IDC machine room operation and maintenance management system and method based on knowledge graph

Abstract

The application discloses an IDC machine room operation and maintenance management system and method based on a knowledge graph, and comprises the following steps: step one, collecting IDC machine room operation and maintenance related data, and constructing an IDC operation and maintenance knowledge graph; step two, extracting an influence closure subgraph; step three, generating an operation and maintenance closure unit and scheduling parameters; step four, generating an initial scheduling sequence; step five, calculating a graph risk release efficiency; step six, forming an improved Moore-Hodgson scheduling rule and generating an updated scheduling sequence; step seven, including a target operation and maintenance closure unit into a delay pool and generating a back-rolling feasible path; and step eight, re-including the updated scheduling sequence and outputting an operation and maintenance scheduling result. The application improves the correlation analysis capability, time limit control capability and dynamic recovery capability of IDC machine room operation and maintenance scheduling.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance management technology, and in particular to an IDC data center operation and maintenance management system and method based on knowledge graph. Background Technology

[0002] With the continuous expansion of data center infrastructure and the increasing demands for continuous operation of business systems, intelligent operation and maintenance management technologies targeting the relationships between equipment, links, alarms, work orders, and business processes within IDC data centers have received widespread attention. Existing IDC data center operation and maintenance platforms primarily rely on static asset ledgers, manual experience-based rules, or simple priority ranking methods for orchestrating operation and maintenance tasks. However, these methods commonly suffer from the following problems in practical applications: In IDC data centers, the relationships between device connections, service carrying, and state changes are complex and intertwined. Existing technologies often struggle to accurately depict the impact scope of objects to be processed from a global perspective, leading to fragmented granularity of operation and maintenance tasks and insufficient identification of linked risks. Faced with deadline constraints and resource usage conflicts, traditional scheduling methods typically rely solely on processing time, manual priority, or fixed rules to select tasks, making it difficult to balance risk propagation suppression and time limit control. This can easily result in critical operation and maintenance units being mistakenly removed or scheduling results deviating from the actual risk structure. In scenarios where the operation and maintenance status is constantly changing, existing methods mostly rely on manual re-scheduling or overall recalculation for delayed tasks, lacking a feasible back-insertion mechanism based on related state changes. This results in poor continuity of the scheduling process, low response efficiency, and insufficient stability of the results.

[0003] Therefore, how to provide an IDC data center operation and maintenance management system and method based on knowledge graphs is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] One objective of this invention is to propose an IDC (Internet Data Center) data center operation and maintenance management system and method based on knowledge graphs. This invention combines knowledge graph modeling, closure extraction, compression mapping, and improved Moore-Hodgson scheduling rules to describe in detail the association analysis, risk release scheduling, and rollback and insertion processing of IDC data center operation and maintenance objects. It has the advantages of accurate association characterization, high scheduling efficiency, and stable operation and maintenance results.

[0005] According to an embodiment of the present invention, an IDC data center operation and maintenance management method based on knowledge graphs includes the following steps: Step 1: Collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; Step 2: Extract the influence closure subgraph from the IDC operations and maintenance knowledge graph around the object to be processed; Step 3: Perform compression mapping on the influencing closure subgraph to generate operation and maintenance closure units and form the scheduling parameters corresponding to the operation and maintenance closure units; Step 4: Based on the deadline in the scheduling parameters, arrange the execution order of the operation and maintenance closure units to generate the initial schedule; Step 5: According to the initial scheduling sequence, the operation and maintenance closure units are sequentially included into the current scheduling set, and the cumulative occupied time is updated. When the cumulative occupied time exceeds the deadline, the graph risk release efficiency corresponding to the operation and maintenance closure units in the current scheduling set is calculated. Step 6: Retain the rule of inclusion of deadline time in order, replace the rule of removal of longest processing volume with the rule of priority removal of graph risk release efficiency, form an improved Moore-Hodgson scheduling rule, and remove the target operation and maintenance closure unit from the current scheduling set according to the improved Moore-Hodgson scheduling rule to generate an updated scheduling sequence; Step 7: Add the target operation and maintenance closure unit to the delay pool, and generate the corresponding rollback feasible path for the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph; Step 8: Re-incorporate the target operation and maintenance closure unit that meets the back-insertion condition in the delay pool into the update schedule along the rollback feasible path, and output the operation and maintenance scheduling result.

[0006] Optionally, step one specifically includes: Collect equipment records, connection records, alarm records, work order records, service records, status records, and maintenance records in the IDC data center to form IDC data center operation and maintenance related data; The operation and maintenance related data of the IDC data center are aligned in time, unified in identification and related in relation to form an operation and maintenance related record set. Based on the operation and maintenance related record set, the entity node set, relationship edge set, node attribute set and edge attribute set are determined. Perform graph organization and relation mapping on the entity node set, relation edge set, node attribute set, and edge attribute set to generate an IDC operation and maintenance knowledge graph.

[0007] Optionally, step two specifically includes: Locate the target entity node corresponding to the operation and maintenance object to be processed in the IDC operation and maintenance knowledge graph, and extract the relationship edges and entity nodes connected to the target entity node to form a target association record set; Based on the direction, type, and state of the relation edges in the target associated record set, perform an influence propagation search along the relation edges on the target entity nodes, extract entity nodes and relation edges that meet the influence propagation conditions, and form an influence propagation record set; Perform node merging, boundary closure, and graph structure extraction on the influence propagation record set to generate an influence closure subgraph.

[0008] Optionally, step three specifically includes: Extract entity nodes, relation edges, and target entity nodes from the influence closure subgraph. Perform connectivity traversal along the connection paths between entity nodes and the association paths between relation edges to determine the range of entity nodes and the range of relation edges that maintain influence transition connections with the target entity node, forming a closure range record set. The entity nodes in the closure range record set are aggregated and organized according to their connection relationships to form an entity node set. The relation edges in the closure range record set are merged and organized according to their association relationships to form a relation edge set. The entity node set and the relation edge set are then mapped and associated to form a closure mapping record set. Perform node shrinking on the entity node set in the closure mapping record set, perform edge relationship merging on the relation edge set in the closure mapping record set, compress the closure mapping record set into a single graph structure unit, and generate the operation and maintenance closure unit. The time constraint information in the operation and maintenance closure unit is extracted to meet the deadline boundary, and the occupancy association information in the operation and maintenance closure unit is accumulated to form scheduling parameters, which include the deadline time and the occupancy duration.

[0009] Optionally, step four specifically includes: Extract the unit identifier from the operation and maintenance closure unit, and extract the deadline and duration from the scheduling parameters. Perform pairing and aggregation according to the correspondence between the unit identifier and the scheduling parameters to form a scheduling record set. The deadline times in the scheduling record set are compared, and the records are arranged from the beginning to the end according to the deadline times to form a sequential record set; The operation and maintenance closure units in the sequential record set are arranged in the order of their positions to generate an initial sequence of programs.

[0010] Optionally, step five specifically includes: Extract the operation and maintenance closure units sequentially according to their positions in the initial schedule, extract the scheduling parameters corresponding to each operation and maintenance closure unit, and add the operation and maintenance closure units into the current schedule set item by item according to their positions. The occupancy time of the maintenance closure units included in the current schedule set is continuously accumulated according to their arrangement position to form a cumulative occupancy time that corresponds one-to-one with the arrangement position. The cumulative occupancy time is then compared with the cutoff time of the current arrangement position to determine the over-limit position where the cumulative occupancy time exceeds the cutoff time. Extract the operation and maintenance closure unit from the current scheduling set corresponding to the over-limit location, and extract the set of entity nodes, the set of relation edges, and the state change relationship connected to the operation and maintenance closure unit from the IDC operation and maintenance knowledge graph to form a risk calculation record set; Perform correlation statistics and ratio calculations on the entity node set, relation edge set, state change relation, deadline and duration of occupation in the risk calculation record set to generate the graph risk release efficiency corresponding to the operation and maintenance closure unit in the current schedule set.

[0011] Optionally, step six specifically includes: Extract the operation and maintenance closure unit, deadline, duration of occupation, and graph risk release efficiency from the current schedule set, and perform corresponding sorting according to the arrangement position of the operation and maintenance closure unit in the current schedule set to form a scheduling record set; In the scheduling record set, the operation and maintenance closure units are included in the order formed at the deadline. The graph risk release efficiency is compared item by item to determine the target operation and maintenance closure unit corresponding to the maximum graph risk release efficiency. The sequential inclusion relationship is determined as the deadline sequential inclusion rule, the target operation and maintenance closure unit is determined as the graph risk release efficiency priority removal rule, and the graph risk release efficiency priority removal rule is replaced with the longest processing volume removal rule to form the improved Moore-Hodgson scheduling rule. Based on the improved Moore-Hodgson scheduling rules, the target operation and maintenance closure unit is removed from the current scheduling set, and the remaining operation and maintenance closure units are included in the relational execution arrangement in order to generate an updated scheduling sequence; The occupied time in the updated schedule is re-accumulated according to its sorting position, and the re-accumulated total occupied time is compared with the cutoff time in the updated schedule.

[0012] Optionally, step seven specifically includes: Extract the target operation and maintenance closure unit from the update schedule, and perform aggregation according to the correspondence between the target operation and maintenance closure unit and the scheduling parameters, and include it in the delay pool; Extract the set of entity nodes, the set of relation edges, and the state change relationships connected to the target operation and maintenance closure unit from the IDC operation and maintenance knowledge graph to form a state association record set; Based on the connection direction and time position of the state change relationship, the path is expanded on the set of entity nodes and the set of relationship edges in the state association record set to form a rollback path record set; Perform connectivity and status matching checks on the rollback path record set, retain paths that meet the connectivity and status change conditions, and generate rollback feasible paths corresponding to the target operation and maintenance closure unit.

[0013] Optionally, step eight specifically includes: Extract the target operation and maintenance closure unit and the scheduling parameters corresponding to the target operation and maintenance closure unit from the delay pool, and extract the connection order from the rollback feasible path to form the back-insertion record set; Based on the connection order in the feasible rollback path, the target operation and maintenance closure unit is matched with the execution order of the operation and maintenance closure units in the update schedule to determine the insertion position of the target operation and maintenance closure unit in the update schedule. The target maintenance closure unit is included in the update schedule according to the insertion position. The occupied time after inclusion is re-accumulated according to the arrangement position. The re-accumulated cumulative occupied time is compared with the cutoff time to determine whether the target maintenance closure unit meets the insertion condition. The target operation and maintenance closure unit that meets the back-insertion condition is reintroduced into the update scheduling sequence along the rollback feasible path to form the operation and maintenance closure unit arrangement result, and the operation and maintenance scheduling result is output based on the operation and maintenance closure unit arrangement result.

[0014] According to an embodiment of the present invention, an IDC data center operation and maintenance management system based on a knowledge graph includes: The knowledge graph construction module is used to collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; The closure extraction module is used to extract the influence closure subgraph from the IDC operation and maintenance knowledge graph around the operation and maintenance object to be processed. The closure mapping module is used to perform compression mapping on the influencing closure subgraph, generate operation and maintenance closure units, and form the scheduling parameters corresponding to the operation and maintenance closure units; The initial scheduling module is used to arrange the execution order of the operation and maintenance closure units according to the deadline in the scheduling parameters and generate the initial schedule. The risk calculation module is used to sequentially include the operation and maintenance closure units into the current schedule set according to the initial schedule sequence and update the cumulative occupied time. When the cumulative occupied time exceeds the deadline, it calculates the graph risk release efficiency corresponding to the operation and maintenance closure units in the current schedule set. The scheduling rules module is used to retain the rules for inclusion in order of deadline time, and replaces the rule for removing the longest processing volume with the rule for removing the graph risk release efficiency first, forming an improved Moore-Hodgson scheduling rule. Based on the improved Moore-Hodgson scheduling rule, the target operation and maintenance closure unit is removed from the current scheduling set, and an updated scheduling sequence is generated. The rollback generation module is used to include the target operation and maintenance closure unit into the delay pool and generate the rollback feasible path corresponding to the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph. The back-in output module is used to re-include the target operation and maintenance closure units that meet the back-in conditions in the delay pool into the update schedule along the rollback feasible path and output the operation and maintenance scheduling results.

[0015] The beneficial effects of this invention are: Compared to existing IDC (Internet Data Center) operation and maintenance (O&M) methods that rely on static asset ledgers, manual experience rules, or simple priority ranking, this invention constructs an IDC O&M knowledge graph by collecting related O&M data. It then extracts influence closure subgraphs around the O&M objects to be processed, compresses and maps these subgraphs into O&M closure units. This allows O&M scheduling objects to move beyond isolated devices, single alarms, or individual work orders, forming a unified scheduling granularity constrained by device connectivity, service carrying relationships, and state change relationships. Based on this approach, this invention can more accurately characterize the influence scope and linkage boundaries of the O&M objects to be processed, avoiding the problems of insufficient risk identification, one-sided scheduling basis, and omission of key relationships caused by task fragmentation in traditional methods. This improves the correlation analysis capabilities and overall scheduling accuracy in complex IDC O&M scenarios.

[0016] Furthermore, in the scheduling process of operation and maintenance closure units, this invention retains the rule of sequential inclusion by deadline and replaces the rule of removal by longest processing volume with a priority removal rule based on graph risk release efficiency, forming an improved Moore-Hodgson scheduling rule. When the cumulative occupied time exceeds the deadline, it can combine the entity node set, relation edge set, and state change relationship to conduct a risk release assessment on the current scheduling set, obtaining a target operation and maintenance closure unit removal result that more closely matches the actual operation and maintenance risk structure. At the same time, the target operation and maintenance closure unit is included in the delay pool, and a rollback feasible path is generated based on the state change relationship in the IDC operation and maintenance knowledge graph. Then, the target operation and maintenance closure unit that meets the rollback condition is reinstated into the updated scheduling sequence along the rollback feasible path, so that the entire operation and maintenance management process has both time limit control capability and dynamic recovery capability. Therefore, this invention can not only improve the efficiency of operation and maintenance task sorting and conflict resolution, but also enhance the continuity, stability, and adaptability of operation and maintenance scheduling results. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an IDC data center operation and maintenance management method based on knowledge graphs proposed in this invention; Figure 2 This is a schematic diagram illustrating the construction of an IDC (Internet Data Center) operation and maintenance knowledge graph, which is part of the IDC operation and maintenance management method proposed in this invention. Figure 3 This is a schematic diagram of the improved Moore-Hodgson scheduling method for IDC data center operation and maintenance management based on knowledge graphs proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figures 1-3 A knowledge graph-based IDC (Internet Data Center) operation and maintenance management method includes the following steps: Step 1: Collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; Step 2: Extract the influence closure subgraph from the IDC operations and maintenance knowledge graph around the object to be processed; Step 3: Perform compression mapping on the influencing closure subgraph to generate operation and maintenance closure units and form the scheduling parameters corresponding to the operation and maintenance closure units; Step 4: Based on the deadline in the scheduling parameters, arrange the execution order of the operation and maintenance closure units to generate the initial schedule; Step 5: According to the initial scheduling sequence, the operation and maintenance closure units are sequentially included into the current scheduling set, and the cumulative occupied time is updated. When the cumulative occupied time exceeds the deadline, the graph risk release efficiency corresponding to the operation and maintenance closure units in the current scheduling set is calculated. Step 6: Retain the rule of inclusion of deadline time in order, replace the rule of removal of longest processing volume with the rule of priority removal of graph risk release efficiency, form an improved Moore-Hodgson scheduling rule, and remove the target operation and maintenance closure unit from the current scheduling set according to the improved Moore-Hodgson scheduling rule to generate an updated scheduling sequence; Step 7: Add the target operation and maintenance closure unit to the delay pool, and generate the corresponding rollback feasible path for the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph; Step 8: Re-incorporate the target operation and maintenance closure unit that meets the back-insertion condition in the delay pool into the update schedule along the rollback feasible path, and output the operation and maintenance scheduling result.

[0020] In this embodiment, step one specifically includes: The system collects device records, connection records, alarm records, work order records, service records, status records, and maintenance records from the IDC data center to form IDC data center operation and maintenance related data. Device records represent the basic and operational information of the operation and maintenance objects within the IDC data center; connection records represent the connection relationships between operation and maintenance objects; alarm records represent abnormal event information during the operation of operation and maintenance objects; work order records represent the handling task information generated in response to abnormal events; service records represent the association information between operation and maintenance objects and the services they support; status records represent the status change information of operation and maintenance objects at different times and locations; and maintenance records represent the constraint information corresponding to operation and maintenance activities. All types of records in the IDC data center operation and maintenance related data are collected according to a unified data format. Time alignment, identifier unification, and relationship association are performed on the IDC data center operation and maintenance related data to form an operation and maintenance related record set. Based on the operation and maintenance related record set, the entity node set, relationship edge set, node attribute set, and edge attribute set are determined. Time alignment is used to map the records in the IDC data center operation and maintenance related data to a unified time base. Identifier unification is used to merge different identifiers pointing to the same operation and maintenance object into a unified object identifier. Relationship association is used to establish the correspondence between records according to object identifier, time position, and related fields. In the operation and maintenance related record set, the records used to represent operation and maintenance objects are assigned to the entity node set, the records used to represent the relationship between operation and maintenance objects are assigned to the relationship edge set, the fields used to represent the characteristics of entity nodes are assigned to the node attribute set, and the fields used to represent the characteristics of relationship edges are assigned to the edge attribute set. By performing graph organization and relation mapping on the entity node set, relation edge set, node attribute set, and edge attribute set, an IDC operation and maintenance knowledge graph is generated. Graph organization is used to build the graph structure according to the entity node set and relation edge set. Relation mapping is used to configure the node attribute set to the corresponding entity node and the edge attribute set to the corresponding relation edge. In the IDC operation and maintenance knowledge graph, entity nodes are associated with each other through relation edges. Node attributes and edge attributes are kept to correspond to the corresponding entity nodes and relation edges. The IDC operation and maintenance knowledge graph is used to represent the object relationships and state relationships in the IDC data center operation and maintenance related data.

[0021] In this embodiment, step two specifically includes: In the IDC operations and maintenance knowledge graph, the target entity node corresponding to the operation and maintenance object to be processed is located, and the relationship edges and entity nodes connected to the target entity node are extracted to form a target association record set. The location process searches for matching entity nodes in the IDC operations and maintenance knowledge graph according to the object identifier corresponding to the operation and maintenance object to be processed. After the matching is completed, the relationship edges that have a connection relationship with the target entity node are extracted one by one, and the entity nodes connected to the relationship edges are extracted simultaneously. The target association record set is used to record the correspondence between the target entity node, relationship edges and entity nodes. Based on the direction, type, and state of the relation edges in the target associated record set, an influence propagation search is performed along the relation edges on the target entity nodes. Entity nodes and relation edges that meet the influence propagation conditions are extracted to form an influence propagation record set. The influence propagation search determines the propagation path according to the relation edge direction, the propagation category according to the relation edge type, and the propagation validity according to the node state. Entity nodes and relation edges that meet the influence propagation conditions are written into the influence propagation record set in sequence. The influence propagation record set is used to characterize the influence propagation range obtained by expanding outward from the target entity nodes. The system performs node merging, boundary closure, and graph structure extraction on the impact propagation record set to generate an impact closure subgraph. Node merging is used to merge duplicate entity nodes in the impact propagation record set into a unified node record. Boundary closure is used to determine the node boundaries and relationship edge boundaries corresponding to the impact propagation range. Graph structure extraction is used to extract the corresponding graph structure from the IDC operation and maintenance knowledge graph according to the node boundaries and relationship edge boundaries. The impact closure subgraph is used to represent the impact propagation association structure corresponding to the operation and maintenance object to be processed.

[0022] In this embodiment, step three specifically includes: Extract entity nodes, relation edges, and target entity nodes from the influence closure subgraph. Perform connectivity traversal along the connection paths between entity nodes and the association paths between relation edges to determine the range of entity nodes and the range of relation edges that maintain influence transitive connection with the target entity node, forming a closure range record set. The connectivity traversal starts with the target entity node as the starting node and gradually expands along the connection paths and association paths to include entity nodes and relation edges that meet the influence transitive connection conditions into the closure range record set. The entity nodes in the closure range record set are aggregated and organized according to their connection relationships to form an entity node set. The relation edges in the closure range record set are merged and organized according to their association relationships to form a relation edge set. The entity node set and the relation edge set are mapped and associated to form a closure mapping record set. The aggregation and organization organizes the entity nodes according to the continuity of the connection relationships, and the merging and organization organizes the relation edges according to the consistency of the association relationships. The mapping and association writes the correspondence between the entity node set and the relation edge set into the closure mapping record set. Node shrinking is performed on the entity node set in the closure mapping record set, and edge merging is performed on the relation edge set in the closure mapping record set. The closure mapping record set is compressed into a single graph structure unit, generating an operation and maintenance closure unit. Node shrinking maps the node identifiers in the entity node set to a single node identifier. Edge merging maps the edge connection relationships in the relation edge set to a single edge connection relationship. The compressed single graph structure unit maintains the influence transmission structure and connection boundary corresponding to the target entity node. The time constraint information in the operation and maintenance closure unit is extracted to determine the deadline boundary. The occupancy association information in the operation and maintenance closure unit is accumulated to form scheduling parameters, which include the deadline time and the occupancy duration. The deadline boundary extraction determines the time termination boundary for the operation and maintenance closure unit to participate in the scheduling from the time constraint information. The occupancy accumulation calculates the occupancy records in the occupancy association information according to the time position to form the deadline time and the occupancy duration.

[0023] In this embodiment, step four specifically includes: Extract the unit identifier from the operation and maintenance closure unit, and extract the deadline and duration of use from the scheduling parameters. Perform pairing and aggregation according to the correspondence between the unit identifier and the scheduling parameters to form a scheduling record set. During the extraction process, associate and organize the unit identifier corresponding to each operation and maintenance closure unit with the deadline and duration of use corresponding to the same operation and maintenance closure unit, so that each record in the scheduling record set corresponds to an operation and maintenance closure unit and a combination of scheduling parameters. The deadline times in the scheduling record set are compared, and the records are arranged from the beginning to the end of the deadline times to form a sequential record set. During the time value comparison process, the deadline times in the scheduling record set are read, and the time sequence corresponding to the deadline times is compared item by item. Then, the record positions in the scheduling record set are adjusted according to the comparison results, so that the operation and maintenance closure units with earlier deadline times are arranged first, and the operation and maintenance closure units with later deadline times are arranged last. The operation and maintenance closure units in the sequential record set are arranged in order according to their positions to generate an initial sequence. During the sequential arrangement process, the operation and maintenance closure units are extracted sequentially according to their positions in the sequential record set, and the extraction results are continuously written into a unified sequence structure in the order of arrangement, so that the arrangement order of the operation and maintenance closure units in the initial sequence is consistent with the arrangement order in the sequential record set.

[0024] In this embodiment, step five specifically includes: The operation and maintenance closure units are extracted sequentially according to their positions in the initial scheduling sequence. When extracting each operation and maintenance closure unit, the corresponding scheduling parameters are extracted simultaneously. The deadline and duration of the scheduling parameters correspond one-to-one with the operation and maintenance closure units. Then, the operation and maintenance closure units are added to the current scheduling set item by item according to their positions, so that the order of the operation and maintenance closure units in the current scheduling set is consistent with the order in the initial scheduling sequence. The occupancy time of the maintenance closure units included in the current schedule set is continuously accumulated according to their arrangement position. The occupancy time corresponding to the first arrangement position is directly taken as the cumulative occupancy time. In subsequent arrangement positions, the cumulative occupancy time corresponding to the previous arrangement position is added to the occupancy time corresponding to the current arrangement position to form a cumulative occupancy time that corresponds one-to-one with the arrangement position. The cumulative occupancy time is compared with the deadline of the current arrangement position. If the cumulative occupancy time is greater than the deadline, the position that exceeds the limit is determined. The operation and maintenance closure unit is extracted from the current scheduling set corresponding to the out-of-limit location. Then, the entity node set, relation edge set, and state change relation are extracted from the IDC operation and maintenance knowledge graph according to the connection relationship of the operation and maintenance closure unit. The entity node set represents the range of nodes connected by the operation and maintenance closure unit, the relation edge set represents the range of edges connected by the operation and maintenance closure unit, and the state change relation represents the state change association corresponding to the operation and maintenance closure unit. Finally, the operation and maintenance closure unit, entity node set, relation edge set, and state change relation are aggregated to form a risk calculation record set. For the entity node set, relation edge set, state change relation, deadline and occupancy duration in the risk calculation record set, perform correlation statistics and ratio calculation. The correlation statistics are performed according to the connection correspondence between the operation and maintenance closure unit and the entity node set, relation edge set, and state change relation, and obtain the correlation statistics result for each operation and maintenance closure unit. The ratio calculation combines the correlation statistics result with the deadline and occupancy duration to generate the graph risk release efficiency corresponding to the operation and maintenance closure unit in the current schedule set.

[0025] In this embodiment, step six specifically includes: Extract the operation and maintenance closure unit, deadline, duration of occupation, and graph risk release efficiency from the current schedule set. Organize the operation and maintenance closure unit according to its position in the current schedule set to form a scheduling record set. Each record in the scheduling record set corresponds to a position and includes an operation and maintenance closure unit, a deadline, a duration of occupation, and a graph risk release efficiency. The scheduling record set maintains the inclusion relationship of operation and maintenance closure units according to the order formed at the deadline. The graph risk release efficiency is compared item by item according to the arrangement position. When comparing, the graph risk release efficiency corresponding to the first arrangement position is first determined as the current maximum value. Then, the current maximum value is compared with the graph risk release efficiency corresponding to the next arrangement position. When the current value is greater than the current maximum value, the current maximum value is updated until the comparison of all arrangement positions is completed, and the target operation and maintenance closure unit corresponding to the maximum graph risk release efficiency is obtained. The sequential inclusion relationship is determined as the deadline sequential inclusion rule, and the removal relationship of the target operation and maintenance closure unit is determined as the graph risk release efficiency priority removal rule. The graph risk release efficiency priority removal rule replaces the longest processing volume removal rule, forming an improved Moore-Hodgson scheduling rule. The deadline sequential inclusion rule limits the arrangement basis when operation and maintenance closure units enter the current scheduling set, and the graph risk release efficiency priority removal rule limits the objects to be removed when the cumulative occupied time exceeds the limit. According to the improved Moore-Hodgson scheduling rules, the target operation and maintenance closure unit is removed from the current scheduling set. The remaining operation and maintenance closure units are arranged and maintained in order, and an updated scheduling sequence is generated. After the removal is completed, the operation and maintenance closure units after the target operation and maintenance closure unit are moved forward in the original arrangement direction, and the operation and maintenance closure units before the target operation and maintenance closure unit maintain their original arrangement position relationship. The occupied time in the updated schedule is re-accumulated according to the arrangement position, and the re-accumulated cumulative occupied time is compared with the cutoff time in the updated schedule. The re-accumulation starts from the first arrangement position. The occupied time corresponding to the first arrangement position is directly taken as the cumulative occupied time. In any subsequent arrangement position, the cumulative occupied time corresponding to the previous arrangement position is added to the occupied time corresponding to the current arrangement position. Then, the cumulative occupied time corresponding to the current arrangement position is compared with the cutoff time corresponding to the current arrangement position.

[0026] This invention addresses the limitations of the traditional Moore-Hodgson algorithm, which is only applicable to single-machine delay control scenarios where processing volume is the core criterion and is ill-suited to the complex relationships between maintenance objects in IDC data centers, the intertwined state propagation paths, and the dynamic changes in the scope of business impact. While retaining the rule of sequential inclusion of deadlines, it transforms the original rule of removing the longest processing volume from the algorithm into a rule prioritizing removal based on graph risk release efficiency. This shifts scheduling decisions from a purely mechanical selection based on occupancy time to a comprehensive judgment that considers the set of entity nodes, relational edges, and state change relationships associated with the maintenance closure unit in the IDC maintenance knowledge graph. This elevates the traditional isolated job processing approach to a graph-based scheduling approach oriented towards associated closure units. Specifically, it extracts maintenance closure units, deadlines, occupancy times, and graph risk release efficiency from the current schedule set to form a scheduling record set, providing a unified, complete, and comparable evaluation basis for each scheduled object. Furthermore, it compares graph risks item by item. The release efficiency determines the target maintenance closure unit, enabling the removal of objects to prioritize maintenance closure units that contribute more to the risk release of the current schedule set and cause less disturbance to the overall relational structure. This avoids the problems of traditional longest processing volume removal methods in IDC data center scenarios, which are prone to mistakenly removing key related tasks, amplifying business propagation risks, and disrupting state connection links. Furthermore, after removing the target maintenance closure unit, this invention maintains the remaining maintenance closure units in the order of inclusion relationship and re-accumulates and compares the occupied time in the updated schedule sequence. This makes the improved Moore-Hodgson scheduling rule not only have the scheduling correction capability under the deadline constraint, but also the risk suppression capability and continuous reordering capability for knowledge graph relational structures. Therefore, the improved algorithm can significantly improve the sorting accuracy, over-limit processing targeting, and scheduling result stability of IDC data center maintenance tasks in complex relational scenarios, reduce the frequency of manual intervention, shorten the duration of risk propagation, and improve the identification accuracy of key maintenance objects and the efficiency of resource occupation conflict resolution.

[0027] In this embodiment, step seven specifically includes: Extract the target operation and maintenance closure unit from the updated schedule list, and perform aggregation according to the correspondence between the target operation and maintenance closure unit and the scheduling parameters, and include it in the delay pool. The extraction process locates the target operation and maintenance closure unit that has been moved out according to the arrangement content in the updated schedule list, and then pairs and organizes it according to the scheduling parameters corresponding to the target operation and maintenance closure unit, so that each record in the delay pool contains a target operation and maintenance closure unit and a set of scheduling parameters. Extract the set of entity nodes, the set of relation edges, and the state change relations connected to the target operation and maintenance closure unit from the IDC operation and maintenance knowledge graph to form a state association record set. The extraction process starts from the connection position of the target operation and maintenance closure unit in the IDC operation and maintenance knowledge graph, expands outward along the connection relations, classifies the entity nodes that maintain the connection relations into the entity node set, classifies the connection relations between entity nodes into the relation edge set, and classifies the state change content corresponding to the entity nodes and relation edges into the state change relations. According to the connection direction and time position of the state change relationship, the path expansion is performed on the entity node set and the relation edge set in the state association record set to form a rollback path record set. The path expansion process first determines the path extension direction according to the connection direction of the state change relationship, then determines the path arrangement order according to the time position, and connects the entity nodes in the entity node set and the relation edges in the relation edge set in sequence along the extension direction and arrangement order to form continuous paths, which are then written into the rollback path record set. Perform connectivity and state matching checks on the rollback path record set, retain paths that meet both connectivity and state change conditions, and generate rollback feasible paths corresponding to the target operation and maintenance closure unit. The connectivity check examines the integrity of the connection between entity nodes and the continuity of the connection between relation edges in the path. The state matching check compares the state change relationships in the path with the state changes corresponding to the target operation and maintenance closure unit. When both connectivity and state matching checks are met, the corresponding path is determined as a rollback feasible path.

[0028] In this embodiment, step eight specifically includes: Extract the target operation and maintenance closure unit and the scheduling parameters corresponding to the target operation and maintenance closure unit from the delay pool, and extract the connection order from the rollback feasible path to form an insertion record set. The extraction process reads the correspondence between the target operation and maintenance closure unit and the scheduling parameters according to the record content in the delay pool, and then reads the connection order according to the path connection order in the rollback feasible path. Finally, the target operation and maintenance closure unit, scheduling parameters and connection order are organized into the same record. Based on the connection order in the feasible path of rollback, the target operation and maintenance closure unit is sequentially connected with the operation and maintenance closure units in the update schedule to determine the insertion position of the target operation and maintenance closure unit in the update schedule. The sequential connection process compares the arrangement position relationship in the update schedule item by item according to the connection order. The insertion point of the target operation and maintenance closure unit is determined at the front, back, or adjacent position of the position corresponding to the connection order, and the insertion point is determined as the insertion position. The target maintenance closure unit is included in the update schedule according to the insertion position. The occupied time after inclusion is re-accumulated according to the arrangement position. The re-accumulated cumulative occupied time is compared with the deadline to determine if the target maintenance closure unit meets the insertion condition. The re-accumulation starts from the first arrangement position of the update schedule. The corresponding occupied time at the first arrangement position is directly taken as the cumulative occupied time. At any subsequent arrangement position, the cumulative occupied time corresponding to the previous arrangement position is added to the occupied time corresponding to the current arrangement position. Then, the cumulative occupied time corresponding to each arrangement position is compared with the deadline corresponding to the same arrangement position. If the comparison result meets the deadline constraint, the target maintenance closure unit is determined to meet the insertion condition. The target maintenance closure unit that meets the back-insertion condition is re-incorporated into the update scheduling sequence along the rollback feasible path to form the maintenance closure unit arrangement result. The maintenance scheduling result is output based on the maintenance closure unit arrangement result. The re-incorporation process maintains the connection relationship between the target maintenance closure unit and the update scheduling sequence according to the connection order in the rollback feasible path. The maintenance closure unit arrangement result records the arrangement content after the target maintenance closure unit is back-inserted. The maintenance scheduling result is determined based on the maintenance closure unit arrangement result.

[0029] A knowledge graph-based IDC (Internet Data Center) operation and maintenance management system includes: The knowledge graph construction module is used to collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; The closure extraction module is used to extract the influence closure subgraph from the IDC operation and maintenance knowledge graph around the operation and maintenance object to be processed. The closure mapping module is used to perform compression mapping on the influencing closure subgraph, generate operation and maintenance closure units, and form the scheduling parameters corresponding to the operation and maintenance closure units; The initial scheduling module is used to arrange the execution order of the operation and maintenance closure units according to the deadline in the scheduling parameters and generate the initial schedule. The risk calculation module is used to sequentially include the operation and maintenance closure units into the current schedule set according to the initial schedule sequence and update the cumulative occupied time. When the cumulative occupied time exceeds the deadline, it calculates the graph risk release efficiency corresponding to the operation and maintenance closure units in the current schedule set. The scheduling rules module is used to retain the rules for inclusion in order of deadline time, and replaces the rule for removing the longest processing volume with the rule for removing the graph risk release efficiency first, forming an improved Moore-Hodgson scheduling rule. Based on the improved Moore-Hodgson scheduling rule, the target operation and maintenance closure unit is removed from the current scheduling set, and an updated scheduling sequence is generated. The rollback generation module is used to include the target operation and maintenance closure unit into the delay pool and generate the rollback feasible path corresponding to the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph. The back-in output module is used to re-include the target operation and maintenance closure units that meet the back-in conditions in the delay pool into the update schedule along the rollback feasible path and output the operation and maintenance scheduling results.

[0030] Example 1: To verify the feasibility of this invention in practice, it was applied to an IDC (Internet Data Center) data center operation and maintenance scenario that handles virtualization computing, database storage, service access, and environmental monitoring. This IDC data center includes server racks, servers, switching equipment, storage devices, power supply and distribution equipment, cooling equipment, and supporting monitoring terminals. During daily operation, it continuously generates device records, connection records, alarm records, work order records, service records, status records, and maintenance records. Existing operation and maintenance methods primarily rely on manual experience, static priorities, and fixed work order sequences. When a pending operation and maintenance object is simultaneously associated with link anomalies, service load fluctuations, and status change conflicts, problems often arise such as incomplete assessment of the impact scope, unbalanced scheduling, and the inability to promptly reinsert temporarily removed tasks. Especially when the accumulated usage time approaches the deadline, traditional methods often only mechanically postpone work orders, easily leading to slow recovery of critical services and frequent manual adjustments.

[0031] In practical applications, various operation and maintenance records generated during the data center's operation phase are first collected and organized to form IDC data center operation and maintenance related data. Then, an IDC operation and maintenance knowledge graph is constructed based on object identifiers, connection relationships, service carrying relationships, and status change relationships. For a pending operation and maintenance event that includes switching equipment anomalies, uplink fluctuations, storage access latency, and service instance alarms, the impact closure subgraph is extracted after locating the target entity node in the graph. After compression mapping, the related impacts, originally scattered across 28 entity nodes and 41 relationship edges, are compressed into 6 operation and maintenance closure units, each corresponding to a deadline and duration. Subsequently, an initial scheduling sequence is generated based on the deadline, and the 6 operation and maintenance closure units are sequentially incorporated into the current scheduling set. When the fourth maintenance closure unit was included, the cumulative occupied time reached 117 minutes, exceeding the 102 minutes corresponding to the current deadline. The system then extracted the set of entity nodes, relational edges, and state change relationships connected to the current schedule set from the IDC maintenance knowledge graph, calculated the graph risk release efficiency, and replaced the longest processing volume removal rule with an improved Moore-Hodgson scheduling rule. This allowed the target maintenance closure unit, with a long occupied time but low business propagation impact and a clear state change rollback path, to be included in the delay pool. After the delay pool was formed, the system expanded the rollback path along the state change relationships, obtaining a feasible rollback path that met both connectivity and state matching conditions. After the remaining maintenance closure units completed processing and released their occupancy, the target maintenance closure unit was re-included in the update schedule, thus completing a round of maintenance scheduling that balanced deadline constraints and risk propagation suppression.

[0032] To verify the effectiveness of this invention, under the same data center load level and the same scale of maintenance events, 30 consecutive rounds of maintenance scheduling tasks were selected, and the method of this invention was compared with the traditional fixed-priority scheduling method. The total number of tasks involved in the comparison was 180, the total number of associated entity nodes was 842, and the total number of associated relationship edges was 1296, with an average of 6 maintenance closure units to be scheduled per round. The statistical results are shown in Table 1. Table 1. Comparison of Scheduling Effectiveness

[0033] As shown in Table 1, the method of this invention significantly outperforms traditional fixed-priority scheduling methods in terms of accuracy in impact range identification, scheduling generation efficiency, frequency of manual intervention, deadline control, business recovery speed, and delayed task reinsertion capability. This is because the present invention first uses an IDC operations and maintenance knowledge graph to model the operations and maintenance objects to be processed, and then compresses the impact closure subgraph to form operations and maintenance closure units, enabling the scheduling objects to cover the actual impact boundaries. When the cumulative occupancy time exceeds the limit, the graph risk release efficiency priority removal rule replaces the longest processing volume removal rule, preventing key impact units from being simply removed. Simultaneously, the delayed pool and rollback feasible paths enable controllable reinsertion of target operations and maintenance closure units, ultimately resulting in better stability and practicality of the operations and maintenance scheduling results in terms of time limit control, risk suppression, and dynamic recovery.

[0034] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A knowledge graph-based IDC (Internet Data Center) operation and maintenance management method, characterized in that, Includes the following steps: Step 1: Collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; Step 2: Extract the influence closure subgraph from the IDC operations and maintenance knowledge graph around the object to be processed; Step 3: Perform compression mapping on the influencing closure subgraph to generate operation and maintenance closure units and form the scheduling parameters corresponding to the operation and maintenance closure units; Step 4: Based on the deadline in the scheduling parameters, arrange the execution order of the operation and maintenance closure units to generate the initial schedule; Step 5: According to the initial scheduling sequence, the operation and maintenance closure units are sequentially included into the current scheduling set, and the cumulative occupied time is updated. When the cumulative occupied time exceeds the deadline, the graph risk release efficiency corresponding to the operation and maintenance closure units in the current scheduling set is calculated. Step 6: Retain the rule of inclusion of deadline time in order, replace the rule of removal of longest processing volume with the rule of priority removal of graph risk release efficiency, form an improved Moore-Hodgson scheduling rule, and remove the target operation and maintenance closure unit from the current scheduling set according to the improved Moore-Hodgson scheduling rule to generate an updated scheduling sequence; Step 7: Add the target operation and maintenance closure unit to the delay pool, and generate the corresponding rollback feasible path for the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph; Step 8: Re-incorporate the target operation and maintenance closure unit that meets the back-insertion condition in the delay pool into the update schedule along the rollback feasible path, and output the operation and maintenance scheduling result.

2. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step one specifically involves: Collect equipment records, connection records, alarm records, work order records, service records, status records, and maintenance records in the IDC data center to form IDC data center operation and maintenance related data; The operation and maintenance related data of the IDC data center are aligned in time, unified in identification and related in relation to form an operation and maintenance related record set. Based on the operation and maintenance related record set, the entity node set, relationship edge set, node attribute set and edge attribute set are determined. Perform graph organization and relation mapping on the entity node set, relation edge set, node attribute set, and edge attribute set to generate an IDC operation and maintenance knowledge graph.

3. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step two specifically involves: Locate the target entity node corresponding to the operation and maintenance object to be processed in the IDC operation and maintenance knowledge graph, and extract the relationship edges and entity nodes connected to the target entity node to form a target association record set; Based on the direction, type, and state of the relation edges in the target associated record set, perform an influence propagation search along the relation edges on the target entity nodes, extract entity nodes and relation edges that meet the influence propagation conditions, and form an influence propagation record set; Perform node merging, boundary closure, and graph structure extraction on the influence propagation record set to generate an influence closure subgraph.

4. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step three specifically involves: Extract entity nodes, relation edges, and target entity nodes from the influence closure subgraph. Perform connectivity traversal along the connection paths between entity nodes and the association paths between relation edges to determine the range of entity nodes and the range of relation edges that maintain influence transition connections with the target entity node, forming a closure range record set. The entity nodes in the closure range record set are aggregated and organized according to their connection relationships to form an entity node set. The relation edges in the closure range record set are merged and organized according to their association relationships to form a relation edge set. The entity node set and the relation edge set are then mapped and associated to form a closure mapping record set. Perform node shrinking on the entity node set in the closure mapping record set, perform edge relationship merging on the relation edge set in the closure mapping record set, compress the closure mapping record set into a single graph structure unit, and generate the operation and maintenance closure unit. The time constraint information in the operation and maintenance closure unit is extracted to meet the deadline boundary, and the occupancy association information in the operation and maintenance closure unit is accumulated to form scheduling parameters, which include the deadline time and the occupancy duration.

5. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step four specifically involves: Extract the unit identifier from the operation and maintenance closure unit, and extract the deadline and duration from the scheduling parameters. Perform pairing and aggregation according to the correspondence between the unit identifier and the scheduling parameters to form a scheduling record set. The deadline times in the scheduling record set are compared, and the records are arranged from the beginning to the end according to the deadline times to form a sequential record set; The operation and maintenance closure units in the sequential record set are arranged in the order of their positions to generate an initial sequence of programs.

6. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step five specifically involves: Extract the operation and maintenance closure units sequentially according to their positions in the initial schedule, extract the scheduling parameters corresponding to each operation and maintenance closure unit, and add the operation and maintenance closure units into the current schedule set item by item according to their positions. The occupancy time of the maintenance closure units included in the current schedule set is continuously accumulated according to their arrangement position to form a cumulative occupancy time that corresponds one-to-one with the arrangement position. The cumulative occupancy time is then compared with the cutoff time of the current arrangement position to determine the over-limit position where the cumulative occupancy time exceeds the cutoff time. Extract the operation and maintenance closure unit from the current scheduling set corresponding to the over-limit location, and extract the set of entity nodes, the set of relation edges, and the state change relationship connected to the operation and maintenance closure unit from the IDC operation and maintenance knowledge graph to form a risk calculation record set; Perform correlation statistics and ratio calculations on the entity node set, relation edge set, state change relation, deadline and duration of occupation in the risk calculation record set to generate the graph risk release efficiency corresponding to the operation and maintenance closure unit in the current schedule set.

7. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step six specifically involves: Extract the operation and maintenance closure unit, deadline, duration of occupation, and graph risk release efficiency from the current schedule set, and perform corresponding sorting according to the arrangement position of the operation and maintenance closure unit in the current schedule set to form a scheduling record set; In the scheduling record set, the operation and maintenance closure units are included in the order formed at the deadline. The graph risk release efficiency is compared item by item to determine the target operation and maintenance closure unit corresponding to the maximum graph risk release efficiency. The sequential inclusion relationship is determined as the deadline sequential inclusion rule, the target operation and maintenance closure unit is determined as the graph risk release efficiency priority removal rule, and the graph risk release efficiency priority removal rule is replaced with the longest processing volume removal rule to form the improved Moore-Hodgson scheduling rule. Based on the improved Moore-Hodgson scheduling rules, the target operation and maintenance closure unit is removed from the current scheduling set, and the remaining operation and maintenance closure units are included in the relational execution arrangement in order to generate an updated scheduling sequence; The occupied time in the updated schedule is re-accumulated according to its sorting position, and the re-accumulated total occupied time is compared with the cutoff time in the updated schedule.

8. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step seven specifically involves: Extract the target operation and maintenance closure unit from the update schedule, and perform aggregation according to the correspondence between the target operation and maintenance closure unit and the scheduling parameters, and include it in the delay pool; Extract the set of entity nodes, the set of relation edges, and the state change relationships connected to the target operation and maintenance closure unit from the IDC operation and maintenance knowledge graph to form a state association record set; Based on the connection direction and time position of the state change relationship, the path is expanded on the set of entity nodes and the set of relationship edges in the state association record set to form a rollback path record set; Perform connectivity and status matching checks on the rollback path record set, retain paths that meet the connectivity and status change conditions, and generate rollback feasible paths corresponding to the target operation and maintenance closure unit.

9. The knowledge graph-based IDC data center operation and maintenance management method according to claim 1, characterized in that, Step eight specifically involves: Extract the target operation and maintenance closure unit and the scheduling parameters corresponding to the target operation and maintenance closure unit from the delay pool, and extract the connection order from the rollback feasible path to form the back-insertion record set; Based on the connection order in the feasible rollback path, the target operation and maintenance closure unit is matched with the execution order of the operation and maintenance closure units in the update schedule to determine the insertion position of the target operation and maintenance closure unit in the update schedule. The target maintenance closure unit is included in the update schedule according to the insertion position. The occupied time after inclusion is re-accumulated according to the arrangement position. The re-accumulated cumulative occupied time is compared with the cutoff time to determine whether the target maintenance closure unit meets the insertion condition. The target operation and maintenance closure unit that meets the back-insertion condition is reintroduced into the update scheduling sequence along the rollback feasible path to form the operation and maintenance closure unit arrangement result, and the operation and maintenance scheduling result is output based on the operation and maintenance closure unit arrangement result.

10. A knowledge graph-based IDC data center operation and maintenance management system, comprising the knowledge graph-based IDC data center operation and maintenance management method according to any one of claims 1 to 9, characterized in that, include: The knowledge graph construction module is used to collect IDC data center operation and maintenance related data and build an IDC operation and maintenance knowledge graph; The closure extraction module is used to extract the influence closure subgraph from the IDC operation and maintenance knowledge graph around the operation and maintenance object to be processed. The closure mapping module is used to perform compression mapping on the influencing closure subgraph, generate operation and maintenance closure units, and form the scheduling parameters corresponding to the operation and maintenance closure units; The initial scheduling module is used to arrange the execution order of the operation and maintenance closure units according to the deadline in the scheduling parameters and generate the initial schedule. The risk calculation module is used to sequentially include the operation and maintenance closure units into the current schedule set according to the initial schedule sequence and update the cumulative occupied time. When the cumulative occupied time exceeds the deadline, it calculates the graph risk release efficiency corresponding to the operation and maintenance closure units in the current schedule set. The scheduling rules module is used to retain the rules for inclusion in order of deadline time, and replaces the rule for removing the longest processing volume with the rule for removing the graph risk release efficiency first, forming an improved Moore-Hodgson scheduling rule. Based on the improved Moore-Hodgson scheduling rule, the target operation and maintenance closure unit is removed from the current scheduling set, and an updated scheduling sequence is generated. The rollback generation module is used to include the target operation and maintenance closure unit into the delay pool and generate the rollback feasible path corresponding to the target operation and maintenance closure unit based on the state change relationship in the IDC operation and maintenance knowledge graph. The back-in output module is used to re-include the target operation and maintenance closure units that meet the back-in conditions in the delay pool into the update schedule along the rollback feasible path and output the operation and maintenance scheduling results.

Images