Land affair intelligent decision support method oriented to cross-department collaboration

Abstract

The invention discloses a cross-department collaboration-oriented land transaction intelligent decision support method, particularly relates to the technical field of land transaction management, and is used for solving the problems of decision tense chaos and low approval efficiency caused by data timeliness difference in the prior art. The method comprises the following steps: acquiring real-time monitoring and periodic revision data of multiple service units, adding timestamps, and identifying time window differences to extract an associated data set; dynamically aligning timestamps and generating timeliness weights on the basis of flow topological characteristics and information orderliness analysis, constructing a space-time propagation path map, screening high-weight paths, and realizing dynamic splicing of lagged data and real-time data; an approval path scheme with timeliness alignment is generated through rule base matching and a risk early warning mechanism, and risks are detected in real time during execution to dynamically adjust a priority sequence, so that the consistency and execution efficiency of cross-department decision making are improved, and the compliance risk of land resource management and control is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of land affair management. More specifically, the present invention relates to an intelligent decision-making support method for land affairs oriented to cross-departmental collaboration. Background Art

[0002] Within large enterprise groups, land affair management for construction projects involves data collaboration and process integration among multiple business units such as land use approval, demolition coordination, and compliance review. Existing technologies usually achieve internal data sharing by constructing a unified management platform and perform process-based control of land resources based on preset rules. However, due to differences in data collection mechanisms and business objectives among different business units, there are significant differences in their data update cycles and timeliness. For example, real-time environmental monitoring data at the project site needs to be dynamically updated to reflect sudden changes, while land use planning data may be revised at fixed intervals. Such data is synchronously called and used as an analysis basis during joint decision-making, resulting in the forced alignment of information in different time dimensions to the same reference system during the decision-making process.

[0003] When dealing with the timeliness differences in cross-business unit data collaboration, existing technologies lack dynamic adaptation capabilities. The update frequencies and time windows of data from different business units are inconsistent, leading to conflicts in the temporal logic of decision-making bases, and further causing the results of land affair management to deviate from actual business needs. This contradiction is particularly prominent in emergency projects or complex land use scenarios that require the integration of real-time monitoring data and historical planning data, resulting in reduced approval efficiency and increased compliance risks, thus affecting the intensive management and control of enterprise land resources. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an intelligent decision-making support method for land affairs oriented to cross-departmental collaboration to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions: An intelligent decision-making support method for land affairs oriented to cross-departmental collaboration, comprising the following steps: S1. Obtain land affair data of multiple business units within the enterprise group. The land affair data includes real-time monitoring data and periodically revised data, and attach time stamps to each data item; S2. Identify the time window differences between the real-time monitoring data and the periodically revised data based on the time stamps, and extract an associated data set with overlapping time windows and inconsistent update cycles; S3. Dynamically align the time stamps of each data item in the associated data set, and generate the timeliness weights of each data item based on the process topology characteristics and information order characteristics of the multi-department approval process; S4. Construct a spatio-temporal propagation path map based on timeliness weights. After screening the propagation paths, dynamically splice the lagging and real-time data items along the high-weight paths to generate a decision input data set with timeliness alignment. S5. Call the preset multi-department collaborative approval rule library according to the decision input data set to generate an approval path plan and a risk warning list that match the timeliness weights. S6. Execute the approval path plan, and in the approval process nodes, detect the triggering conditions of the risk warning list in real time to dynamically adjust the priority order of the subsequent approval paths.

[0006] In a preferred embodiment, S1 includes: Obtain real-time monitoring data and mark the acquisition time point as the reference time of the time stamp. Obtain periodically revised data. The periodically revised data is batch exported by the enterprise resource planning system according to a preset period, and time stamps are attached according to the effective time in the data revision record. Verify the legality of the time stamps of the real-time monitoring data and the periodically revised data. Classify and store the verified land transaction data into a distributed database according to the business unit identifier, and establish a time stamp index to support subsequent data queries.

[0007] In a preferred embodiment, S2 includes: Calculate the real-time data window length according to the time stamp of the real-time monitoring data. Calculate the revised data window length according to the time stamp of the periodically revised data. Traverse the time windows of the real-time monitoring data and the periodically revised data, and extract the data item set where the time windows overlap and the ratio of the real-time data window length to the revised data window length exceeds a preset ratio threshold. Based on the business unit identifier and data item attributes of each data item in the data item set, establish cross-business unit association rules, and screen the data items that meet the association rules and have inconsistent update cycles to generate an associated data set.

[0008] In a preferred embodiment, the real-time data window length is the first preset time threshold pushed forward from the acquisition time point to the second preset time threshold pushed backward; the revised data window length is the third preset time threshold pushed forward from the effective time to the fourth preset time threshold pushed backward.

[0009] In a preferred embodiment, S3 includes: Construct a collaborative path topology network for the multi-department approval process. The nodes represent the approval departments, the edges represent the cross-department data interaction paths, and calculate the betweenness centrality of each node as the topological vulnerability in the process topology features. Parse the historical operation logs of the approval path, count the number of decision divergence events and the number of redundant process operations at each approval node, and calculate the entropy increase dissipation value of the approval path based on the information entropy formula as the information orderliness feature; According to the quantization results of topological vulnerability and entropy increase dissipation value, dynamically weight and correct the timestamp offset of each data item in the associated data set; Generate the timeliness weight of each data item based on the corrected timestamp offset. The timeliness weight is negatively correlated with topological vulnerability and positively correlated with the entropy increase dissipation value.

[0010] In a preferred embodiment, S4 includes: Calculate the path weight of each data item in the associated data set based on the timeliness weight. The path weight is the product of the timeliness weight and the timestamp adjacency of the data item; Construct a spatio-temporal propagation path map, where nodes represent data items, edges represent the spatio-temporal dependence relationship between data items, and the edge weight is the normalized value of the path weight; Screen high-weight paths according to the edge weight. High-weight paths are defined as continuous node sequences whose cumulative sum of edge weights exceeds a preset path threshold and whose path length is less than a preset hop threshold; Dynamically splice lagging data items and real-time data items along the high-weight path according to the timestamp adjacency; Generate a timeliness-aligned decision input data set in the order of the spatio-temporal propagation path after splicing.

[0011] In a preferred embodiment, the splicing rule is that when the time difference between the timestamp of the lagging data item and the timestamp of the real-time data item is less than a preset alignment tolerance, they are merged into the same decision input data unit.

[0012] In a preferred embodiment, S5 includes: Parse the timeliness weight and data item attributes in the decision input data set, and match the applicable approval rules in the multi-department collaborative approval rule library based on the rule priority mapping table; Generate an initial approval path plan according to the applicable approval rules. The initial approval path plan includes the approval department node sequence and the approval time limit constraint between nodes; Detect the timeliness weight attenuation rate of the node sequence in the initial approval path plan. If the timeliness weight attenuation rate exceeds the preset risk threshold, trigger a risk warning event and mark the risk node; Adjust the node priority of the approval path plan according to the risk warning event, and generate an optimized approval path plan and a risk warning list; Output the optimized approval path plan and the risk warning list to the approval execution terminal in the order of the timestamp of the approval department node sequence.

[0013] In a preferred embodiment, the risk warning list includes risk node identifiers, risk types, and mitigation strategies.

[0014] In a preferred embodiment, S6 includes: Distribute the approval department node sequence in the optimized approval path plan to the corresponding approval execution terminals, and monitor the approval status of each node and the triggering conditions of the risk warning list; When it is detected that the risk node identifier in the risk warning list matches the current approval node, extract the risk type and mitigation strategy, and calculate the priority adjustment coefficient of the current node according to the risk level; Reorder the subsequent approval department node sequence based on the priority adjustment coefficient to generate a dynamically adjusted approval path plan, and push the adjusted task list to the relevant approval execution terminals; Record the priority adjustment event in the approval process log, including the risk node identifier, adjustment timestamp, comparison data of the original path plan and the adjusted path plan; Update the risk status in the risk warning list according to the adjusted path plan. If the risk mitigation strategy has been executed and the risk level has dropped below the threshold, remove the corresponding risk node identifier.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By dynamically integrating the timeliness characteristics of multi-source data and the process topology relationship, a closed-loop optimization system for cross-department land transaction decision-making is constructed; the timeliness weights generated based on the process topology characteristics and information order analysis can quantitatively reflect the dynamic credibility and urgency of different data items in the approval process, replacing the manual experience setting under traditional static rules; through the dynamic pruning and splicing mechanism of the spatio-temporal propagation path map, the spatio-temporal correlation of lagged data and real-time data is effectively integrated, avoiding information distortion caused by forced time alignment; while ensuring the timeliness of data, the integrity and consistency of decision-making basis are significantly improved. Especially in scenarios involving sudden environmental changes or complex planning adjustments, decision-making inputs matching the business reality can be quickly generated, reducing the cost of manual secondary verification.

[0016] 2. Through the dynamic feedback mechanism of the risk warning list and the approval path, real-time correction of the decision execution process is realized; the path priority adjustment driven by the timeliness weight can adaptively identify and respond to potential risk nodes in the process, such as approval delays or resource conflicts, thereby optimizing the resource allocation efficiency while ensuring compliance; the synergistic effect of the rule base and real-time data enables the approval path plan to not only comply with the preset management specifications but also flexibly adapt to dynamic business requirements, reducing the compliance risk caused by temporal conflicts while improving the efficiency of land transaction disposal, providing reliable support for the enterprise group to achieve intensive management of land resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flowchart of the intelligent decision-making support method for land affairs oriented to cross-departmental collaboration of the present invention; Figure 2 This is a flowchart for generating the timeliness weights of each data item of the present invention. Detailed implementation manners

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] Embodiment: Figure 1 The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration of the present invention is given, including the following steps: S1. Obtain the land affair data of multiple business units within the enterprise group. The land affair data includes real-time monitoring data and periodically revised data, and attach a time stamp to each data item; S2. Based on the time stamp, identify the time window differences between the real-time monitoring data and the periodically revised data, and extract the associated data set with overlapping time windows and inconsistent update cycles; S3. Dynamically align the time stamps of each data item in the associated data set, and generate the timeliness weights of each data item based on the process topology characteristics and information order characteristics of the multi-department approval process; S4. Construct a spatio-temporal propagation path map based on the timeliness weights, screen the propagation paths, and then dynamically splice the lagging and real-time data items along the high-weight paths to generate a decision input data set with timeliness alignment; S5. According to the decision input data set, call the preset multi-department collaborative approval rule library to generate an approval path plan and a risk warning list that match the timeliness weights; S6. Execute the approval path plan, and in the approval process nodes, real-time detect the triggering conditions of the risk warning list to dynamically adjust the priority order of the subsequent approval paths.

[0020] S1. Obtain the land affair data of multiple business units within the enterprise group. The land affair data includes real-time monitoring data and periodically revised data, and attach a time stamp to each data item. The specific implementation is as follows: When obtaining land transaction data of multiple business units within an enterprise group, real-time monitoring data is collected in real time through environmental sensors and engineering equipment deployed at the project site. Environmental sensors include soil moisture sensors, air quality monitors, and noise detectors, while engineering equipment includes positioning terminals and vibration monitors for construction machinery. The real-time collected data is transmitted to the central data platform through Internet of Things communication protocols, including the MQTT protocol and the LoRaWAN protocol.

[0021] The collection time point is based on Coordinated Universal Time (UTC), and a timestamp is generated by the clock chip built into the sensor. The timestamp format is "YYYY-MM-DD HH:MM:SS", accurate to the second level. For example, if a soil moisture sensor at a construction site collects a humidity value of 45% at 09:30:15 on October 1, 2023, the timestamp "2023-10-01 09:30:15" is attached and uploaded to the platform through the MQTT protocol. Periodically revised data is batch exported by the enterprise resource planning system according to a preset cycle, and the preset cycles include monthly, quarterly, or annual, and the specific cycle is determined by the internal management rules of the enterprise group. The effective time in the data revision record is based on the approval completion time entered in the enterprise resource planning system. For example, if the land use nature of a plot is revised to "industrial land" in the third quarter of 2023, and the approval completion time is 17:00:00 on September 30, 2023, the timestamp "2023-09-30 17:00:00" is attached.

[0022] When verifying the legality of the timestamps of real-time monitoring data and periodically revised data, first check whether the timestamp falls within the data valid time range of the corresponding business unit. The data valid time range is preset according to the type of business unit. For example, the data valid time range for the environmental monitoring unit is 24 hours before the collection time point to 1 hour after, and the data valid time range for the land use planning unit is from the effective time of the most recent revision to the effective time of the next revision. If the timestamp exceeds the valid time range, it is determined as illegal data and marked with an abnormal status. Secondly, check whether there is a future timestamp, that is, a timestamp later than the Coordinated Universal Time of the current system time. For example, if the current system time is 2023-10-01 10:00:00, and the timestamp of a data item is 2023-10-01 10:30:00, it is determined as a future timestamp and marked as abnormal.

[0023] The verified land transaction data is classified and stored in a distributed database according to the business unit identifier. The business unit identifier is a unique string code in the format of "department abbreviation_business type_sequence number". For example, "EP_MONITOR_001" represents the data of the first monitoring point in the environmental protection monitoring department, and "LAND_PLAN_003" represents the data of the third planning scheme in the land use planning department. The distributed database adopts a sharding storage mechanism, where each shard corresponds to a business unit identifier, and the data is partitioned by timestamp range. The sharding rule is to divide the timestamp by natural month. For example, the environmental protection monitoring data in October 2023 is stored in the shard "EP_MONITOR_202310". When building the timestamp index, the B+ tree structure is used to sort the timestamp field, the index key is the timestamp value, and the leaf node stores the physical address of the corresponding data item. For example, when querying the environmental protection monitoring data with the timestamp range from 2023-10-01 09:00:00 to 2023-10-01 10:00:00, the corresponding partition in the shard "EP_MONITOR_202310" can be quickly located through the timestamp index, reducing the overhead of full table scanning.

[0024] The storage process of real-time monitoring data and periodically revised data includes data compression and encryption. The data compression uses a lossless compression algorithm, dictionary coding for text fields, and differential coding for numeric fields. For example, the continuous monitoring data sequence of a soil moisture sensor is [45%, 46%, 47%, 46%], and after differential coding, it is stored as [45, +1, +1, -1]. The data encryption uses the AES-256 symmetric encryption algorithm, and the encryption key is distributed according to the business unit identifier. For example, the environmental protection monitoring data is encrypted with the key K_EP_MONITOR, and the land use planning data is encrypted with the key K_LAND_PLAN. The encryption key is rotated regularly through the key management system, and the rotation period is 90 days. The encrypted data generates a checksum through the SHA-256 hash function, and the checksum is stored together with the ciphertext for subsequent data integrity verification.

[0025] When querying data, locate the target data item according to the business unit identifier and timestamp index, and decrypt and decompress it to restore the original data. For example, when querying the vibration monitoring data of a certain engineering equipment from 09:30:00 to 09:35:00 on October 1, 2023, first locate the shard "EQUIP_VIB_202310" through the business unit identifier "EQUIP_VIB_002", and then retrieve the compressed data block within this time period through the timestamp index. After decrypting and decompressing, output the vibration frequency sequence [50Hz, 52Hz, 51Hz]. If the timestamp range of the query request spans multiple data partitions, access each partition in parallel and merge the results to ensure the query efficiency. When accessing in parallel, use multi-threading technology, and the number of threads is dynamically allocated according to the number of CPU cores of the server. For example, an 8-core CPU allocates 8 threads.

[0026] The abnormal data processing mechanism includes automatic re-collection and manual intervention. For data marked as illegal or with a future timestamp, the system automatically triggers a re-collection instruction to obtain data from the sensor or enterprise resource planning system again. If legal data still cannot be obtained after three consecutive re-collections, an abnormal report is generated and pushed to the management terminal for manual verification of the data source configuration or device status. The abnormal report includes the data item identifier, timestamp, abnormal type, and re-collection log. For example, the report number "ERR_20231001_001" records that the timestamp of a certain noise sensor is incorrect due to a clock chip failure. During manual intervention, the operation and maintenance personnel issue device calibration instructions or revise data entry rules through the management terminal. For example, calibrate the sensor clock chip to the current Coordinated Universal Time, or adjust the timestamp generation logic of the approval process in the enterprise resource planning system.

[0027] The data backup and recovery mechanism ensures storage reliability. A full backup is performed at 02:00 every morning, and the backup data is stored in a remote disaster recovery center, and the backup period is reserved for 30 days. An incremental backup is performed once an hour, and only newly added or modified data items are backed up. For example, the full backup file for October 1, 2023 is "BACKUP_20231001_FULL", and the incremental backup file is "BACKUP_20231001_0200_INCR". When recovering data, select the most recent full backup and subsequent incremental backups according to the timestamp range of the backup files, and restore them to the distributed database in sequence.

[0028] S2. Identify the time window differences between real-time monitoring data and periodically revised data based on timestamps, and extract the associated data set with overlapping time windows and inconsistent update cycles. The specific implementation is as follows: When calculating the real-time data window length based on the timestamps of real-time monitoring data, the preset first time threshold and second time threshold are dynamically adjusted according to the business unit type. For example, for the real-time data window length of the project environmental compliance monitoring unit, it is from 24 hours before the collection time point to 1 hour after the collection time point. The first time threshold is 24 hours, and the second time threshold is 1 hour to ensure the timeliness of data on sudden pollution incidents at the construction site; for the real-time data window length of the construction vehicle scheduling monitoring unit, it is from 12 hours before the collection time point to 30 minutes after the collection time point. The first time threshold is 12 hours, and the second time threshold is 30 minutes to adapt to the short-cycle fluctuation characteristics of vehicle scheduling data. The calculation method of the real-time data window length is: the start time of the real-time data window is equal to the collection time point minus the first time threshold, and the end time of the real-time data window is equal to the collection time point plus the second time threshold. For example, if the collection time point of a certain construction vehicle scheduling data is 08:00:00 on October 1, 2023, then the real-time data window range is from 20:00:00 on September 30, 2023 to 08:30:00 on October 1, 2023.

[0029] When calculating the revised data window length based on the timestamps of periodically revised data, the preset third time threshold and fourth time threshold are set according to the type of revised data. For example, for the revised data window length of land use planning revised data, it is from 1 quarter before the effective time to 1 month after the effective time. The third time threshold is 1 quarter, and the fourth time threshold is 1 month to cover the transition period of planning adjustments; for the revised data window length of the demolition progress revised data, it is from 2 weeks before the effective time to 2 weeks after the effective time. The third time threshold is 2 weeks, and the fourth time threshold is 2 weeks to balance the flexibility of coordinating the demolition progress. The calculation method of the revised data window length is: the start time of the revised data window is equal to the effective time minus the third time threshold, and the end time of the revised data window is equal to the effective time plus the fourth time threshold. For example, if the effective time of a certain demolition progress revision is 09:00:00 on September 30, 2023, then the revised data window range is from 09:00:00 on September 16, 2023 to 09:00:00 on October 14, 2023.

[0030] When traversing the time windows of real-time monitoring data and periodically revised data, extract the data items where the time windows overlap. The determination condition for time window overlap is: the start time of the real-time data window is less than or equal to the end time of the revised data window, and the end time of the real-time data window is greater than or equal to the start time of the revised data window. For example, for a certain project, the real-time window of environmental compliance monitoring data is from 10:00:00 on September 30, 2023 to 11:00:00 on October 1, 2023, and the window of land use planning revised data is from 17:00:00 on June 30, 2023 to 17:00:00 on October 30, 2023, and there is an overlap between the two. Further filter out the data items where the ratio of the length of the real-time data window to the length of the revised data window exceeds a preset ratio threshold. The ratio threshold is set according to the business collaboration requirements. For example, in the scenario of environmental compliance and land use planning collaboration, the ratio threshold is 0.1, that is, the length of the real-time window needs to be greater than 10% of the length of the revised window. The length of the real-time window is calculated as the end time of the real-time data window minus the start time of the real-time data window, and the length of the revised window is calculated as the end time of the revised data window minus the start time of the revised data window. For example, the length of the above-mentioned environmental compliance real-time window is 25 hours, and the length of the land use planning revised window is 3 months (about 2160 hours), and the ratio is 25 / 2160≈0.0116, which exceeds the ratio threshold of 0.01, so this data item is retained.

[0031] Based on the business unit identifiers and data item attributes of each data item in the data item set, establish cross-business unit association rules. The business unit identifier follows the format of "department abbreviation_business type_sequence number". For example, "ENV_COMPLIANCE_001" represents the environmental compliance monitoring data of a project, and "LAND_PLAN_003" represents the land use planning data. The data item attributes include data type, project number, geographical coordinates, and revision version number. The association rules include data type matching, geographical coordinate overlap, and project number association. For example, when a data item with a data type of "soil pollution monitoring" is associated with a data item with a data type of "land use nature", it is necessary to meet the conditions that the geographical coordinates fall within the same plot range and the project numbers belong to the same engineering project. The determination method for geographical coordinate overlap is: if the difference in longitude and latitude of the geographical coordinates of the two data items is less than 0.001 degrees, they are regarded as the same plot. For example, the geographical coordinates of a certain soil pollution monitoring data are 39.9042 degrees north latitude and 116.4074 degrees east longitude, and the geographical coordinates of a certain land use planning data are 39.9045 degrees north latitude and 116.4078 degrees east longitude, and the differences in longitude and latitude are 0.0003 degrees and 0.0004 degrees respectively, meeting the overlap conditions.

[0032] Screen data items that meet the association rules and have inconsistent update cycles to generate an associated data set. The determination condition for inconsistent update cycles is that the update cycle of real-time monitored data is less than the update cycle of periodically revised data. For example, if the project environment compliance monitoring data is updated hourly and the land use planning data is updated quarterly, then the update cycles are inconsistent; if the demolition progress monitoring data is updated daily and the land use planning data is updated weekly, it is also considered inconsistent. For data items with the same update cycle (such as both being updated on a daily basis), even if they meet other association rules, they are still excluded from the associated data set. The time unit of the update cycle is unified as hours. For example, daily update is converted to 24 hours, and weekly update is converted to 168 hours to ensure calculation consistency.

[0033] Figure 2 The flowchart for generating the timeliness weights of each data item in the present invention is given. In S3, the timestamps of each data item in the associated data set are dynamically aligned, and the timeliness weights of each data item are generated based on the process topology characteristics and information order characteristics of the multi-department approval process. The specific implementation is as follows: When constructing the collaborative path topology network of the multi-department approval process, the nodes represent the approval departments, including the land use planning department, the demolition coordination department, and the compliance review department, and the edges represent the cross-department data interaction paths. The determination basis for the cross-department data interaction path is the data transfer record between departments in the approval process. For example, if the land use planning department sends plot boundary data to the demolition coordination department, there is an edge between the two departments. The edge weight is defined as the data interaction frequency. For example, if the land use planning department and the demolition coordination department interact 100 times in historical tasks, the edge weight is 100. The calculation method for the shortest approval path is: the sum of the edge weights in the path from the start node to the end node is the smallest. When calculating the betweenness centrality of each node, first count the number of paths passing through this node in all the shortest approval paths, and then divide it by the total number of all the shortest approval paths to obtain the betweenness centrality value. For example, if the land use planning department passes through this node in 80 out of 100 shortest approval paths, the betweenness centrality is 0.8. The topological vulnerability is defined as the reciprocal of the betweenness centrality value. For example, when the betweenness centrality is 0.8, the topological vulnerability is 1.25, indicating that the higher the criticality of this node in the collaborative path, the lower its vulnerability.

[0034] When parsing the historical operation logs of the approval path, the number of decision divergence events is counted as the number of times of opinion conflicts generated by the same approval node in different historical tasks. For example, the compliance review department rejected the land use planning application 3 times due to environmental protection standard disputes in 10 historical approval tasks, and the number of decision divergence events is 3. The number of redundant operation times is counted as the number of times of repeated submission and repeated review in the approval path. For example, the demolition and relocation coordination department required the resubmission of data due to incomplete materials, and the data was resubmitted 3 times in the same task, then the number of redundant operation times is 3. When calculating the entropy increase dissipation value based on the information entropy formula, the decision divergence probability is calculated as the number of decision divergence events divided by the total number of historical tasks, and the redundant operation probability is calculated as the number of redundant operation times divided by the total number of operations. For example, the total number of historical tasks of a certain approval node is 10 times, and the number of decision divergence events is 3 times, then the decision divergence probability is 0.3; the total number of operations is 50 times, and the number of redundant operation times is 3 times, then the redundant operation probability is 0.06. The entropy increase dissipation value is calculated as the negative of the decision divergence probability multiplied by the logarithm of the decision divergence probability, minus the redundant operation probability multiplied by the logarithm of the redundant operation probability.

[0035] The calculation formula for the entropy increase dissipation value is as follows:

[0036] Among them, represents the entropy increase dissipation value; is the decision divergence probability (the ratio of the number of decision divergence events to the total number of historical approval tasks); is the redundant operation probability (the ratio of the number of redundant operation times to the total number of operations).

[0037] According to the quantization results of topological vulnerability and entropy increase dissipation value, the timestamp offset of each data item in the associated data set is dynamically weighted and corrected. The timestamp offset is defined as the difference between the actual timestamp of the data item and the theoretical alignment timestamp, and the theoretical alignment timestamp is generated according to the preset time nodes of the approval process. For example, the theoretical alignment timestamp of a land use planning data item is 10:00:00 on October 1, 2023, and the actual timestamp is 10:05:00 on October 1, 2023, then the offset is 5 minutes. The dynamic weighted correction formula is: the corrected offset is equal to the original offset multiplied by the topological vulnerability coefficient and then divided by the entropy increase dissipation coefficient. The value of the topological vulnerability coefficient is the normalized result of the topological vulnerability value, and the normalization method is to linearly map the topological vulnerability value to the interval from 0 to 1. For example, if the topological vulnerability range is from 1 to 5, then the topological vulnerability value 1 is mapped to 0, 5 is mapped to 1, and the intermediate values are calculated according to the linear ratio. The value of the entropy increase dissipation coefficient is the normalized result of the entropy increase dissipation value, and the normalization method is to linearly map the entropy increase dissipation value to the interval from 1 to 10. For example, if the entropy increase dissipation value range is from 0 to 2.5, then the entropy increase dissipation value 0 is mapped to 1, 2.5 is mapped to 10, and the intermediate values are calculated according to the linear ratio. For example, if the topological vulnerability of a data item is 1.25 (0.8 after normalization) and the entropy increase dissipation value is 2.5 (10 after normalization), then the corrected offset is 5 minutes * (0.8 / 10) = 0.4 minutes.

[0038] Generate the timeliness weight of each data item based on the corrected timestamp offset. The timeliness weight calculation formula is: the timeliness weight is equal to the baseline weight divided by (1 + the corrected offset), and the baseline weight is preset according to the business priority. For example, the baseline weight of a land use planning data item is 1.0, and the baseline weight of a compliance review data item is 0.8. For example, if the corrected offset of a land use planning data item is 0.4 minutes, then the timeliness weight is 1.0 / (1 + 0.4) = 0.714. The negative correlation between the weight and the topological vulnerability is reflected as: when the topological vulnerability decreases (i.e., the node criticality increases), the weight attenuation of the same offset decreases. For example, when the topological vulnerability decreases from 1.25 to 1.0 (0.8 → 0.6 after normalization), the corrected offset decreases and the weight increases; the positive correlation between the weight and the entropy increase dissipation value is reflected as: when the entropy increase dissipation value increases (i.e., the confusion degree of the approval path increases), the weight attenuation of the same offset increases. For example, when the entropy increase dissipation value increases from 2.0 to 2.5 (8 → 10 after normalization), the corrected offset increases and the weight decreases.

[0039] The exception handling mechanism includes the default weight assignment when the node is isolated and the fault-tolerant processing when the entropy dissipation value returns to zero. If a node has no edge connection (isolated node) in the collaborative path topology network, its topological vulnerability is set to the maximum value (for example, 10) by default to avoid missing key nodes. If the entropy dissipation value is calculated to be zero (no decision divergence and redundant operations), the entropy dissipation coefficient is set to 1 by default to prevent division by zero errors. Hardware dependencies include server clusters that support parallel computing to accelerate topology network construction and entropy value calculation. The server cluster is configured with at least 16-core CPUs and 64GB of memory to meet high concurrent computing requirements; software dependencies are the time series analysis library Pandas 1.3.0 and above, which is used to efficiently process the timestamp data of historical operation logs, and NetworkX 2.6.0 and above, which is used to build and calculate collaborative path topology networks.

[0040] Boundary conditions cover extreme offsets and abnormal inputs. If the corrected offset exceeds the preset maximum tolerance value (for example, 30 minutes), it is forced to be set to the maximum tolerance value to prevent the data from being mistakenly ignored due to too small a weight. If the calculation result of the entropy increase dissipation value is negative (due to probability calculation error), the absolute value is taken and recalculated. In the data preprocessing stage, historical operation logs are denoised to remove test data and invalid operation records. For example, task logs marked as "debugging" are not included in the statistics, and data items that are repeated more than 10 times in a single task are considered abnormal inputs and trigger manual review.

[0041] The parameter adjustment range and basis are determined through historical data analysis and business rules. The setting basis of the topological vulnerability normalization range of 0 to 1 is the statistical results of the node criticality distribution in the historical approval tasks, ensuring that 90% of the node topological vulnerability values ​​fall within the range of 1 to 5. The setting basis of the entropy increase dissipation value normalization range of 1 to 10 is the chaos tolerance threshold of different approval scenarios, for example, the upper tolerance limit of regular projects is 2.5, and the emergency projects are temporarily adjusted to 3.0. The basis for setting the benchmark weight is the business unit priority table, for example, the land planning department has a high priority (benchmark weight 1.0), the compliance review department has a medium priority (benchmark weight 0.8), and the acquisition and demolition coordination department has a low priority (benchmark weight 0.6).

[0042] S4. Based on the timeliness weight, a spatiotemporal propagation path map is constructed. After screening the propagation paths, the delayed and real-time data items are dynamically spliced ​​along the high-weight paths to generate a timeliness-aligned decision input data set. The specific implementation is as follows: When calculating the path weight of each data item in the associated data set based on the timeliness weight, the path weight is defined as the product of the timeliness weight and the timestamp adjacency of the data item. The timestamp adjacency is obtained by calculating the reciprocal of the difference between the timestamps of two data items. The formula is that the timestamp adjacency is equal to 1 divided by (the timestamp difference + 1), where the unit of the timestamp difference is minutes, and adding 1 is used to avoid division by zero errors. For example, if the timestamp of data item A is 10:00:00 on October 1, 2023, and the timestamp of data item B is 10:05:00 on October 1, 2023, the timestamp difference is 5 minutes, and the timestamp adjacency is 1 / (5 + 1) = 0.1667. If the timeliness weight of data item A is 0.8, then the path weight is 0.8×0.1667≈0.133. The calculation results of all path weights are stored in the weight matrix. The rows and columns of the matrix respectively correspond to different data items in the associated data set, and the missing path weights in the matrix are default filled with zeros.

[0043] When constructing the spatio-temporal propagation path graph, the nodes represent data items, and the node attributes include data item identifiers, timestamps, and geographical coordinates; the edges represent the spatio-temporal dependence relationships between data items. The determination condition of the spatio-temporal dependence relationship is that the geographical coordinate difference between data items is less than the preset distance threshold and the timestamp difference is less than the preset time correlation window. The geographical coordinate difference threshold is set according to the business scenario. For example, it is set to 50 meters in the land expropriation and demolition scenario and 500 meters in the land use planning scenario. The time correlation window is set according to the data update frequency. For example, it is set to 30 minutes for real-time monitoring data and 7 days for periodically revised data. The edge weight is the normalized value of the path weight, and the normalization method is to linearly map the path weight to the interval from 0 to 1. The formula is that the normalized edge weight is equal to (the original path weight - the minimum path weight) divided by (the maximum path weight - the minimum path weight). For example, if the path weight range of a batch of data items is from 0 to 0.5, then the minimum path weight is 0, the maximum path weight is 0.5, and the path weight 0.2 is normalized to 0.4.

[0044] When screening high-weight paths according to edge weights, the high-weight paths need to meet the conditions that the cumulative sum of edge weights exceeds the preset path threshold and the path length is less than the preset hop threshold. The path threshold is set according to the weight distribution of valid paths in historical data. For example, after analyzing the data of the past year, the cumulative sum of weights of 90% of the valid paths is in the range of 1 to 5, and the path threshold is set to 3. The hop threshold is used to limit the maximum number of nodes in the path. For example, it is set to 5 to avoid the complication of decision-making logic caused by too long paths. When traversing the spatio-temporal propagation path graph, the depth-first search algorithm is used to start from the starting node, and the starting node selection rule is the data item with the highest timeliness weight. The edge weights are accumulated and the path length is recorded. When the cumulative weight exceeds 3 and the path length is less than 5, it is marked as a high-weight path. For example, a certain path contains the node sequence A→B→C, and the edge weights are 0.5, 0.8, and 0.6 respectively. The cumulative weight is 1.9 and the path length is 3, which does not reach the threshold and is not included in the high-weight paths.

[0045] When dynamically splicing lagged data items and real-time data items along high-weight paths according to the timestamp adjacency, the splicing rule is that the difference between the timestamps of the lagged data item and the real-time data item is less than the preset alignment tolerance. The alignment tolerance is dynamically adjusted according to the business scenario. For example, it is set to 10 minutes for regular projects and 5 minutes for emergency projects. The adjustment needs to be approved by the project management terminal and the operation log is recorded. For example, the timestamp of a certain lagged data item is 09:50:00 on October 1, 2023, and the timestamp of the real-time data item is 10:00:00 on October 1, 2023. The timestamp difference is 10 minutes. When the alignment tolerance is 10 minutes, they are merged into the same decision input data unit. The merged data unit inherits the geographical coordinates and the latest attributes of the real-time data item, and at the same time retains the historical revision records of the lagged data item. If multiple data items in the same path meet the alignment tolerance, they are merged according to the principle of the nearest timestamp first. For example, data item C (with a difference of 8 minutes) is merged with the real-time data item prior to data item B (with a difference of 12 minutes).

[0046] When generating a decision input data set with timeliness alignment in the order of the spatio-temporal propagation path for the spliced data units, the data units are arranged in ascending order of the timestamps of the nodes in the path. For example, if a high-weight path contains nodes A (timestamp 10:00), B (timestamp 10:05), and C (timestamp 10:10), the generated data set is stored in the order of A→B→C. The data set storage format is a time series structure, and each data unit contains a timestamp, geographical coordinates, data value, and associated path identifier. The data unit identifier generation rule is "path ID_timestamp_geographical coordinate hash value", for example, "PATH_001_202310011000_39.9042N116.4074E", to ensure uniqueness. The geographical coordinate hash value is generated by rounding the latitude and longitude to four decimal places and concatenating them. For example, 39.9042 degrees north latitude and 116.4074 degrees east longitude are hashed to "39.9042N116.4074E".

[0047] The exception handling mechanism includes filtering of invalid paths and forced splitting when the alignment tolerance is exceeded. If the edge weight of a path is negative or zero (due to calculation errors), the path is automatically filtered and an exception log is recorded. The log format is "exception path ID_exception weight value_timestamp". If the timestamp difference of data items exceeds the alignment tolerance, even if they belong to the same high-weight path, they are still split into independent data units and marked as unaligned, and at the same time, an alarm notification is triggered to the project management terminal. For example, when the timestamp difference is 15 minutes and the tolerance is 10 minutes, it is split into two data units and an alarm message "unaligned data items: ID1234 and ID5678" is pushed. Hardware dependencies include the distributed computing framework Apache Spark 3.0 and above for efficiently processing large-scale spatio-temporal propagation path graphs; software dependencies are the graph database Neo4j 4.0 and above for storing and traversing graph nodes and edges, and the numerical calculation library NumPy 1.20 and above for matrix operations and normalization processing.

[0048] Boundary conditions cover extreme geographical coordinates and timestamp inputs. If the difference in geographical coordinates exceeds the maximum range of the Earth's longitude and latitude (longitude from -180 to 180 degrees, latitude from -90 to 90 degrees), it is determined as illegal input and discarded, and the source IP address of the anomaly is recorded simultaneously. If the timestamp is earlier than the earliest time supported by the system (such as 1970-01-01 00:00:00) or later than the latest time (such as 2038-01-19 03:14:07), it is automatically truncated to the time boundary value supported by the system. For example, the timestamp 1969-12-31 23:59:59 is truncated to 1970-01-01 00:00:00. The parameter adjustment is verified based on historical data. For example, the setting of the path threshold 3 enables 85% of the high-weight paths to cover key decision-making data, and the setting of the hop count threshold 5 controls the path search time within 1 second. The verification data is sourced from the land transaction management records of the enterprise group in the past three years.

[0049] It is worth noting that the timeliness weight generated by S3 provides a dynamic quantification basis for the path construction of S4, forming a closed-loop technical link from data feature analysis to decision input. S3 calculates the timeliness weight through the fusion of process topology features (such as node betweenness centrality) and information order features (such as decision divergence entropy value), quantitatively reflecting the criticality and credibility of data items in the approval process; S4 then selects high-weight spatio-temporal propagation paths based on this weight, and fuses multi-source data through path pruning and dynamic splicing mechanisms. This association enables the collaborative action of the timeliness evaluation of data items and spatio-temporal dependence relationships. For example, data items with high timeliness weights preferentially form the main propagation paths to ensure the integrity and timeliness of core decision-making bases, while low-weight data is supplemented along the branch paths as auxiliary information to avoid redundant interference. At the same time, the path optimization result of S4 can be reversely fed back to the weight generation logic of S3 (such as the decay of path weights triggering the update of weight coefficients), forming a dynamic adaptive data alignment mechanism. This closed-loop association significantly improves the flexibility and decision-making accuracy of cross-departmental data collaboration. Especially when dealing with sudden land affairs, it can quickly identify key data paths and avoid temporal conflicts, thereby reducing decision-making risks caused by information lag or logical confusion and supporting efficient resource management and control in complex scenarios.

[0050] S5. Call the preset multi-department collaborative approval rule library according to the decision input data set to generate an approval path plan and a risk warning list that match the timeliness weight. The specific implementation is as follows: When analyzing the timeliness weight and data item attributes in the decision-making input dataset, the rule priority mapping table is a two-dimensional matrix. The rows represent the timeliness weight intervals, the columns represent the data item attribute types, and the cell content is the corresponding approval rule number. The timeliness weight intervals are divided at a step of 0.1, such as [0.0, 0.1), [0.1, 0.2),..., [0.9, 1.0]. The data item attribute types include data types (real-time monitoring, periodic revision), geographical coordinate ranges (plot number, administrative division), and project phases (project establishment, construction, acceptance). For example, the cell where the timeliness weight interval [0.7, 0.8) intersects with the attribute type "real-time monitoring - Plot A - project establishment phase" stores the rule number R_001, indicating a high-priority project establishment approval rule. When matching the applicable approval rules in the multi-department collaborative approval rule library, the rule priority mapping table is queried based on the timeliness weight and attribute type of the data item. If there are multiple matching rules, the rule with the highest upper limit of the timeliness weight interval is selected. For example, if the timeliness weight of a data item is 0.75 and it matches R_001 (interval [0.7, 0.8)) and R_002 (interval [0.6, 0.7)), R_001 is finally selected.

[0051] When generating the initial approval path plan according to the applicable approval rules, the initial approval path plan includes the sequence of approval department nodes and the approval time limit constraints between the nodes. The sequence of department nodes is generated according to the approval order in the rule library. For example, the order defined by rule R_001 is "land use planning department → land expropriation and demolition coordination department → compliance review department → financial audit department". The approval time limit constraints between the nodes are set according to the department's processing capacity and rule priority. For example, the benchmark time limit of the land use planning department is 3 working days, the land expropriation and demolition coordination department is 5 working days, the compliance review department is 2 working days, and the financial audit department is 1 working day. The calculation method of the time limit constraint is the benchmark time limit multiplied by (1 - timeliness weight). For example, if the benchmark time limit is 3 days and the timeliness weight is 0.8, the actual time limit is 3*(1 - 0.8) = 0.6 days (about 14 hours). The time limit calculation result is rounded to the nearest whole hour. For example, 0.6 days is converted to 14 hours, and 0.7 days is converted to 17 hours.

[0052] When detecting the timeliness weight decay rate of the node sequence in the initial approval path plan, the decay rate is calculated as the decline amplitude of the timeliness weight per unit time. The unit time is set according to the approval stage. It is calculated by the hour in the project establishment stage, by the day in the construction stage, and by the week in the acceptance stage. For example, in the project establishment stage, the timeliness weight of a certain data item drops from 0.8 to 0.6 within 24 hours, and the decay rate is (0.8 - 0.6) / 24 = 0.0083 / hour. The preset risk threshold is set according to the decay rate distribution of high-risk events in historical data. For example, after analyzing the data of the past two years, the decay rate of 95% of high-risk events is 0.005 / hour, and the risk threshold is set to 0.005 / hour. If the detected decay rate exceeds the threshold, a risk warning event is triggered and the risk node is marked. For example, the decay rate of a node in a demolition and relocation coordination department is 0.01 / hour, exceeding the threshold of 0.005 / hour, it is marked as a risk node and the risk type is recorded as "approval delay" and the risk level is "high".

[0053] When adjusting the node priority of the approval path plan according to the risk warning event, the optimization rules include moving the risk node forward, reallocating resources, and adjusting the time limit elasticity. Moving the risk node forward means moving the high-risk node to a more forward position in the approval path. For example, the demolition and relocation coordination department was originally the second node and is adjusted to the first node after adjustment. Reallocating resources means allocating additional approval resources to high-risk nodes. For example, increasing the number of approval personnel or extending the daily working hours. The resource allocation coefficient is set according to the risk level. For example, the high-risk level coefficient is 1.5 (i.e., the resources are increased by 50%). Adjusting the time limit elasticity means extending the time limit constraint for high-risk nodes. For example, the original time limit is 14 hours, and when the risk level is high, it is extended by 20%, that is, 14 * 1.2 = 16.8 hours (about 17 hours). After generating the optimized approval path plan, the risk warning list includes the risk node identifier, risk type, and mitigation strategy. For example, the risk node identifier is "demolition_and_relocation_002", the risk type is "approval delay", and the mitigation strategy is "add 2 coordination specialists and enable the fast approval channel", and the deadline for strategy execution is within 2 hours after the risk is triggered.

[0054] When outputting the optimized approval path plan and the risk warning list to the approval execution terminal in the timestamp order of the approval department node sequence, the timestamp order is arranged in ascending order according to the deadline of the node processing time limit. For example, the optimized approval path plan is "Demolition and Relocation Coordination Department (deadline: 2023-10-05 18:00) → Land Use Planning Department (deadline: 2023-10-06 09:00) → Compliance Review Department (deadline: 2023-10-07 12:00) → Financial Audit Department (deadline: 2023-10-08 09:00)", then it is output to the terminal in the order of the deadline. The output format is a structured JSON file, including the node name, time limit, risk identification, and mitigation strategy. For example, the JSON field of the node "Demolition and Relocation Coordination_002" is {"node": "Demolition and Relocation Coordination_002", "deadline": "2023-10-05 18:00", "risk_id": "R002", "action": "Add 2 commissioners", "action_deadline": "2023-10-05 20:00"}. The terminal device needs to support JSON parsing and task reminder functions. For example, push tasks to the to-do list of the approval system of the department head through a message queue (such as RabbitMQ).

[0055] The exception handling mechanism includes the default rule fallback when the rule matching fails and the retry logic for the abnormal decay rate calculation. If there is no matching item in the rule priority mapping table, the default rule R_DEFAULT is enabled. The rule content is "All departments approve according to the benchmark time limit", and the benchmark time limit is set according to the internal management manual of the enterprise group. For example, the default time limit for the land use planning department is 3 days, and the demolition and relocation coordination department is 5 days. If the decay rate calculation fails due to data anomalies (such as the weight suddenly becoming negative), re-extract the last 3 weight values for moving average calculation. For example, if the weight sequence is [0.8, 0.6, 0.4], the moving average value is (0.8 + 0.6 + 0.4) / 3 = 0.6, and the decay rate calculation is (0.8 - 0.6) / 24 = 0.0083 / hour. The hardware dependency is a relational database MySQL 8.0 or above that supports high-concurrency queries, which is used to store the rule library and mapping table. The database cluster configuration has at least 3 nodes to achieve disaster recovery backup; the software dependency is a process engine Activiti 7.0 or above, which is used to execute the dynamic adjustment of the approval path plan, and a message middleware RabbitMQ 3.9 or above, which is used for task distribution and status synchronization.

[0056] Boundary condition coverage invalid rule number and extremely long approval time limit. If the rule number does not exist in the rule library, discard the rule and record an exception log in the format of "invalid rule ID_data item identifier_timestamp", e.g., "invalid rule ID_R999_Plot A_20231001". If the calculated result of the approval time limit exceeds the maximum value supported by the system (such as 30 days), it is forcibly truncated to 30 days and an alarm notification is triggered, and the alarm message is pushed to the system administrator's terminal. The parameter adjustment basis is deduced by the historical approval success rate. For example, the setting of the risk threshold of 0.005 / hour enables 90% of high-risk events to be identified in a timely manner. The verification data is sourced from the records of 500 land transaction projects of the enterprise group in the past two years. Abnormal projects (such as projects suspended due to policy changes) are excluded during data cleaning.

[0057] S6. Execute the approval path plan and detect the triggering conditions of the risk warning list in real time at the approval process nodes to dynamically adjust the priority order of the subsequent approval paths. The specific implementation is as follows: When distributing the approval department node sequence in the optimized approval path plan to the corresponding approval execution terminals, the approval department node sequence is pushed to the to-do task list of each approval department through the message middleware in the priority order. For example, if the optimized approval path plan contains the node sequence "Demolition and Relocation Coordination Department → Land Use Planning Department → Compliance Review Department", the tasks are sent to the approval system interface of the Demolition and Relocation Coordination Department in sequence through the message queue, and the task status is marked as "pending". The message middleware uses RabbitMQ version 3.9 and above, which supports message persistence and transaction rollback to ensure that task distribution is not lost. When monitoring the approval status of each node and the triggering conditions of the risk warning list, the approval status includes "pending", "processing", "completed", "timed out", and the status data is synchronized every 5 minutes through the polling interface. The triggering conditions of the risk warning list include risk node identifier, risk type (such as approval delay, resource shortage), and risk level (high, medium, low). For example, if the approval task of the Demolition and Relocation Coordination Department triggers the "approval delay" risk and the risk level is high, it is detected that the node matches the identifier in the risk warning list.

[0058] When the risk node identifier in the risk warning list is detected to match the current approval node, extract the risk type and mitigation strategy. For example, if the risk node identifier is "Demolition and Relocation Coordination_002", the risk type is "Approval Delay", the mitigation strategy is "Add 2 coordination specialists", and the risk level is high. When calculating the priority adjustment coefficient of the current node according to the risk level, the mapping rule between the risk level and the adjustment coefficient is: high risk corresponds to a coefficient of 2.0, medium risk 1.5, and low risk 1.2. The mapping rule is derived from the statistical analysis of the risk level and processing efficiency in historical data. For example, when the processing efficiency of high-risk tasks increased by 2 times in the past year, the priority coefficient was 2.0. The calculation method is: the priority adjustment coefficient is equal to the baseline coefficient multiplied by the risk level weight. The baseline coefficient is 1.0, and the risk level weight is obtained by looking up a table. For example, if the high-risk level weight is 2.0, the adjustment coefficient is 1.0×2.0 = 2.0.

[0059] When reordering the subsequent approval department node sequence based on the priority adjustment coefficient, the sorting rule for the subsequent node sequence is: give priority to processing nodes with a high adjustment coefficient, and if the coefficients are the same, arrange them in the original order. For example, the original subsequent node sequence is "Land Use Planning Department → Compliance Review Department → Financial Audit Department". If the Compliance Review Department needs to be preposed due to a high risk level (coefficient 2.0), the adjusted sequence is "Compliance Review Department → Land Use Planning Department → Financial Audit Department". After generating the dynamically adjusted approval path plan, push the adjusted task list to the relevant approval execution terminals. For example, push an urgent task to the Compliance Review Department through a message middleware, mark the task status as "Urgent", and attach the mitigation strategy "Add 2 coordination specialists". When pushing the task, it is necessary to verify the processing capacity of the approval terminal. If the current load of the terminal exceeds 80% (such as the number of tasks to be processed is greater than 10), it will be automatically diverted to the standby terminal.

[0060] When recording the priority adjustment event in the approval process log, the log content includes the risk node identifier, adjustment timestamp, comparison data between the original path plan and the adjusted path plan. For example, the log entry is "The risk node Demolition and Relocation Coordination_002 triggered a priority adjustment at 14:30 on October 5, 2023. The original sequence [Demolition and Relocation Coordination → Land Use Planning → Compliance Review], the adjusted sequence [Compliance Review → Demolition and Relocation Coordination → Land Use Planning]". The log is stored in a time series database InfluxDB version 2.0 and above, and is indexed by timestamp to support subsequent auditing and backtracking. When the log recording fails (such as database connection timeout), the data is temporarily stored in a local cache file in CSV format, and retried for writing every 5 minutes until successful.

[0061] When updating the risk status in the risk warning list according to the adjusted path plan, if the risk mitigation strategy has been executed and the risk level has dropped below the threshold, the corresponding risk node identifier is removed. For example, after adding 2 commissioners to the demolition and relocation coordination department, the approval task is completed within 4 hours, and the risk level drops from high to low, and the system automatically removes the risk identifier of "demolition and relocation coordination_002". The risk level threshold is dynamically set according to the business scenario. For example, the low threshold of the risk level for regular projects is 0.3, and for urgent projects is 0.5. If the risk level has not dropped below the threshold, the risk type is upgraded and a new mitigation strategy is generated. For example, when the risk of insufficient resources is not mitigated, a new strategy "apply for cross-departmental collaboration support" is added and the risk level is raised to urgent, and at the same time, a secondary alarm is triggered to the superior management department.

[0062] The exception handling mechanism includes task reassignment when a node fails and data recovery when logging fails. If the approval execution terminal cannot receive tasks (such as network interruption exceeding 3 retries), task reassignment is triggered to the backup terminal. For example, when the main system of the demolition and relocation coordination department fails, the task is automatically forwarded to the backup approval team terminal, and the backup terminal needs to confirm receipt within 30 minutes. If logging fails (such as database write error), it is temporarily cached to a local file, and the file naming rule is "log date_terminal IP_sequence number". After the database is restored, it is resynchronized through an incremental synchronization tool (such as Debezium). The hardware dependency is a highly available message middleware cluster (such as RabbitMQ mirror queue) to ensure the fault tolerance of task distribution and status synchronization; the software dependency is the approval process engine Activiti 7.0 and above, which supports dynamic task adjustment and status rollback, and the log analysis tool ELK (Elasticsearch, Logstash, Kibana) stack for log retrieval and visualization.

[0063] The boundary conditions cover extreme risk levels and abnormal node sequences. If the risk level calculation exceeds the preset range (such as the coefficient 2.5 exceeding the maximum value 2.0), it is forced to be truncated to the maximum value and a system alarm is triggered. The alarm message push format is JSON, including the reason for the exception and the handling suggestion. If a circular dependency appears in the adjusted node sequence (such as A→B→A), the adjustment is automatically terminated and rolled back to the original sequence, and the reason for the exception is recorded in the log. The reason for the exception is classified as "logical conflict". The parameter adjustment basis is verified through historical response efficiency. For example, the setting of the priority coefficient 2.0 shortens the processing time of 90% of high-risk tasks by 50%. The verification data is based on the processing records of 200 high-risk tasks of the enterprise group in the past year.

[0064] In the cross - department land affairs decision - making scenario, traditional technologies usually rely on matching a static rule base with a fixed time window, making it difficult to solve the problem of decision - making temporal chaos caused by the timeliness conflicts of multi - source data. In this embodiment, by dynamically aligning timestamps and fusing process topology features to generate timeliness weights, replacing manual rule setting, the timeliness quantification of data items can be dynamically adjusted according to the stability of the approval process and the degree of information chaos. Further, through pruning and dynamic splicing of the spatio - temporal propagation path map, the non - linear fusion of lagged data and real - time data is achieved, avoiding information distortion caused by traditional linear interpolation. In addition, the dynamic priority adjustment of the risk warning list and the approval path forms a closed - loop feedback. By real - time detecting the risk level and reversely optimizing the path, the defect that the traditional static process cannot adapt to sudden risks is solved. Each step constructs a complete technical chain for cross - department land affairs decision - making through multi - dimensional feature fusion, non - linear data fusion and dynamic feedback mechanisms.

[0065] The calculations involved in the embodiments are all dimensionless numerical calculations. The preset parameters and threshold selections in the calculations are set by those skilled in the art according to the actual situation.

[0066] It should be noted that the present invention can be deployed on the device itself to achieve embedded applications, or can run on a PC with a user interface or other terminals, so as to meet various hardware environments and usage requirements.

[0067] The above - mentioned embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above - mentioned embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general - purpose computer, a special - purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer - readable storage medium, or transmitted from one computer - readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wire (such as infrared, wireless, microwave, etc.). The computer - readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center containing one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid - state drive.

[0068] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0069] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or modules can be in electrical, mechanical, or other forms.

[0070] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0071] In addition, in each embodiment of the present application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0072] If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0073] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0074] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent decision-making support method for land affairs oriented to cross-departmental collaboration, characterized in that It includes the following steps: S1. Obtain the land transaction data of multiple business units within the enterprise group. The land transaction data includes real-time monitoring data and periodic revision data, and attach timestamps to each data item; S2. Identify the time window differences between the real-time monitoring data and the periodic revision data based on the timestamps, and extract the associated data set where there is an overlap in the time window and the update cycles are inconsistent; S3. Dynamically align the timestamps of each data item in the associated data set, and generate the timeliness weights of each data item based on the process topology characteristics and information order characteristics of the multi-department approval process; S4. Construct a spatio-temporal propagation path map based on the timeliness weights, filter the propagation paths, and then dynamically splice the lagged and real-time data items along the high-weight paths to generate a decision input data set with timeliness alignment; S5. Call the preset multi-department collaborative approval rule library according to the decision input data set, and generate an approval path plan and a risk warning list that match the timeliness weights; S6. Execute the approval path plan, and detect the triggering conditions of the risk warning list in real time at the approval process nodes to dynamically adjust the priority order of the subsequent approval paths.

2. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, wherein S1 includes: Obtain the real-time monitoring data and mark the acquisition time point as the reference time of the timestamp; Obtain the periodic revision data. The periodic revision data is batch exported by the enterprise resource planning system according to a preset cycle, and timestamps are attached according to the effective time in the data revision record; Verify the legality of the timestamps of the real-time monitoring data and the periodic revision data; Classify and store the verified land transaction data into a distributed database according to the business unit identifier, and establish a timestamp index to support subsequent data queries.

3. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, characterized in that S2 It includes: Calculate the length of the real-time data window according to the timestamp of the real-time monitoring data; Calculate the length of the revision data window according to the timestamp of the periodic revision data; Traverse the time windows of the real-time monitoring data and the periodic revision data, and extract the data item set where the time windows overlap and the ratio of the length of the real-time data window to the length of the revision data window exceeds a preset ratio threshold; Based on the business unit identifier and data item attributes of each data item in the data item set, establish cross-business unit association rules, and filter the data items that meet the association rules and have inconsistent update cycles to generate an associated data set.

4. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 3, wherein The length of the real-time data window is from pushing a preset first time threshold forward from the acquisition time point to pushing a preset second time threshold backward; the length of the revision data window is from pushing a preset third time threshold forward from the effective time to pushing a preset fourth time threshold backward.

5. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, wherein S3 includes: Construct a collaborative path topology network for the multi-department approval process. The nodes represent the approval departments, and the edges represent the cross-department data interaction paths. Calculate the betweenness centrality of each node as the topological vulnerability in the process topology characteristics; Parse the historical operation logs of the approval paths, count the number of decision divergence events and the number of redundant operation times of each approval node, and calculate the entropy increase dissipation value of the approval path based on the information entropy formula as the information order characteristics; According to the quantization results of the topological vulnerability and the entropy increase dissipation value, dynamically weight and correct the timestamp offsets of each data item in the associated data set; Generate the timeliness weights of each data item based on the corrected timestamp offset. The timeliness weight is negatively correlated with the topological vulnerability and positively correlated with the entropy increase dissipation value.

6. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, characterized in that S4 Including: Calculate the path weights of each data item in the associated data set based on the timeliness weights. The path weight is the product of the timeliness weight and the timestamp adjacency of the data item; Construct a spatio-temporal propagation path graph, where the nodes represent data items and the edges represent the spatio-temporal dependence relationships between data items. The edge weight is the normalized value of the path weight; Screen high-weight paths according to the edge weights. A high-weight path is defined as a continuous node sequence whose cumulative sum of edge weights exceeds a preset path threshold and whose path length is less than a preset hop threshold; Dynamically splice the lagging data items and the real-time data items along the high-weight paths according to the timestamp adjacency; After splicing, generate a decision input data set with timeliness alignment in the order of the spatio-temporal propagation paths.

7. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 6, characterized in that The splicing rule is that when the difference between the timestamps of the lagging data item and the real-time data item is less than a preset alignment tolerance, they are merged into the same decision input data unit.

8. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, wherein S5 Including: Analyze the timeliness weights and data item attributes in the decision input data set, and match the applicable approval rules in the multi-department collaborative approval rule library based on the rule priority mapping table; Generate an initial approval path plan according to the applicable approval rules. The initial approval path plan includes a sequence of approval department nodes and the approval time limit constraints between nodes; Detect the timeliness weight decay rate of the node sequence in the initial approval path plan. If the timeliness weight decay rate exceeds a preset risk threshold, trigger a risk warning event and mark the risk nodes; Adjust the node priorities of the approval path plan according to the risk warning event to generate an optimized approval path plan and a risk warning list; Output the optimized approval path plan and the risk warning list to the approval execution terminal in the order of the timestamps of the approval department node sequence.

9. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 8, characterized in that The risk warning list includes risk node identifiers, risk types, and mitigation strategies.

10. The intelligent decision-making support method for land affairs oriented to cross-departmental collaboration according to claim 1, characterized in that S6 Including: Distribute the sequence of approval department nodes in the optimized approval path plan to the corresponding approval execution terminals, and monitor the approval status of each node and the triggering conditions of the risk warning list; When it is detected that the risk node identifier in the risk warning list matches the current approval node, extract the risk type and mitigation strategy, and calculate the priority adjustment coefficient of the current node according to the risk level; Reorder the subsequent approval department node sequence based on the priority adjustment coefficient to generate a dynamically adjusted approval path plan, and push the adjusted task list to the relevant approval execution terminals; Record the priority adjustment event in the approval process log, including the risk node identifier, the adjustment timestamp, the comparison data between the original path plan and the adjusted path plan; Update the risk status in the risk warning list according to the adjusted path plan. If the risk mitigation strategy has been executed and the risk level has dropped below the threshold, remove the corresponding risk node identifier.

Images