shared objects, missing objects, and newly added objects. At the same time, adjacent processing nodes are also affected by factors such as differences in processing completion time, processing state continuity, and differences in node carrying capacity. If existing technologies rely solely on the reading results of a single node, simple quantity comparison results, or isolated state records for judgment, they can usually only discover local record inconsistencies and find it difficult to establish a stable and closed object correspondence between adjacent processing nodes.

[0004] Furthermore, in the absence of stable object correspondences, existing technologies struggle to connect the scattered local correspondences between adjacent processing nodes into a process chain structure that oriented towards the entire processing chain. Consequently, they cannot accurately recover the true process sequence, process type, and location of the processing nodes involved for the target object across multiple processing nodes. Therefore, when anomalies such as mid-process separation, external incorporation, reverse flow, or cross-node jump occur in the processing chain, existing technologies are prone to confusion between different anomaly types, leading to insufficient accuracy in anomaly identification results. Subsequent response measures such as review, flow restriction, rollback, or suspension of processing lack reliable basis.

[0005] Therefore, how to establish a stable object correspondence between adjacent processing nodes when there is a lack of explicit transfer records and there are simultaneous object missing, addition, timing deviation and state difference between adjacent processing nodes, and on this basis restore the processing topology of the target object among multiple processing nodes, so as to accurately identify abnormal detachment, abnormal merging, abnormal backflow and abnormal jump, has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of existing technologies, this application provides the following technical solution: By reading data from the object identifiers corresponding to the target object at multiple processing nodes and performing node processing parameters, the application determines the node object set, constructs inter-node constraints, generates and filters candidate mapping relationships, reconstructs the processing topology, and identifies anomaly types. This results in anomaly identification of the processing chain during the multi-processing node flow process, thereby achieving accurate identification of anomaly detachment, anomaly incorporation, anomaly backflow, and anomaly jump, and outputting corresponding processing results. In the first aspect, this application discloses a method for identifying parameter composite deviation anomalies, comprising: The system retrieves the object identifier data corresponding to the target object in multiple processing nodes and the node processing parameters corresponding to each processing node. Based on the object identifier data corresponding to each processing node, the system determines the set of node objects corresponding to each processing node. Based on the node processing parameters corresponding to each processing node, the system determines the time window connection constraints and processing state continuation constraints between adjacent processing nodes. For sequentially adjacent processing nodes, candidate mapping relationships between adjacent processing nodes are generated based on the common object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding processing node and the node object sets corresponding to the following processing node. Based on the time window connection constraints, processing state continuation constraints, node carrying capacity constraints, and object conservation constraints corresponding to the candidate mapping relationships, conflict resolution and feasibility screening are performed on the candidate mapping relationships to generate target mapping relationships between adjacent processing nodes; Based on the target mapping relationship between adjacent processing nodes in each time series, the processing topology reconstruction result corresponding to the target object is generated; Based on the processing topology reconstruction results, identify abnormal detachment, abnormal incorporation, abnormal backflow, or abnormal jump, and generate processing chain anomaly identification results; Output the corresponding processing result based on the anomaly identification result of the processing chain.

[0007] Secondly, this application discloses a parameter composite deviation anomaly identification system, comprising: The node preprocessing module is used to obtain the object identifier reading data of the target object corresponding to multiple processing nodes and the node processing parameters corresponding to each processing node, and to determine the node object set corresponding to each processing node based on the object identifier reading data corresponding to each processing node, and to determine the time window connection constraints and processing state continuation constraints between adjacent processing nodes based on the node processing parameters corresponding to each processing node. The candidate generation module is used to generate candidate mapping relationships between adjacent processing nodes based on the common object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding and following processing nodes in time sequence. The mapping filtering module is used to perform conflict resolution and feasibility filtering on candidate mapping relationships based on the time window connection constraints, processing state continuation constraints, node carrying quantity constraints and object conservation constraints corresponding to the candidate mapping relationships, and to generate target mapping relationships between adjacent processing nodes. The topology reconstruction module is used to generate the processing topology reconstruction result corresponding to the target object based on the target mapping relationship between each temporally adjacent processing node; The anomaly detection module is used to identify abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump based on the processing topology reconstruction results, and generate processing chain anomaly identification results. The result output module is used to output the corresponding processing result based on the anomaly identification result of the processing chain.

[0008] Compared with related technologies, this application has the following advantages: This application obtains object identifiers for the target object across multiple processing nodes and node processing parameters for each node. First, it determines the set of node objects for each processing node and establishes time window connection constraints and processing state continuity constraints between adjacent processing nodes. Then, it generates candidate mapping relationships based on shared object relationships, missing object relationships, and newly added object relationships. Combining time window connection constraints, processing state continuity constraints, node carrying capacity constraints, and object conservation constraints, it performs conflict resolution and feasibility screening on the candidate mapping relationships to generate the target mapping relationship. Further, it generates processing topology reconstruction results for the target object based on the target mapping relationship, and identifies abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump based on the processing topology reconstruction results. Finally, it outputs the corresponding processing results based on the abnormal identification results of the processing chain, thus forming a closed loop for abnormal identification and differentiated handling in the multi-processing node flow process.

[0009] This application has at least the following beneficial effects: This application first organizes the object identifier reading data scattered across processing nodes into a set of node objects, and then converts the node processing parameters corresponding to each processing node into time window connection constraints and processing state continuation constraints. This unifies the originally isolated reading and processing information between multiple processing nodes onto a continuously comparable constraint basis, thereby improving the problem in the prior art where node data is fragmented and it is difficult to directly support cross-node flow judgment.

[0010] This application generates candidate mapping relationships for shared object relationships, missing object relationships, and newly added object relationships between temporally adjacent processing nodes. Furthermore, it combines time window connection constraints, processing state continuation constraints, node carrying quantity constraints, and object conservation constraints to perform conflict resolution and feasibility screening on the candidate mapping relationships. This can screen out the target mapping relationship that is more in line with the actual flow from multiple possible object correspondences, thereby reducing ambiguity in object correspondences between adjacent processing nodes and improving the accuracy and stability of determining object transmission relationships between adjacent processing nodes.

[0011] Based on the target mapping relationship between adjacent processing nodes in each time sequence, this application concatenates the continuous target mapping relationship of the same object identifier between multiple processing nodes, and further organizes the object's passing order, object passing type and the positions of the processing nodes involved to generate a processing topology reconstruction result. This restores the originally scattered local correspondence to a passing structure facing the entire processing chain, thereby providing a unified and stable judgment basis for distinguishing and identifying abnormal departure, abnormal merging, abnormal backflow and abnormal jump.

[0012] After obtaining the topology reconstruction results, this application further classifies and identifies different anomaly types, and assigns different subsequent processing methods to anomaly detachment, anomaly incorporation, anomaly backflow, anomaly jump, and normal passage. This allows the identification results to be directly converted into differentiated handling content such as review, flow restriction, rollback review, postponement of subsequent processing, or normal passage, thereby improving the usability of anomaly identification results and enhancing the consistency and reliability of subsequent processing in the entire processing chain. Attached Figure Description

[0013] Figure 1 A schematic flowchart of a parameter composite deviation anomaly identification method provided in this application; Figure 2 The node preprocessing flowchart provided in this application; Figure 3 This application provides a structural block diagram of a parameter composite deviation anomaly identification system. Detailed Implementation

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Example 1

[0015] Please see Figure 1 As shown, this embodiment provides a method for identifying parameter composite deviation anomalies, including the following steps: The purpose of this study is to collect object identification data corresponding to the target object on multiple processing nodes, as well as node processing parameters corresponding to each processing node. This will enable the unified collection of object identification information and processing status information of the target object on multiple processing nodes, and to form the basic node data results and node adjacency relationships required for subsequent determination of node object set, time window connection constraints and processing status continuation constraints.

[0016] In some implementations, the steps of obtaining object identifier read data corresponding to the target object at multiple processing nodes and node processing parameters corresponding to each processing node include: In some implementations, the target object can be a flow object that has been configured with an object identifier and has passed through multiple processing nodes in sequence; for example, the target object can be linens, surgical instruments, instrument packaging units, or turnover containers; wherein, the multiple processing nodes can be collection nodes, processing nodes, transfer nodes, or output nodes corresponding to the above-mentioned flow object processing process; the node processing parameters corresponding to each processing node can include node completion time parameters and status parameters characterizing the processing status of the target object at the current processing node.

[0017] Based on the above implementation scenario, data and node processing parameters can be read through the object identifiers corresponding to each processing node, the set of node objects corresponding to each processing node, the time window connection constraints and processing state continuation constraints between adjacent processing nodes can be determined, and candidate mapping relationships, target mapping relationships, processing topology reconstruction results, processing chain anomaly identification results and corresponding processing results can be generated.

[0018] See Figure 2 As shown, in step 101, the node adjacency relationship corresponding to multiple processing nodes is determined; the pre-configured sequential order of multiple processing nodes is read; based on the sequential order of multiple processing nodes, the node identifier, collection location identifier, effective collection period, and the preceding and following processing nodes that are sequentially adjacent to the current processing node are determined for each processing node, and the node adjacency relationship is generated; the node adjacency relationship is used to subsequently determine the time window connection constraint, processing state continuation constraint, and candidate mapping relationship between adjacent processing nodes.

[0019] Step 102: Obtain object identifier reading data; at the collection location corresponding to each processing node, call the object identifier reading device to identify the target object passing through the current processing node, and obtain the raw object identifier reading result. The raw object identifier reading result includes at least the object identifier, reading time, and node identifier.

[0020] For the original read results of object identifiers within the same processing node, intra-node deduplication and time-series processing are performed. Intra-node deduplication merges duplicate read results for the same object identifier within a preset short-time acquisition interval, while time-series processing sorts the retained read results according to the read time. Based on the results of intra-node deduplication and time-series processing, object identifier read data corresponding to each processing node is generated. This object identifier read data is used in subsequent steps to determine the node object set corresponding to each processing node.

[0021] The method for setting the preset short-term acquisition interval can be determined based on the normal transit time of the target object at the current acquisition position, so that one normal transit process of the same target object corresponds to one valid reading result.

[0022] Step 103: Obtain node processing parameters; at the parameter acquisition location corresponding to each processing node, call the parameter acquisition device to obtain the original results of the node processing parameters of the current processing node; the original results of the node processing parameters include at least the node completion time parameter and the status parameter; the node completion time parameter is used to characterize the processing completion time of the target object at the current processing node, and the status parameter is used to characterize the processing state formed by the target object after passing through the current processing node; based on the node identifier and acquisition time, assign the original results of the node processing parameters to the corresponding processing node, and generate the node processing parameters corresponding to each processing node. The node processing parameters are used in subsequent steps to determine the time window connection constraints and processing state continuation constraints between adjacent processing nodes; among them, the node completion time parameter can take any of the forms of processing end time, processing release time, or processing confirmation time; the status parameter can take any of the forms of numerical parameters, level parameters, or category parameters that can characterize the processing state of the target object.

[0023] Step 104: Generate basic node data results; The object identifier data obtained in Step 102 and the node processing parameters obtained in Step 103 are organized according to the node identifiers to generate basic node data results. The basic node data results include the object identifier data corresponding to each processing node and the node processing parameters corresponding to each processing node; the basic node data results are used in subsequent steps to determine the node object set results, and also used in subsequent steps to determine the constraint results between nodes based on node adjacency relationships.

[0024] The purpose of determining the set of node objects corresponding to each processing node by reading data based on the object identifiers of each processing node is to organize the data read from the object identifiers of each processing node into a set of node objects that can directly participate in the comparison of relationships between adjacent processing node sets, thus providing a unified input for subsequently determining shared object relationships, missing object relationships, and newly added object relationships.

[0025] In some implementations, the steps of determining the set of node objects corresponding to each processing node by reading data based on the object identifier corresponding to each processing node include: Step 201: Read the object identifier reading data corresponding to each processing node; read the object identifier reading data corresponding to each processing node in the basic node data result; each object identifier reading data includes at least the object identifier, node identifier, and reading time; collect the object identifier reading data according to the processing node to provide input for the subsequent extraction of the node object set.

[0026] Step 202: Extract the node object set; Read data for the object identifier corresponding to each processing node, extract all different object identifiers, and retain only one instance of the same object identifier that appears repeatedly, to generate the node object set corresponding to the current processing node; Each object identifier in the node object set is used to represent that the corresponding target object has been effectively read more than once in the current processing node.

[0027] Step 203: Perform a validity check on the set of node objects. For each set of node objects corresponding to a processing node, perform a validity check by combining the valid collection period and the distribution of reading times in the data read by the object identifier corresponding to the current processing node.

[0028] The validity verification process includes: removing object identifiers whose reading time is earlier than the start time of the current processing node's valid data collection period or later than the end time of the current processing node's valid data collection period; and removing object identifiers that only form isolated reads and whose consecutive reading times do not have the characteristic of passing through consecutively.

[0029] After validity verification, the object identifiers corresponding to the actual objects passed by the current processing node are retained, and a set of node objects corresponding to the current processing node is generated.

[0030] Step 204: Generate a set of node objects. Arrange the set of node objects corresponding to each processing node in the order of processing nodes to generate a set of node objects. The set of node objects is used in subsequent steps to determine shared object relationships, missing object relationships, and newly added object relationships by combining the adjacency relationships of nodes, and to generate candidate mapping relationship results.

[0031] In some implementations, to illustrate the process of forming the node object set result, the following example data is used for explanation: For example, the original reading results of the object identifiers corresponding to the first processing node include A1, A1, A2, A3 and A3, and the original reading results of the object identifiers corresponding to the second processing node include A1, A3, A4 and A4.

[0032] After performing intra-node deduplication on the original read results of the object identifiers corresponding to the first processing node, A1, A2, and A3 are retained; after performing intra-node deduplication on the original read results of the object identifiers corresponding to the second processing node, A1, A3, and A4 are retained.

[0033] Furthermore, based on the valid collection time periods corresponding to the first and second processing nodes, the above object identifiers are validated. When the reading times of A1, A2, A3, and A4 are all within the corresponding valid collection time periods and there are no isolated readings, the node object set corresponding to the first processing node is generated as {A1, A2, A3}, and the node object set corresponding to the second processing node is generated as {A1, A3, A4}.

[0034] Based on the node processing parameters corresponding to each processing node, time window connection constraints and processing state continuation constraints between adjacent processing nodes are determined. The purpose is to convert the node processing parameters corresponding to each processing node into node constraint results between adjacent processing nodes, so as to perform conflict resolution and feasibility screening on the candidate mapping relationship results in the future.

[0035] In some implementations, the steps for determining the time window connection constraints and processing state continuation constraints between adjacent processing nodes based on the node processing parameters corresponding to each processing node include: Step 301: Read the node processing parameters corresponding to each processing node; read the node processing parameters corresponding to each processing node in the basic node data results, and read the node adjacency relationship generated in step 101; wherein, the node processing parameters corresponding to each processing node include at least the node completion time parameter and the state parameter; according to the node adjacency relationship, read the node processing parameters corresponding to the preceding and following processing nodes that are temporally adjacent, providing input for the subsequent determination of the constraints between nodes.

[0036] Step 302: Determine the time window connection constraints; For the preceding and following processing nodes that are time-adjacent, read the node completion time parameters corresponding to the preceding and following processing nodes, and determine the time window connection constraints between adjacent processing nodes based on the time sequence relationship between the two and the preset flow time range between nodes.

[0037] The time window connection constraint is used to characterize that after the object corresponding to the previous processing node completes the current processing, the time when the object corresponding to the next processing node enters the subsequent processing should fall within the effective connection time interval corresponding to the preset flow duration range.

[0038] The method for setting the preset flow time range between nodes is determined based on the statistical results of the normal flow time between adjacent processing nodes in the historical normal processing records, or it can be determined based on the shortest and longest flow time corresponding to the process sequence.

[0039] Step 303: Determine the processing state continuation constraint; For the preceding and following processing nodes that are temporally adjacent, read the state parameters corresponding to the preceding and following processing nodes, and based on the pre-established correspondence between the state parameters of adjacent processing nodes, determine whether the state parameters corresponding to the preceding processing node can be continued to the state parameters corresponding to the following processing node, and generate the processing state continuation constraint between adjacent processing nodes.

[0040] The correspondence between adjacent processing node state parameters is used to characterize the sustainable correspondence between the state parameters formed by the previous processing node and the state parameters received or formed by the next processing node.

[0041] The correspondence between state parameters of adjacent processing nodes can be established in the following ways: mapping the state interval corresponding to the state parameter of the previous processing node to the state interval corresponding to the state parameter of the next processing node; or mapping the state category corresponding to the state parameter of the previous processing node to the state category corresponding to the state parameter of the next processing node; or mapping the direction of change of the state parameter of the previous processing node to the direction of change of the state parameter of the next processing node.

[0042] Step 304: Generate inter-node constraint results; organize the time window connection constraints and processing state continuation constraints corresponding to each time-series adjacent processing nodes according to the order of processing nodes to generate inter-node constraint results; the inter-node constraint results are used to perform conflict resolution and feasibility screening on the candidate mapping relationship results and generate the target mapping relationship.

[0043] Furthermore, to illustrate the formation process of the constraints between nodes, the following example data will be used as a case study: For example, if the node completion time parameter corresponding to the first processing node is 10:00 and the node completion time parameter corresponding to the second processing node is 10:18, and the preset flow time between nodes is between 5 minutes and 30 minutes, then the time window connection constraint is satisfied between the first processing node and the second processing node.

[0044] For example, if the state parameter corresponding to the first processing node is state interval S1 and the state parameter corresponding to the second processing node is state interval S2, and the pre-established correspondence between the state parameters of adjacent processing nodes indicates that state interval S1 can continue to state interval S2, then the processing state continuity constraint is satisfied between the first processing node and the second processing node.

[0045] For sequentially adjacent preceding and following processing nodes, candidate mapping relationships are generated based on the shared object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding and following processing nodes. The purpose is to generate candidate mapping relationship results based on the node object set relationships between adjacent processing nodes in the absence of explicit transfer records, so as to provide candidate inputs for determining the target mapping relationship by combining the node constraint results, node carrying quantity constraints, and object conservation constraints.

[0046] In some implementations, the steps for generating candidate mapping relationships between adjacent processing nodes include: Step 401: Read the set of node objects corresponding to adjacent processing nodes; read the generated set of node objects and read the node adjacency relationship generated in step 101; based on the node adjacency relationship, determine the sequentially adjacent preceding and following processing nodes, read the corresponding set of node objects of the preceding and following processing nodes, and use the two as inputs for calculating the adjacent processing node set relationship.

[0047] Step 402: Determine the shared object relationship; read the set of node objects corresponding to the previous processing node and the set of node objects corresponding to the next processing node; read each object identifier in the set of node objects corresponding to the previous processing node item by item, and compare the currently read object identifier with all object identifiers in the set of node objects corresponding to the next processing node; when the currently read object identifier has the same object identifier in the set of node objects corresponding to the next processing node, determine the currently read object identifier as the shared object identifier.

[0048] After comparing all object identifiers in the set of node objects corresponding to the previous processing node, all common object identifiers are matched with their corresponding previous and next processing node identifiers to generate common object relationships. Common object relationships are used to characterize the object correspondence where the same object identifier is read in both the previous and next processing nodes, and are used to generate subsequent continuous mapping candidate results.

[0049] Step 403: Determine the missing object relationship; read the node object set corresponding to the previous processing node and the node object set corresponding to the next processing node; read each object identifier in the node object set corresponding to the previous processing node item by item, and compare the consistency of the currently read object identifier with all object identifiers in the node object set corresponding to the next processing node; when the currently read object identifier does not have the same object identifier in the node object set corresponding to the next processing node, determine the currently read object identifier as the missing object identifier.

[0050] After comparing all object identifiers in the set of node objects corresponding to the previous processing node, all missing object identifiers are matched with their corresponding previous processing node identifiers to generate missing object relationships. Missing object relationships are used to characterize the correspondence between objects that were read in the previous processing node but not in the subsequent processing node, and are used to generate candidate results for detached mapping and candidate results for conflict resolution mapping.

[0051] Step 404: Determine the new object relationship; read the node object set corresponding to the previous processing node and the node object set corresponding to the next processing node; read each object identifier in the node object set corresponding to the next processing node item by item, and compare the consistency of the currently read object identifier with all object identifiers in the node object set corresponding to the previous processing node; when the currently read object identifier does not have the same object identifier in the node object set corresponding to the previous processing node, determine the currently read object identifier as the new object identifier.

[0052] After comparing all object identifiers in the set of node objects corresponding to the next processing node, all newly added object identifiers are matched with their corresponding identifiers of the next processing node to generate new object relationships. The new object relationships are used to represent the correspondence between objects that are read in the next processing node but not in the previous processing node, and are used to generate candidate results for incorporation mapping and candidate results for conflict resolution mapping.

[0053] Step 405: Generate candidate mapping relationships; read the common object relationships generated in step 402, the missing object relationships generated in step 403, and the newly added object relationships generated in step 404; based on the common object relationships, missing object relationships, and newly added object relationships, classify the object transfer possibilities between adjacent processing nodes and generate candidate mapping relationships between adjacent processing nodes.

[0054] Specifically, for each object identifier in the shared object relationship, the identifier of the previous processing node and the identifier of the next processing node corresponding to the object identifier are read. The occurrence relationship of the object identifier in the previous processing node is directly mapped to the occurrence relationship of the object identifier in the next processing node, generating a continuous mapping candidate result from the object identifier corresponding to the previous processing node to the same object identifier corresponding to the next processing node. The continuous mapping candidate result is used to characterize that the same object identifier is continuously passed between adjacent processing nodes.

[0055] For each object identifier in the missing object relationship, read the identifier of the previous processing node corresponding to the object identifier, and match the occurrence relationship of the object identifier in the previous processing node with the missing state in the next processing node to generate a detached mapping candidate result from the object identifier corresponding to the previous processing node to the missing position of the next processing node; the detached mapping candidate result is used to characterize that the current object identifier does not continue to appear in the next processing node after the previous processing node.

[0056] For each object identifier in the newly added object relationship, read the identifier of the next processing node corresponding to the object identifier, and match the vacancy status in the previous processing node with the appearance relationship of the object identifier in the next processing node to generate a candidate result of the merged mapping from the vacancy position of the previous processing node to the corresponding object identifier of the next processing node; the candidate result of the merged mapping is used to characterize the current object identifier appearing in the next processing node but not having a corresponding source in the previous processing node.

[0057] When both missing object relationships and newly added object relationships exist between adjacent processing nodes, the reading time and status parameters of the object identifier corresponding to the missing object relationship in the previous processing node are read item by item, and the reading time and status parameters of the object identifier corresponding to the newly added object relationship in the subsequent processing node are read item by item. When the reading time of the newly added object is later than the reading time of the missing object, and the difference between the two reading times is not greater than the preset initial matching time limit, and the status parameters of the previous processing node corresponding to the missing object and the status parameters of the subsequent processing node corresponding to the newly added object satisfy the pre-established correspondence of adjacent processing node status parameters, the object identifier corresponding to the missing object is initially matched with the object identifier corresponding to the newly added object, and a candidate result for conflict resolution mapping is generated.

[0058] The candidate mapping results, detached mapping results, merged mapping results, and conflict resolution mapping results corresponding to adjacent processing nodes in each time series are arranged in the order of processing nodes to generate candidate mapping relationships. The candidate mapping relationships are used to perform conflict resolution and feasibility screening in combination with the constraint results between nodes, the constraint of the number of nodes to be carried, and the object conservation constraint, and to generate the target mapping relationship.

[0059] Furthermore, to illustrate the formation process of candidate mapping relationships, the following example data will be used as a case study: For example, the set of node objects corresponding to the first processing node is {A1, A2, A3}, and the set of node objects corresponding to the second processing node is {A1, A3, A4}. After comparing the two sets of node objects, it can be determined that the common object relationships include A1 and A3, the missing object relationships include A2, and the newly added object relationships include A4. Further, based on the common object relationships, continuous mapping candidate results corresponding to A1 and A3 are generated respectively; based on the missing object relationships, detached mapping candidate results corresponding to A2 are generated; based on the newly added object relationships, merged mapping candidate results corresponding to A4 are generated. When the reading time of A2 in the first processing node is earlier than the reading time of A4 in the second processing node, and the difference between the two reading times is not greater than the preset initial matching time limit, and the state parameters of A2 in the first processing node and the state parameters of A4 in the second processing node satisfy the correspondence relationship of adjacent processing node state parameters, further conflict resolution mapping candidate results corresponding to A2 and A4 can be generated.

[0060] Based on the time window connection constraints, processing state continuation constraints, node carrying capacity constraints, and object conservation constraints corresponding to the candidate mapping relationships, conflict resolution and feasibility screening are performed on the candidate mapping relationships to generate target mapping relationships between adjacent processing nodes. The purpose is to perform constraint verification, conflict identification, and screening on the generated candidate mapping relationships, eliminate candidate mapping relationships with unreasonable time transmission, unreasonable processing state continuation, unreasonable node carrying capacity allocation, or non-closed object correspondence, and retain target mapping relationships that can represent the real object transmission relationship between adjacent processing nodes, so as to provide stable input for the subsequent generation of processing topology reconstruction results.

[0061] In some implementations, the steps for generating target mapping relationships between adjacent processing nodes include: Step 501: Read the candidate mapping relationships and the constraints between nodes; read the candidate mapping relationships, read the constraints between nodes, and read the node adjacency relationships generated in step 101; based on the node adjacency relationships, determine the preceding and following processing nodes that are temporally adjacent, and extract the candidate mapping relationships, time window connection constraints, and processing state continuation constraints corresponding to the current adjacent processing nodes. The current candidate mapping relationships are used to subsequently determine the node carrying quantity constraints, object conservation constraints, and perform conflict resolution.

[0062] Step 502: Determine the node carrying capacity constraint; for the current adjacent processing node, read the node object set corresponding to the previous processing node and the node object set corresponding to the next processing node, and count the total number of object identifiers in the node object set corresponding to the previous processing node and the total number of object identifiers in the node object set corresponding to the next processing node, respectively, and generate the number of objects in the previous processing node and the number of objects in the next processing node.

[0063] Then read the pre-configured node carrying capacity parameter corresponding to the current adjacent processing node; the node carrying capacity parameter is used to characterize the upper limit of the number of objects that the previous processing node is allowed to transfer out during the current effective collection period, and the upper limit of the number of objects that the next processing node is allowed to receive during the current effective collection period.

[0064] When the node carrying capacity parameter is in the form of a fixed quantity, the upper limit of the number of objects that the previous processing node can transfer out and the upper limit of the number of objects that the next processing node can receive are directly read. When the node carrying capacity parameter is in the form of a processing capacity parameter, the single processing quantity, single processing duration and current effective collection period duration corresponding to the current processing node are read. Based on the single processing quantity, single processing duration and current effective collection period duration, the upper limit of the number of objects that the previous processing node can transfer out and the upper limit of the number of objects that the next processing node can receive are calculated.

[0065] Next, compare the number of objects in the previous processing node with the maximum number of objects that the previous processing node is allowed to transfer out, and take the smaller value as the current maximum number of objects that the previous processing node can transfer out; compare the number of objects in the next processing node with the maximum number of objects that the next processing node is allowed to receive, and take the smaller value as the current maximum number of objects that the next processing node can receive.

[0066] Next, compare the upper limit of the number of objects that can be transferred out by the previous processing node with the upper limit of the number of objects that can be received by the next processing node, and take the smaller value as the upper limit of the number of objects corresponding to the candidate mapping relationship that can be retained between the current adjacent processing nodes; and determine that the number of object identifiers covered by all the candidate mapping relationships that can be retained between the current adjacent processing nodes shall not exceed the upper limit of the number of objects corresponding to the candidate mapping relationships that can be retained between the current adjacent processing nodes.

[0067] Based on the comparison results above, a node capacity constraint is generated for the current adjacent processing nodes. The node capacity constraint is used in the next step to identify candidate conflict mapping relationships formed due to an excessive number of objects, and is also used in step 505 to perform conflict resolution and feasibility screening.

[0068] Step 503: Determine object conservation constraints; for the current adjacent processing node, read the set of node objects corresponding to the previous processing node, the set of node objects corresponding to the next processing node, and the current candidate mapping relationship; wherein, the current candidate mapping relationship includes at least continuous mapping candidate relationship, detached mapping candidate relationship, merged mapping candidate relationship and conflict resolution candidate relationship.

[0069] First, for each object identifier in the set of node objects corresponding to the previous processing node, count the number of candidate mapping relationships in the current candidate mapping relationship that use the current object identifier as the preceding object identifier. When the count result is one, the current object identifier is determined as the unique corresponding object in the forward direction. When the count result is more than one, the current object identifier is determined as the object in the forward direction that is in conflict. When the count result is zero, the current object identifier is determined as the object in the forward direction that is missing.

[0070] For each object identifier in the set of node objects corresponding to the next processing node, count the number of candidate mapping relationships in the current candidate mapping relationship that use the current object identifier as the subsequent object identifier. When the count result is one, the current object identifier is determined as the unique corresponding object in the backward direction. When the count result is more than one, the current object identifier is determined as the conflicting object in the backward direction. When the count result is zero, the current object identifier is determined as the missing object in the backward direction.

[0071] Then, the forward unique correspondence objects, forward conflict objects, forward missing objects, backward unique correspondence objects, backward conflict objects, and backward missing objects are further classified and organized. Among them, the forward unique correspondence objects and backward unique correspondence objects are used to indicate that the object identifier has a unique correspondence in the current candidate mapping relationship, the forward conflict objects and backward conflict objects are used to indicate that the object identifier has a duplicate correspondence in the current candidate mapping relationship, and the forward missing objects and backward missing objects are used to indicate that the object identifier has no corresponding relationship in the current candidate mapping relationship.

[0072] Based on the above classification and organization results, object conservation constraints are determined. These constraints include: each object identifier corresponding to the previous processing node is allowed to form only one continuous mapping candidate relationship or one detached mapping candidate relationship in the current candidate mapping relationship; each object identifier corresponding to the next processing node is allowed to form only one continuous mapping candidate relationship or one merged mapping candidate relationship in the current candidate mapping relationship; the candidate mapping relationships corresponding to forward conflicting objects and backward conflicting objects are both determined as object conservation conflicting relationships; forward missing objects are only allowed to retain detached mapping candidate relationships corresponding to the current object identifier, and backward missing objects are only allowed to retain merged mapping candidate relationships corresponding to the current object identifier; except for the object identifiers corresponding to detached mapping candidate relationships and merged mapping candidate relationships, all other object identifiers should form a unique continuous correspondence between the previous and next processing nodes; the object conservation constraint results are used to identify conflicting candidate mapping relationships in the next step and to perform conflict resolution and feasibility screening in step 505.

[0073] Step 504: Perform conflict identification on the current candidate mapping relationship; read the current candidate mapping relationship extracted in step 501, and perform conflict verification on each of the current candidate mapping relationships.

[0074] For each candidate mapping relationship in the current candidate mapping relationship, read the object identifier of the previous processing node, the object identifier of the next processing node, the object reading time before and after, and the state parameters before and after. Compare the difference in object reading time before and after with the effective connection time interval corresponding to the time window connection constraint, compare the state parameters before and after with the processing state continuation constraint, and compare the object identifier occupancy status corresponding to the current candidate mapping relationship with the node carrying quantity constraint and the object conservation constraint respectively.

[0075] The current candidate mapping relationship is determined as a conflict candidate mapping relationship when any of the following conditions are met: the difference between the reading time of the preceding and following objects exceeds the effective connection time interval; the preceding and following state parameters do not meet the processing state continuation constraint; multiple candidate mapping relationships correspond to the same preceding processing node object identifier; multiple candidate mapping relationships correspond to the same following processing node object identifier; or the addition of the current candidate mapping relationship results in the number of objects corresponding to the retained candidate mapping relationship being higher than the number of objects allowed by the node carrying capacity constraint.

[0076] Candidate mappings that do not match the above conditions are identified as non-conflicting candidate mappings.

[0077] The conflict candidate mapping relationships and non-conflict candidate mapping relationships are organized according to the object identifier and the processing node position to generate conflict identification results; the conflict identification results include at least conflict candidate mapping relationships and non-conflict candidate mapping relationships; the conflict identification results are used to perform conflict resolution and feasibility screening in the next step.

[0078] Step 505: Perform conflict resolution and feasibility screening on the conflict identification results; read the conflict identification results and perform conflict resolution and feasibility screening around the conflict candidate mapping relationship and non-conflict candidate mapping relationship in the conflict identification results.

[0079] First, the non-conflict candidate mapping relationships in the conflict identification results are directly retained to form the initial retained mapping relationships; then, the conflict candidate mapping relationships in the conflict identification results are resolved item by item according to the object identifier.

[0080] When the same previous processing node object identifier corresponds to multiple conflict candidate mapping relationships, read the difference in reading time between the previous and next objects and the comparison results of the previous and next state parameters corresponding to the multiple conflict candidate mapping relationships respectively, retain the conflict candidate mapping relationship with the smallest difference in reading time between the previous and next objects and the previous and next state parameters satisfy the processing state continuation constraint, and delete the remaining conflict candidate mapping relationships; when the same subsequent processing node object identifier corresponds to multiple conflict candidate mapping relationships, retain one conflict candidate mapping relationship and delete the remaining conflict candidate mapping relationships in the same way.

[0081] After resolving object identifier-level conflicts, the retained candidate mapping relationships are merged into the initial retained mapping relationships to form the mapping relationships to be screened. Then, the node carrying capacity constraint is checked on the mapping relationships to be screened. When the number of object identifiers corresponding to the mapping relationship to be screened is higher than the upper limit of the number of objects allowed by the node carrying capacity constraint, the candidate mapping relationships are retained in order of increasing difference between the time of object reading before and after, until the number of object identifiers corresponding to the retained candidate mapping relationships is no longer higher than the upper limit of the number of objects allowed by the node carrying capacity constraint.

[0082] Next, object conservation constraint verification is performed on the retained candidate mapping relationships. Candidate mapping relationships that still result in duplicate correspondences of object identifiers of the previous processing node, duplicate correspondences of object identifiers of the subsequent processing node, or non-closed correspondences between objects are deleted, generating feasible mapping screening results. The feasible mapping screening results are used to generate the target mapping relationship in step 506.

[0083] Step 506: Generate target mapping relationship; determine the retained mapping relationship in the feasible mapping screening results corresponding to the current adjacent processing node as the target mapping relationship between the current adjacent processing nodes; then organize the target mapping relationships corresponding to each temporal adjacent processing node according to the order of processing nodes to generate target mapping relationship; the target mapping relationship is used to generate the processing topology reconstruction result in subsequent steps.

[0084] Furthermore, to illustrate the formation process of the target mapping relationship, the following example data will be used as a case study: For example, between the first processing node and the second processing node, continuous mapping candidate results corresponding to A1 and A3, detached mapping candidate results corresponding to A2, merged mapping candidate results corresponding to A4, and conflict-resolving mapping candidate results corresponding to A2 and A4 have been generated. Further, time window connection constraint verification, processing state continuation constraint verification, node carrying quantity constraint verification, and object conservation constraint verification are performed on the above candidate mapping relationships. If the conflict-resolving mapping candidate results corresponding to A2 and A4 do not meet the time window connection constraint, or if retaining the conflict-resolving mapping candidate results and the detached mapping candidate results corresponding to A2 at the same time would cause the object conservation constraint to be violated, then the conflict-resolving mapping candidate results corresponding to A2 and A4 are deleted. The continuous mapping candidate results corresponding to A1 and A3, the detached mapping candidate results corresponding to A2, and the merged mapping candidate results corresponding to A4 are retained to generate the target mapping relationship between the first processing node and the second processing node.

[0085] Based on the target mapping relationship between adjacent processing nodes in each time series, the processing topology reconstruction result corresponding to the target object is generated. The purpose is to connect and integrate the already determined target mapping relationship between adjacent processing nodes according to the order of the processing nodes, restore the actual passage structure of the target object between multiple processing nodes, and form the processing topology reconstruction result required for subsequent identification of abnormal departure, abnormal incorporation, abnormal backflow and abnormal jump.

[0086] In some implementations, the steps for generating the processing topology reconstruction result corresponding to the target object include: Step 601: Read the target mapping relationship and node adjacency relationship.

[0087] Read the target mapping relationship and the node adjacency relationship generated in step 101; based on the node adjacency relationship, extract the target mapping relationship between each temporally adjacent processing node in the order of processing nodes; the target mapping relationship is used to subsequently determine the continuous traversal relationship of the target object between multiple processing nodes.

[0088] Step 602: Determine the object traversal order; for each object identifier, read its target mapping relationship at each processing node; according to the order of processing nodes, concatenate the continuous target mapping relationships of the same object identifier between multiple processing nodes to generate the object traversal order corresponding to the current object identifier; the object traversal order is used to characterize the continuous traversal path of the current object identifier from the previous processing node to the next processing node, and is used to determine the object traversal type in the future.

[0089] Step 603: Determine the object traversal type; for each object identifier, based on the object traversal order, and combining the node adjacency relationship and the target mapping relationship, identify the object traversal type corresponding to the current object identifier.

[0090] When there is a target mapping relationship between the current object identifier and the preceding and following processing nodes in the temporal sequence, the corresponding continuous traversal type of the current object identifier between the adjacent processing nodes is determined. When the current object identifier appears in a previous processing node, but there is no corresponding target mapping relationship in the next processing node indicated by the node adjacency relationship, and the object passing sequence is interrupted at the previous processing node, it is determined that the current object identifier corresponds to the detachment passing type. When the current object identifier appears in a subsequent processing node, but there is no corresponding target mapping relationship in the preceding processing node indicated by the node adjacency relationship, and the object passes through the order for the first time at the subsequent processing node, it is determined that the current object identifier corresponds to the passing type. When the current object identifier has a node position change in the order in which the subsequent processing node returns to the previous processing node, the return flow type corresponding to the current object identifier is determined. When the current object identifier's traversal sequence contains a process node that directly enters a non-adjacent subsequent process node, and the intermediate processing node between the two does not appear in the current object identifier's traversal sequence, the jump traversal type corresponding to the current object identifier is determined.

[0091] Through the above processing, the object traversal order and object traversal type corresponding to each object identifier are obtained.

[0092] Step 604: Perform topology reorganization on the object traversal order and object traversal type to generate the processed topology reconstruction result; read the object traversal order corresponding to each object identifier generated in step 602, and the object traversal type corresponding to each object identifier determined in step 603; for each object identifier, read the position of each processing node in the object traversal order corresponding to the object identifier, and read the target mapping relationship between each adjacent processing node corresponding to the object identifier, and then combine the object traversal order, object traversal type, position of each processing node and target mapping relationship between each adjacent processing node corresponding to the object identifier to form the topology reorganization content corresponding to the object identifier.

[0093] Next, the topology organization content corresponding to each object identifier is summarized and organized according to the object identifier to generate the processing topology reconstruction result. The processing topology reconstruction result includes at least the object traversal order, object traversal type, the location of the processing nodes involved, and the target mapping relationship between adjacent processing nodes for each object identifier. The processing topology reconstruction result is used to read in step 701 and to identify abnormal detachment, abnormal merging, abnormal backflow, and abnormal jump in steps 702 to 705.

[0094] Furthermore, to illustrate the process of forming the topology reconstruction results, the following example data will be used as a case study: For example, if A1 has a continuous target mapping relationship between the first, second, and third processing nodes, but does not enter the fourth processing node, then the object traversal order corresponding to A1 can be organized as "first processing node - second processing node - third processing node"; if A4 has a target mapping relationship between the second and third processing nodes, and has no corresponding source in the first processing node, then the object traversal order corresponding to A4 can be organized as "second processing node - third processing node"; if A2 only appears in the first processing node, then the object traversal order corresponding to A2 can be organized as "first processing node"; furthermore, based on the object traversal order corresponding to A1, A2, and A4 respectively, it can be determined that A1 corresponds to a continuous traversal type, A2 corresponds to a detached traversal type, and A4 corresponds to a merged traversal type; then, the object traversal order, object traversal type, the positions of the processing nodes involved, and the target mapping relationship between adjacent processing nodes corresponding to A1, A2, and A4 are organized to generate the processing topology reconstruction result.

[0095] Based on the processing topology reconstruction results, abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump are identified, and processing chain anomaly identification results are generated. The purpose is to classify and identify the abnormal passage of the target object between multiple processing nodes based on the processing topology reconstruction results, and obtain processing chain anomaly identification results that can be directly used to output the corresponding processing results in the subsequent process.

[0096] In some implementations, the steps for generating the processing chain anomaly identification result include: Step 701: Read the processed topology reconstruction results and extract the object traversal order, object traversal type, and processing node positions corresponding to each object identifier; the processed topology reconstruction results are used for subsequent judgment based on the identification type.

[0097] Step 702, identify abnormal detachment; for object identifiers whose object traversal type includes detachment traversal type, read the object traversal order corresponding to the current object identifier, and determine whether the current object identifier is interrupted after a certain previous processing node and before a predetermined next processing node; when the current object identifier is identified in the previous processing node, and there is no corresponding target mapping relationship in the next processing node indicated by the node adjacency relationship, and there is no continuous traversal relationship directly connecting the current previous processing node to the subsequent processing node in the processing topology reconstruction result, it is determined that the current object identifier corresponds to abnormal detachment, and an abnormal detachment identification result is generated.

[0098] Step 703, identify abnormal merging; for objects whose traversal type contains the object identifier of the merging traversal type, read the object traversal order corresponding to the current object identifier, and determine whether the current object identifier appears in a certain subsequent processing node but lacks the corresponding source of the preceding processing node.

[0099] When the current object identifier is identified in the next processing node, and there is no corresponding target mapping relationship in the previous processing node indicated by the node adjacency relationship, and there is no continuous traversal relationship from the previous processing node to the current next processing node in the processing topology reconstruction result, it is determined that the current object identifier corresponds to an abnormal merging, and an abnormal merging identification result is generated.

[0100] Step 704: Identify abnormal backflow; For object identifiers whose object traversal type contains backflow traversal type, read the object traversal order corresponding to the current object identifier, and determine whether the current object identifier has a situation where it is returned from the subsequent processing node to the previous processing node in the order of processing nodes.

[0101] When the node position of the object corresponding to the current object identifier changes after passing through a subsequent processing node and then re-entering the previous processing node, the abnormal backflow corresponding to the current object identifier is determined, and an abnormal backflow identification result is generated.

[0102] Step 705: Identify abnormal jumps; for object identifiers whose object traversal type contains jump traversal type, read the object traversal order corresponding to the current object identifier, and read the node adjacency relationship generated in step 101; based on the node adjacency relationship, determine all intermediate processing nodes between the previous and next processing nodes corresponding to the current object identifier; then determine whether all intermediate processing nodes appear in the object traversal order corresponding to the current object identifier.

[0103] When at least one of the intermediate processing nodes does not appear in the order in which the object corresponding to the current object identifier is processed, and the current object identifier appears directly in the last processing node, an abnormal jump is determined to be corresponding to the current object identifier, and an abnormal jump identification result is generated.

[0104] Step 706: Summarize and organize the abnormal departure identification results, abnormal merging identification results, abnormal backflow identification results, abnormal jump identification results, and normal pass identification results to generate the processing chain abnormal identification results.

[0105] Read the abnormal detachment identification results generated in step 702, the abnormal merging identification results generated in step 703, the abnormal backflow identification results generated in step 704, and the abnormal jump identification results generated in step 705, and read the normal passage identification results corresponding to the object identifiers in the processed topology reconstruction results whose object passage type is continuous passage type and which did not hit abnormal detachment, abnormal merging, abnormal backflow, and abnormal jump.

[0106] For each object identifier, determine whether the current object identifier appears in the abnormal detachment identification result, abnormal merging identification result, abnormal backflow identification result, abnormal jump identification result, or normal pass identification result, and organize the identification type corresponding to the current object identifier with the relevant processing node positions; then summarize and organize the identification types and relevant processing node positions corresponding to all object identifiers according to object identifiers to generate processing chain abnormal identification results; the processing chain abnormal identification results include at least the identification type and relevant processing node positions corresponding to each object identifier; the processing chain abnormal identification results are used for reading in step 801 and for generating object processing results in step 803.

[0107] The corresponding processing result is output based on the anomaly identification result of the processing chain; the purpose is to convert the anomaly identification result of the processing chain into an executable subsequent processing conclusion, so that the identification result can directly affect the subsequent flow control or review processing of the target object.

[0108] In some implementations, the steps of outputting the corresponding processing result based on the anomaly identification result of the processing chain include: Step 801: Read the processing chain anomaly identification results and extract the identification type and the location of the processing node corresponding to each object identifier; the processing chain anomaly identification results are used to determine the processing method corresponding to the current object identifier.

[0109] Step 802: Determine the processing method corresponding to the recognition type; read the pre-configured recognition type processing correspondence; the recognition type processing correspondence is used to represent the processing methods corresponding to different recognition types; the recognition type processing correspondence can be established in the following way: The process of separating an exception is assigned to a review process; merging an exception is assigned to a flow restriction process; returning an exception is assigned to a rollback review process; jumping an exception is assigned to a postponement of subsequent processing; and passing an exception is assigned to a normal processing process. Through the above processes, a correspondence between the identification type and the processing method is formed.

[0110] Step 803: Perform corresponding processing on the abnormal identification results and identification type processing correspondence of the processing chain to generate object processing results; read the identification type and the location of the processing node involved for each object identifier extracted in step 801, and read the identification type processing correspondence determined in step 802; for each object identifier, according to the identification type corresponding to the current object identifier, find the processing method corresponding to the identification type in the identification type processing correspondence, and then organize the current object identifier, the processing method corresponding to the current object identifier and the location of the processing node involved to generate the object processing result corresponding to the current object identifier.

[0111] Next, the processing results corresponding to all object identifiers are summarized and organized according to object identifiers to generate object processing results. The object processing results should include at least the processing method and the location of the processing nodes involved for each object identifier. The object processing results are used to output the corresponding processing results in the next step.

[0112] Step 804: Execute terminal output on the object processing result to generate the corresponding processing result; specifically, read the object processing result; for each object identifier in the object processing result, determine the corresponding output terminal and output content according to the processing method corresponding to the current object identifier.

[0113] When the processing method is review processing, the review prompt content corresponding to the current object identifier is sent to the review terminal; when the processing method is flow restriction processing, the flow restriction content corresponding to the current object identifier is sent to the flow control terminal; when the processing method is rollback review processing, the rollback processing content corresponding to the current object identifier is sent to the rollback processing terminal; when the processing method is postponement of subsequent processing, the postponement processing content corresponding to the current object identifier is sent to the subsequent processing control terminal; when the processing method is normal pass processing, the normal pass content corresponding to the current object identifier is sent to the normal flow terminal.

[0114] Then, the output content and output terminals corresponding to all object identifiers are summarized and organized according to the object identifiers to generate corresponding processing results; the corresponding processing results are the final output results of this implementation method.

[0115] Furthermore, to illustrate the formation process of the anomaly identification results and corresponding processing results in the processing chain, the following example data is used for explanation: For example, based on the processing topology reconstruction results, A2 corresponds to the detachment type, and there is no continuous traversal relationship from the first processing node directly to the subsequent processing node in the processing topology reconstruction results, so A2 is identified as an abnormal detachment; A4 corresponds to the merging type, and there is no continuous traversal relationship from the preceding processing node to the second processing node in the processing topology reconstruction results, so A4 is identified as an abnormal merging; A1 corresponds to the continuous traversal type and does not hit any abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump, so A1 is identified as normal passage; furthermore, the abnormal detachment identification result corresponding to A2 is assigned to the review processing method, the abnormal merging identification result corresponding to A4 is assigned to the flow restriction processing method, and the normal passage identification result corresponding to A1 is assigned to the normal passage processing method; then, based on the processing methods corresponding to A2, A4, and A1 respectively, the output terminal and output content are determined, and the corresponding processing results are generated; among them, the review prompt content corresponding to A2 is sent to the review terminal, the flow restriction content corresponding to A4 is sent to the flow control terminal, and the normal passage content corresponding to A1 is sent to the normal flow terminal. Example 2

[0116] See Figure 3 As shown, this embodiment provides a parameter composite deviation anomaly identification system. Since this system uses a parameter composite deviation anomaly identification method from Embodiment 1, it has the same effect, which will not be repeated here. The system includes: The node preprocessing module is used to obtain the object identifier reading data of the target object corresponding to multiple processing nodes and the node processing parameters corresponding to each processing node, and to determine the node object set corresponding to each processing node based on the object identifier reading data corresponding to each processing node, and to determine the time window connection constraints and processing state continuation constraints between adjacent processing nodes based on the node processing parameters corresponding to each processing node. The candidate generation module is used to generate candidate mapping relationships between adjacent processing nodes based on the common object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding and following processing nodes in time sequence. The mapping filtering module is used to perform conflict resolution and feasibility filtering on candidate mapping relationships based on the time window connection constraints, processing state continuation constraints, node carrying quantity constraints and object conservation constraints corresponding to the candidate mapping relationships, and to generate target mapping relationships between adjacent processing nodes. The topology reconstruction module is used to generate the processing topology reconstruction result corresponding to the target object based on the target mapping relationship between each temporally adjacent processing node; The anomaly detection module is used to identify abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump based on the processing topology reconstruction results, and generate processing chain anomaly identification results. The result output module is used to output the corresponding processing result based on the anomaly identification result of the processing chain.

[0117] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope defined in the claims.

Claims

1. A method for identifying anomalies due to composite parameter deviations, characterized in that, include: The system retrieves the object identifier data corresponding to the target object in multiple processing nodes and the node processing parameters corresponding to each processing node. Based on the object identifier data corresponding to each processing node, the system determines the set of node objects corresponding to each processing node. Based on the node processing parameters corresponding to each processing node, the system determines the time window connection constraints and processing state continuation constraints between adjacent processing nodes. For sequentially adjacent processing nodes, candidate mapping relationships between adjacent processing nodes are generated based on the common object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding processing node and the node object sets corresponding to the following processing node. Based on the time window connection constraints, processing state continuation constraints, node carrying capacity constraints, and object conservation constraints corresponding to the candidate mapping relationships, conflict resolution and feasibility screening are performed on the candidate mapping relationships to generate target mapping relationships between adjacent processing nodes; Based on the target mapping relationship between adjacent processing nodes in each time series, the processing topology reconstruction result corresponding to the target object is generated; Based on the processing topology reconstruction results, identify abnormal detachment, abnormal incorporation, abnormal backflow, or abnormal jump, and generate processing chain anomaly identification results; Output the corresponding processing result based on the anomaly identification result of the processing chain.

2. The method for identifying parameter composite deviation anomalies according to claim 1, characterized in that, Methods for determining the set of node objects corresponding to each processing node include: Read the data by reading the object identifier corresponding to each processing node; Read the data of the object identifiers corresponding to each processing node to extract different object identifiers and obtain the set of node objects corresponding to each processing node. Based on the distribution of valid collection time periods and reading times corresponding to each processing node, perform validity checks on the set of node objects corresponding to each processing node. The node object sets corresponding to each processing node that has passed the validity verification are arranged in the order of the processing nodes to obtain the node object set result.

3. The method for identifying parameter composite deviation anomalies according to claim 1, characterized in that, Methods for determining time window continuity constraints and processing state continuation constraints between adjacent processing nodes include: Read the node processing parameters corresponding to each processing node, and extract the node processing parameters corresponding to the preceding and following processing nodes that are temporally adjacent by combining the node adjacency relationship. Based on the node completion time parameters corresponding to the previous processing node, the node completion time parameters corresponding to the next processing node, and the preset flow time range between nodes, the time window connection constraints are determined. Based on the correspondence between the state parameters of the previous processing node, the state parameters of the next processing node, and the state parameters of adjacent processing nodes, the processing state continuity constraints are determined.

4. The method for identifying parameter composite deviation anomalies according to claim 3, characterized in that, Methods for generating candidate mapping relationships between adjacent processing nodes include: Read the collection of node objects corresponding to the previous processing node and the collection of node objects corresponding to the next processing node; Compare the set of node objects corresponding to the previous processing node with the set of node objects corresponding to the next processing node to determine the common object relationships, missing object relationships, and newly added object relationships; Based on shared object relationships, missing object relationships, and newly added object relationships, generate continuous mapping candidate results, detached mapping candidate results, and merged mapping candidate results, and generate conflict-resolved mapping candidate results when both missing object relationships and newly added object relationships exist simultaneously. The candidate results of continuous mapping, detached mapping, merged mapping, and conflict-resolved mapping are organized to generate candidate mapping relationships.

5. The method for identifying parameter composite deviation anomalies according to claim 4, characterized in that, When both missing object relationships and newly added object relationships exist, the methods for generating candidate results for conflict resolution mappings include: Read the reading time and status parameters of the missing object relationship corresponding to the object identifier in the previous processing node, and read the reading time and status parameters of the newly added object relationship corresponding to the object identifier in the next processing node; When the reading time of the newly added object is later than the reading time of the missing object, and the difference between the two reading times is not greater than the preset initial matching time limit, and the state parameters of the previous processing node corresponding to the missing object and the state parameters of the next processing node corresponding to the newly added object satisfy the correspondence relationship of adjacent processing node state parameters, the object identifier corresponding to the missing object is initially matched with the object identifier corresponding to the newly added object to generate a candidate result for conflict resolution mapping.

6. The method for identifying parameter composite deviation anomalies according to claim 3, characterized in that, Methods for generating target mapping relationships between adjacent processing nodes include: Read the candidate mapping relationship; based on the set of node objects corresponding to the previous processing node and the set of node objects corresponding to the next processing node, determine the node carrying quantity constraint and object conservation constraint; Based on time window connection constraints, processing state continuity constraints, node carrying capacity constraints, and object conservation constraints, conflict identification is performed on candidate mapping relationships to obtain conflict candidate mapping relationships and non-conflict candidate mapping relationships. Conflict resolution and feasibility screening are performed on the conflict candidate mappings, and the target mapping is generated by combining the non-conflict candidate mappings.

7. The method for identifying parameter composite deviation anomalies according to claim 1, characterized in that, The methods for generating the topology reconstruction results corresponding to the target object include: Read the target mapping relationship between adjacent processing nodes in each time sequence; determine the order in which objects corresponding to each object identifier pass through based on the target mapping relationship; Determine the object pass type corresponding to each object identifier based on the object pass order; Based on the object identification, the order of objects passed through, the type of objects passed through, and the location of the processing nodes involved, the results of the processing topology reconstruction are generated.

8. A method for identifying composite parameter deviation anomalies according to claim 1 or 7, characterized in that, Methods for generating processing chain anomaly identification results include: Read the order of objects, the type of objects, and the positions of the processing nodes involved in the topology reconstruction results; Based on the object's passage order, object passage type, and the location of the processing nodes involved, abnormal detachment, abnormal merging, abnormal backflow, and abnormal jump are identified. For objects whose passage type is continuous and are not identified as abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump, the normal passage identification result is determined. The results of abnormal separation identification, abnormal incorporation identification, abnormal backflow identification, abnormal jump identification, and normal pass identification are summarized and organized to generate the abnormal identification results of the processing chain.

9. The method for identifying parameter composite deviation anomalies according to claim 6, characterized in that, Methods for performing conflict resolution and feasibility screening on candidate conflict mappings include: When multiple conflict candidate mappings correspond to the same previous processing node object identifier, or multiple conflict candidate mappings correspond to the same subsequent processing node object identifier, the conflict candidate mapping with the smallest difference between the reading times of the preceding and following objects and whose state parameters satisfy the processing state continuation constraint is retained. After object identifier-level conflict resolution is completed, the node carrying quantity constraint verification is performed on the retained conflict candidate mappings and non-conflict candidate mappings, and the candidate mappings are retained in order of increasing difference between the reading times of the preceding and following objects to generate the target mapping.

10. A parameter composite deviation anomaly identification system, used to implement the parameter composite deviation anomaly identification method according to any one of claims 1-9, characterized in that, The system includes: The node preprocessing module is used to obtain the object identifier reading data of the target object corresponding to multiple processing nodes and the node processing parameters corresponding to each processing node, and to determine the node object set corresponding to each processing node based on the object identifier reading data corresponding to each processing node, and to determine the time window connection constraints and processing state continuation constraints between adjacent processing nodes based on the node processing parameters corresponding to each processing node. The candidate generation module is used to generate candidate mapping relationships between adjacent processing nodes based on the common object relationships, missing object relationships, and newly added object relationships between the node object sets corresponding to the preceding and following processing nodes in time sequence. The mapping filtering module is used to perform conflict resolution and feasibility filtering on candidate mapping relationships based on the time window connection constraints, processing state continuation constraints, node carrying quantity constraints and object conservation constraints corresponding to the candidate mapping relationships, and to generate target mapping relationships between adjacent processing nodes. The topology reconstruction module is used to generate the processing topology reconstruction result corresponding to the target object based on the target mapping relationship between each temporally adjacent processing node; The anomaly detection module is used to identify abnormal detachment, abnormal merging, abnormal backflow, or abnormal jump based on the processing topology reconstruction results, and generate processing chain anomaly identification results. The result output module is used to output the corresponding processing result based on the anomaly identification result of the processing chain.

Images