An RPA robot data synchronization method based on big data analysis

Abstract

The application discloses a kind of RPA robot data synchronization methods based on big data analysis, comprising the following steps: collecting and pre-processing the operation monitoring data in multiple business systems;Extract data synchronization features, construct data synchronization feature vector set;Data synchronization feature vector set is input into improved TimeXer model, and output data synchronization strategy parameter set;Build RPA robot execution task set;Establish data field mapping relationship and execute data mapping processing, generate target synchronization data;Target synchronization data is checked, and synchronization check result is generated and output synchronization result mark;Collect synchronization state feedback data, and update model parameters to realize data synchronization strategy self-adaptive optimization.The application utilizes improved data synchronization strategy modeling method to optimize automated data synchronization process, realizes cross-system data synchronization, with the advantages of strategy self-adaptive optimization and high synchronization stability.

Description

Technical Field

[0001] This invention relates to the field of data synchronization processing technology, and in particular to a data synchronization method for RPA robots based on big data analysis. Background Technology

[0002] With the increasing number of enterprise information systems, the need for data interaction between different business systems is becoming more frequent. Utilizing RPA robots to perform data synchronization tasks has become an important technical means to improve the level of business automation. Existing technologies typically control RPA robots to perform data reading, field mapping, and data writing operations across multiple business systems through preset rules or fixed scheduling strategies to achieve cross-system data synchronization. These methods can reduce the cost of manual intervention to some extent, but most rely on static synchronization strategies or simple rule judgments, making it difficult to dynamically adjust the synchronization strategy based on the operating status of business systems, data change characteristics, and synchronization anomalies.

[0003] In existing technologies, the scheduling and execution of RPA data synchronization tasks typically lack the ability to comprehensively analyze multi-source operational monitoring data. This makes it difficult to effectively identify the impact of factors such as data change frequency, field modification behavior, and system resource consumption on synchronization performance, resulting in insufficient adaptability of synchronization strategies. Furthermore, feedback information such as verification results, execution time, resource consumption, and abnormal events during the synchronization process is often not effectively utilized, making it difficult to establish a strategy optimization mechanism based on historical operational data. Consequently, in complex business environments, problems such as decreased synchronization efficiency, uneven resource utilization, and insufficient anomaly recovery capabilities can easily arise.

[0004] Therefore, how to provide a data synchronization method for RPA robots based on big data analysis is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] One objective of this invention is to propose a data synchronization method for RPA robots based on big data analysis. This invention utilizes an improved data synchronization strategy modeling method to optimize the automated data synchronization process, achieve cross-system data synchronization, and has the advantages of strategy adaptive optimization and high synchronization stability.

[0006] A data synchronization method for RPA robots based on big data analysis according to an embodiment of the present invention includes the following steps: Collect operational monitoring data from multiple business systems, preprocess it, and generate a standardized synchronous dataset; Data synchronization features are extracted from standardized synchronization datasets, and a set of data synchronization feature vectors is constructed. The set of data synchronization feature vectors is input into the improved TimeXer model to model and analyze the synchronization requirements of each synchronization object, and output a set of data synchronization strategy parameters. Construct a set of RPA robot execution tasks based on the data synchronization strategy parameter set; Based on the set of tasks executed by the RPA robot, establish the data field mapping relationship between the source business system and the target business system and perform data mapping processing to generate target synchronization data; Perform verification processing on the target synchronization data, generate synchronization verification results, and output synchronization result identifiers; The system collects synchronization status feedback data during the execution of tasks by the RPA robot, and updates the parameters of the improved TimeXer model based on the synchronization status feedback data to achieve adaptive optimization of the data synchronization strategy.

[0007] Optionally, the operation monitoring data includes source business data, target business data, historical synchronization log data, and system operation status data. The preprocessing includes timestamp alignment, field parsing, missing value handling, outlier removal, data format unification, and structured encoding processing.

[0008] Optionally, the construction of the data synchronization feature vector set specifically includes: The standardized synchronization dataset is divided into time windows and the data is collected according to a preset time interval to generate time window data units of the synchronization objects, and the time window data units of the synchronization objects are arranged in chronological order. Based on the time window data units of the synchronization object, extract the characteristics of data change frequency, data update time interval, and data volume fluctuation; Extract field change characteristics, system resource usage characteristics, and synchronization anomaly event characteristics from the time window data unit of the synchronization object; The data change frequency characteristics, data update time interval characteristics, data volume fluctuation characteristics, field change characteristics, system resource usage characteristics, and synchronization anomaly event characteristics are processed with unified dimensions, and the characteristics are sequentially concatenated according to the synchronization object identifier to generate the data synchronization feature vector corresponding to each synchronization object. The data synchronization feature vectors corresponding to all synchronization objects are summarized to construct a data synchronization feature vector set.

[0009] Optionally, the output of the data synchronization strategy parameter set specifically includes: The data synchronization feature vector set is grouped according to the synchronization object identifier, and further divided into endogenous variable sequences and exogenous variable sequences based on the feature source; The endogenous and exogenous variable sequences are input into the input embedding layer of the improved TimeXer model. The endogenous variable sequence is processed by piecewise encoding to generate endogenous time segment labels, and the exogenous variable sequence is processed by variable-level encoding to generate exogenous variable labels. The improvement of the improved TimeXer model compared with the original model is that the original global endogenous label generation layer is improved into a multi-strategy semantic prototype label generation layer, and the exogenous variable cross-attention encoding layer is improved into a constraint-driven dynamic sparse cross-attention encoding layer. The endogenous variable sequence is input into the multi-policy semantic prototype tag generation layer for semantic aggregation processing to generate multiple policy semantic prototype tags. The endogenous time segment tags are then input into the endogenous time segment self-attention encoding layer for time segment association encoding processing to generate endogenous time segment representation results. Constraint join rules are generated based on field change relationships, primary key relationships, system coupling relationships, and abnormal event relationships. The exogenous variable labels, multiple policy semantic prototype labels, and endogenous time segment representation results are input into the dynamic sparse cross-attention encoding layer. According to the constraint connection rules, dynamic sparse attention connection relationships are established between multiple policy semantic prototype labels and exogenous variable labels. The exogenous variable labels are weighted and interactively encoded through attention weights, and feature fusion processing is performed in combination with the endogenous time segment representation results to generate the synchronous demand association representation results. The output mapping layer maps the representation results of the synchronization requirements, outputting synchronization priority parameters, synchronization time window parameters, synchronization path parameters, synchronization batch parameters, and synchronization trigger condition parameters respectively. The above parameters are then summarized according to the synchronization object identifier to output a set of data synchronization strategy parameters.

[0010] Optionally, the construction of the RPA robot's task set specifically includes: The various strategy parameters in the data synchronization strategy parameter set are grouped according to the synchronization object identifier, and corresponding task parameter configuration units are generated for each synchronization object. The task priority sequence and task scheduling time interval corresponding to each synchronization object are determined based on the task parameter configuration unit. The task execution path sequence and task batch execution sequence corresponding to each synchronization object are determined based on the task parameter configuration unit. Based on the task parameter configuration unit, the task trigger control instructions corresponding to each synchronization object are determined, and the task trigger control instructions are associated and encapsulated with the task priority sequence, task scheduling time interval, task execution path sequence and task batch execution sequence to generate the task execution configuration unit. The task execution configuration units corresponding to each synchronization object are summarized to construct a set of RPA robot execution tasks.

[0011] Optionally, the generation of the target synchronization data specifically includes: Extract the source business system identifier and target business system identifier corresponding to each synchronous task in the RPA robot's task execution set. Read the source business data in the source business system based on the source business system identifier, read the target structure data in the target business system based on the target business system identifier, and establish a data field mapping relationship between the source business system and the target business system. Based on the data field mapping relationship, perform field matching processing on the source business data to obtain a set of field matching results; Based on the field matching result set, perform data type conversion and encoding rule conversion on the source business data to generate a field conversion result set; Based on the field transformation result set and the field matching result set, perform primary key association processing and target structure reorganization processing to generate target synchronized data and write it to the corresponding target business system.

[0012] Optionally, the output of the synchronization result identifier specifically includes: Extract the target synchronization data and the target structure data from the target business system, and collect and process the data of each field in the target synchronization data according to the synchronization object identifier to generate a synchronization verification data unit. Based on the synchronization verification data unit, perform integrity verification and consistency verification on the target synchronization data to obtain the integrity verification result and the consistency verification result. Based on the synchronization verification data unit, perform primary key constraint verification and field value matching verification on the target synchronization data to obtain the primary key constraint verification result and the field value matching verification result; Based on the synchronization verification data unit, the time validity of the target synchronization data is verified, and the synchronization verification result is generated by combining the integrity verification result, consistency verification result, primary key constraint verification result, and field value matching verification result. Based on the synchronization verification result, a result determination process is performed, and a synchronization result identifier is output. The synchronization result identifier includes a synchronization success identifier and a synchronization error identifier.

[0013] Optionally, the adaptive optimization of the data synchronization strategy specifically includes: Synchronization status feedback data is collected during the execution of tasks by the RPA robot, and then aggregated and processed according to the synchronization object identifier to generate a synchronization status feedback data set. The synchronization status feedback data includes task status information, synchronization verification results, synchronization result identifiers, execution duration information, resource consumption information, and abnormal event information. The synchronization state feedback data set is standardized and structured, and then used to generate model feedback data. The parameters of the improved TimeXer model are updated based on the model feedback data, and the data synchronization strategy parameter set is regenerated based on the updated model. The set of tasks to be executed by the RPA robot is regenerated based on the new set of data synchronization strategy parameters, and the RPA robot is scheduled to perform data synchronization processing.

[0014] The beneficial effects of this invention are: This invention proposes a data synchronization method for RPA robots based on big data analytics. By collecting and preprocessing operational monitoring data from multiple business systems, a standardized synchronization dataset is constructed. Based on this dataset, multi-dimensional data synchronization features are extracted, including data change frequency, data update interval, data volume fluctuation, field change behavior, and system resource consumption, forming a set of data synchronization feature vectors. An improved TimeXer model is used to jointly model and analyze these multi-dimensional features. This allows for refined analysis of the synchronization requirements of each synchronization object before synchronization execution, outputting strategy parameters such as synchronization priority, synchronization time window, synchronization path, synchronization batch, and synchronization trigger conditions. This enables the RPA robot to dynamically generate a set of execution tasks based on business data change characteristics and system operating status, improving the rationality and efficiency of data synchronization task scheduling.

[0015] This invention establishes a data field mapping relationship between the source and target business systems during the data synchronization task performed by an RPA robot. It generates target synchronized data through processes such as field matching, data type conversion, encoding rule conversion, primary key association, and target structure reorganization. Simultaneously, it performs integrity, consistency, primary key constraint, field value matching, and time validity checks on the synchronization results, effectively ensuring data structure consistency and data content accuracy during cross-system data synchronization. By comprehensively verifying the synchronization results and outputting synchronization result identifiers, abnormal synchronization situations can be identified promptly during the data synchronization execution phase, improving the reliability of the data synchronization process and data quality control capabilities.

[0016] Furthermore, this invention collects synchronization status feedback data during RPA robot task execution, including task status information, synchronization verification results, synchronization result identifiers, execution duration information, resource consumption information, and abnormal event information. Based on this feedback data, the improved TimeXer model parameters are updated, enabling the model to continuously learn the operational patterns during data synchronization execution, thereby achieving adaptive optimization of the data synchronization strategy. By continuously updating the data synchronization strategy parameters and regenerating the RPA robot's task set, the data synchronization tasks can dynamically adjust in complex business environments, improving the overall data synchronization efficiency, resource utilization, and anomaly handling capabilities of the system. 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 RPA robot data synchronization method based on big data analysis proposed in this invention; Figure 2 This invention presents a flowchart illustrating the output of the data synchronization strategy parameter set for an RPA robot data synchronization method based on big data analysis. Figure 3 This is a flowchart illustrating the target synchronization data generation process of an RPA robot data synchronization method based on big data analysis 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 data synchronization method for RPA robots based on big data analytics includes the following steps: Collect operational monitoring data from multiple business systems, preprocess it, and generate a standardized synchronous dataset; Data synchronization features are extracted from standardized synchronization datasets, and a set of data synchronization feature vectors is constructed. The set of data synchronization feature vectors is input into the improved TimeXer model to model and analyze the synchronization requirements of each synchronization object, and output a set of data synchronization strategy parameters. Construct a set of RPA robot execution tasks based on the data synchronization strategy parameter set; Based on the set of tasks executed by the RPA robot, establish the data field mapping relationship between the source business system and the target business system and perform data mapping processing to generate target synchronization data; Perform verification processing on the target synchronization data, generate synchronization verification results, and output synchronization result identifiers; The system collects synchronization status feedback data during the execution of tasks by the RPA robot, and updates the parameters of the improved TimeXer model based on the synchronization status feedback data to achieve adaptive optimization of the data synchronization strategy.

[0020] In this embodiment, the operation monitoring data includes source business data, target business data, historical synchronization log data, and system operation status data. Preprocessing includes timestamp alignment, field parsing, missing value handling, outlier removal, data format unification, and structured encoding processing.

[0021] In this embodiment, the construction of the data synchronization feature vector set specifically includes: The standardized synchronization dataset is divided into time windows and the data is collected according to a preset time interval to generate time window data units of the synchronization objects, and the time window data units of the synchronization objects are arranged in chronological order. Based on the time window data units of the synchronization object, data change frequency characteristics, data update time interval characteristics, and data volume fluctuation characteristics are extracted. Among them, the data change frequency characteristic is obtained by summing the number of new data records, the number of modified data records, and the number of deleted data records in the current time window and processing them with the corresponding statistical duration of the current time window. The data update time interval characteristic is obtained by calculating the time difference between two adjacent data update times in the current time window. The data volume fluctuation characteristic is obtained by calculating the change in the total number of data records in a continuous time window. Based on the time window data unit of the synchronization object, field change features, system resource usage features, and synchronization anomaly event features are extracted. Among them, field change features are obtained by identifying change events corresponding to changes in field name, field type, field length, field value range, and field mapping relationship. System resource usage features are obtained by extracting processor usage data, memory usage data, storage read / write usage data, and network transmission usage data from the system running status data. Synchronization anomaly event features are obtained by identifying synchronization failure events, field missing events, primary key conflict events, latency exceedance events, write interruption events, and duplicate synchronization events from historical synchronization log data. The data change frequency characteristics, data update time interval characteristics, data volume fluctuation characteristics, field change characteristics, system resource usage characteristics, and synchronization abnormal event characteristics are processed with unified dimensions. The characteristics are then sequentially concatenated according to the synchronization object identifier to generate the data synchronization feature vector corresponding to each synchronization object. The unified dimension processing includes obtaining the unified dimension feature value by subtracting the corresponding feature statistical average value from each feature value and dividing by the corresponding feature discrete scale value. The data synchronization feature vectors corresponding to all synchronization objects are summarized to construct a data synchronization feature vector set.

[0022] In this embodiment, the output of the data synchronization strategy parameter set specifically includes: The data synchronization feature vector set is grouped according to the synchronization object identifier and divided into endogenous variable sequence and exogenous variable sequence based on the feature source. The feature source includes the data source of the synchronization object itself and the data source associated with the system. The data change frequency feature, data update time interval feature and data volume fluctuation feature corresponding to the data source of the synchronization object itself constitute the endogenous variable sequence, while the field change feature, system resource consumption feature and synchronization abnormal event feature corresponding to the data source associated with the system constitute the exogenous variable sequence. The improved TimeXer model inputs the endogenous and exogenous variable sequences into the input embedding layer. The endogenous variable sequence is processed by piecewise encoding to generate endogenous time segment labels, and the exogenous variable sequence is processed by variable-level encoding to generate exogenous variable labels. The improvement of the improved TimeXer model compared to the original model is that the original global endogenous label generation layer is improved into a multi-strategy semantic prototype label generation layer, and the exogenous variable cross-attention encoding layer is improved into a constraint-driven dynamic sparse cross-attention encoding layer. The generation of endogenous time segment labels includes: dividing the endogenous variable sequence into continuous time segments according to a preset time segment length to obtain multiple endogenous time segment data blocks; performing numerical alignment and feature concatenation processing on the data change frequency characteristics, data update time interval characteristics, and data volume fluctuation characteristics in each endogenous time segment data block to generate endogenous time segment feature vectors; performing linear mapping processing on each endogenous time segment feature vector to unify the vector dimension, and fusing and encoding the mapped endogenous time segment feature vectors with the corresponding time segment position identifier information to generate corresponding endogenous time segment labels; The generation of exogenous variable labels includes: performing variable-level aggregation processing on field change features, system resource usage features, and synchronization anomaly event features in the exogenous variable sequence according to variable type to obtain variable feature vectors corresponding to each variable; performing dimension-unified mapping processing on each variable feature vector; and performing fusion encoding processing on the mapped variable feature vectors and corresponding variable category identification information to generate corresponding exogenous variable labels. The endogenous variable sequence is input into the multi-policy semantic prototype tag generation layer for semantic aggregation processing to generate multiple policy semantic prototype tags. The endogenous time segment tags are then input into the endogenous time segment self-attention encoding layer for time segment association encoding processing to generate endogenous time segment representation results. Semantic aggregation processing includes: first, grouping the data change frequency features, data update time interval features, and data volume fluctuation features in the endogenous variable sequence into time segments according to time order, resulting in multiple endogenous time segment feature sets; second, performing feature normalization on various feature values ​​in each endogenous time segment feature set to eliminate differences in the dimensions of different features; third, statistically summarizing the normalized feature values ​​according to feature type, obtaining the average feature value by summing the various feature values ​​within the same time segment and dividing by the number of feature values ​​within that time segment, and obtaining the feature fluctuation value by summing the absolute values ​​of the differences between each feature value and the average feature value and dividing by the number of feature values, thus obtaining a comprehensive feature representation of the time segment; and finally, performing cross-time segment aggregation on the comprehensive feature representations of each time segment according to time order, by weighted summing of the comprehensive feature representations of adjacent time segments and dividing by the time segment involved in the calculation. The quantity yields a time series aggregated feature representation; based on preset strategy semantic partitioning rules, semantic mapping processing is performed on the time series aggregated feature representation, by extracting synchronization frequency change features, synchronization time interval features, data scale change features, and synchronization stability features from the time series aggregated feature representation, and combining the extracted features to form different strategy semantic feature sets; prototype labeling construction processing is performed on each strategy semantic feature set, by concatenating the feature values ​​in each strategy semantic feature set to form a strategy semantic feature vector, performing dimension unification mapping processing on the strategy semantic feature vector to generate a semantic representation vector, and fusing the semantic representation vector with the corresponding strategy semantic category identification information to form multiple strategy semantic prototype labels corresponding to synchronization priority semantic information, synchronization time window semantic information, synchronization path semantic information, synchronization batch semantic information, and synchronization triggering condition semantic information; The temporal segment association encoding process includes: arranging endogenous temporal segment markers in chronological order; calculating the association strength between temporal segments based on the feature similarity between each time segment marker; obtaining the feature difference degree by summing the absolute values ​​of the differences between corresponding feature values ​​of two time segment markers and dividing by the number of features; and then converting the feature difference degree into association strength to obtain the temporal segment association weight; constructing a temporal segment association matrix based on the temporal segment association weight, where each element in the matrix represents the degree of association between two time segment markers; performing association weighting processing on each endogenous temporal segment marker according to the temporal segment association matrix; obtaining the temporal segment interaction feature by multiplying each time segment marker with its corresponding association weight and summing all results; performing multi-time segment fusion processing on the temporal segment interaction feature in chronological order; obtaining the temporal segment fusion feature by accumulating the current temporal segment interaction feature with the interaction features of the preceding and following temporal segments and dividing by the number of time segments participating in the fusion; and performing sequence integration processing on the temporal segment fusion feature to generate an endogenous temporal segment representation result containing the association relationships between multiple time segments. Constraint join rules are generated based on field change relationships, primary key relationships, system coupling relationships, and abnormal event relationships. The acquisition of field change relationships includes: extracting records of field name changes, field type changes, and field value range changes from historical synchronization log data, matching and comparing field change records of the same synchronization object in different time segments, generating a set of field change records by statistically analyzing the field names and field attributes that changed in adjacent time segments, and determining the field change relationship based on the frequency and continuous changes of field changes in the set of field change records. Obtaining primary key relationships includes: extracting primary key field information from each data table from source business data and target business data, performing primary key value matching processing on the primary key fields of different business systems, generating a primary key mapping set by associating data records with the same primary key value, and determining the primary key relationship based on the primary key correspondence between different data tables in the primary key mapping set; The acquisition of system coupling relationships includes: extracting data call records, interface call records, and data synchronization path records between various business systems from system operation status data and historical synchronization log data; generating a system interaction record set by statistically analyzing the number of data calls and synchronization dependency order between various business systems; and determining the system coupling relationship based on the call frequency and dependency order between different systems in the system interaction record set. The acquisition of abnormal event correlations includes: extracting synchronization failure events, field missing events, primary key conflict events, and synchronization interruption events from historical synchronization log data, and collecting abnormal event records in chronological order. By statistically analyzing the co-occurrence of different abnormal events in adjacent time segments, a set of abnormal event co-occurrence records is generated, and the abnormal event correlation is determined based on the co-occurrence frequency of abnormal events in the abnormal event co-occurrence record set. The exogenous variable labels, multiple policy semantic prototype labels, and endogenous time segment representation results are input into the dynamic sparse cross-attention encoding layer. According to the constraint connection rules, dynamic sparse attention connection relationships are established between multiple policy semantic prototype labels and exogenous variable labels. The exogenous variable labels are weighted and interactively encoded through attention weights, and feature fusion processing is performed in combination with the endogenous time segment representation results to generate the synchronous demand association representation results. The generation of the synchronous demand association representation results includes: determining multiple policy semantic prototype tags and exogenous variable tags that are allowed to establish connections based on constraint connection rules, and selecting a corresponding set of exogenous variable tags for each policy semantic prototype tag according to the variable combination, thereby establishing a dynamic sparse attention connection relationship between the policy semantic prototype tag and the exogenous variable tags; calculating the feature correlation degree between each policy semantic prototype tag and its corresponding set of exogenous variable tags, obtaining the correlation value by summing the absolute values ​​of the differences between the feature values ​​of the policy semantic prototype tag and the feature values ​​of the exogenous variable tags and normalizing them according to the number of features, and then... Attention weights are generated from degree values; weighted interactive coding is performed on exogenous variable labels using these attention weights, and interactive coding feature vectors are generated by multiplying the feature values ​​of each exogenous variable label with their corresponding attention weights and accumulating the product results; interactive coding feature vectors are then fused with endogenous time segment representation results, and joint feature vectors are formed by concatenating the interactive coding feature vectors with the endogenous time segment representation results, and dimensionality unification mapping is performed on the joint feature vectors to obtain fused feature representations; finally, time series aggregation processing is performed on the fused feature representations according to the synchronization object identifiers to generate synchronization requirement association representation results; The output mapping layer maps the representation results of the synchronization requirements, outputting synchronization priority parameters, synchronization time window parameters, synchronization path parameters, synchronization batch parameters, and synchronization trigger condition parameters respectively. The above parameters are then summarized according to the synchronization object identifier, and a set of data synchronization strategy parameters is output. The mapping process includes: performing sequence expansion on each feature value in the synchronization demand association representation result according to the feature order, converting the multidimensional feature representation in the synchronization demand association representation result into a continuously arranged representation feature sequence; performing interval division processing on the representation feature sequence according to the preset parameter mapping rules, dividing the representation feature sequence into multiple feature intervals according to the position order in the feature sequence, so that each feature interval corresponds to a synchronization strategy parameter type; performing weighted combination processing on the feature values ​​in each feature interval, multiplying each feature value in the same feature interval by the corresponding weight and accumulating the product results to obtain the parameter representation value; and then performing parameter transformation processing on the parameter representation value according to the parameter type, matching the parameter representation value with the preset parameter threshold range and generating the corresponding synchronization strategy control parameters according to the matching results, obtaining synchronization priority parameters, synchronization time window parameters, synchronization path parameters, synchronization batch parameters, and synchronization trigger condition parameters.

[0023] In this embodiment, the construction of the task set to be executed by the RPA robot specifically includes: The various strategy parameters in the data synchronization strategy parameter set are grouped according to the synchronization object identifier, and corresponding task parameter configuration units are generated for each synchronization object. The task priority sequence and task scheduling time interval corresponding to each synchronization object are determined based on the task parameter configuration unit. The determination of the task priority sequence includes: extracting synchronization priority parameters from the task parameter configuration unit, and aggregating each synchronization priority parameter according to the synchronization object identifier to form a priority parameter set; sorting the synchronization tasks corresponding to each synchronization object according to the priority level value in the priority parameter set, and generating a task priority sequence by arranging the priority level values ​​in descending order and assigning corresponding task execution order to each synchronization object according to the arrangement order. The determination of the task scheduling time interval includes: extracting the task start execution time and task end execution time from the synchronization time window parameters from the task parameter configuration unit, parsing and processing the task start execution time and task end execution time execution time interval, calculating the time window length by calculating the time difference between the task start execution time and the task end execution time, and establishing the correspondence between the task start execution time, the task end execution time and the time window length according to the synchronization object identifier to generate the task scheduling time interval; The task execution path sequence and task batch execution sequence corresponding to each synchronization object are determined based on the task parameter configuration unit. The determination of the task execution path sequence includes: extracting synchronization path parameters from the task parameter configuration unit, and aggregating each synchronization path parameter according to the synchronization object identifier to form a path parameter set; performing sequential parsing on the source business system identifier, target business system identifier, and intermediate transmission node identifier in the path parameter set, arranging the source business system identifier, intermediate transmission node identifier, and target business system identifier in order according to the data transmission direction, and establishing a node order mapping relationship based on the connection relationship between nodes to generate the task execution path sequence; The determination of the task batch execution sequence includes: extracting the single batch data volume, batch execution order, and batch interval time from the synchronous batch parameters in the task parameter configuration unit, and aggregating each synchronous batch parameter according to the synchronous object identifier to form a batch parameter set; performing batch division processing on the total synchronous data record based on the single batch data volume in the batch parameter set, generating multiple data batch units by continuously segmenting the total synchronous data record according to the single batch data volume, sorting each data batch unit according to the batch execution order, and establishing the time interval relationship between adjacent data batch units in combination with the batch interval time to generate the task batch execution sequence; Based on the task parameter configuration unit, the task trigger control instructions corresponding to each synchronization object are determined, and the task trigger control instructions are associated and encapsulated with the task priority sequence, task scheduling time interval, task execution path sequence and task batch execution sequence to generate the task execution configuration unit. The generation of task execution configuration units includes: extracting synchronization trigger condition parameters from the task parameter configuration unit, and performing rule parsing on the trigger time condition, trigger data change condition, trigger resource status condition, and trigger exception recovery condition in the synchronization trigger condition parameters to generate task trigger control instructions corresponding to each synchronization object; simultaneously reading the task priority sequence, task scheduling time interval, task execution path sequence, and task batch execution sequence, and aggregating the various sequence information according to the synchronization object identifier to form a task configuration parameter set; performing parameter association processing on the task trigger control instructions and the task configuration parameter set, by correspondingly combining the task trigger control instructions with the task execution order in the task priority sequence, the task execution time range in the task scheduling time interval, the path node order in the task execution path sequence, and the batch execution order in the task batch execution sequence to generate task execution parameter groups; and then performing structured encapsulation processing on each task execution parameter group, by integrating the task trigger control instructions, task priority sequence, task scheduling time interval, task execution path sequence, and task batch execution sequence according to a unified task structure format to generate task execution configuration units corresponding to each synchronization object. The task execution configuration units corresponding to each synchronization object are summarized to construct a set of RPA robot execution tasks.

[0024] In this embodiment, the generation of target synchronization data specifically includes: Extract the source business system identifier and target business system identifier corresponding to each synchronous task in the RPA robot's task execution set. Read the source business data in the source business system based on the source business system identifier, read the target structure data in the target business system based on the target business system identifier, and establish a data field mapping relationship between the source business system and the target business system. Based on the data field mapping relationship, perform field matching processing on the source business data to obtain a set of field matching results; The field matching process includes: extracting a source field set from the source business data based on the data field mapping relationship, and extracting a target field set from the target structure data of the target business system; performing item-by-item matching between each source field in the source field set and the corresponding target field in the target field set according to the data field mapping relationship, by reading the field name, field type, field length, and field value range of the source field and comparing them with the field name, field type, field length, and field value range of the corresponding target field; determining a name matching result when the field name is consistent or the field name is consistent with the target field name after being transformed by the field mapping rules; determining a type matching result when the field type is consistent with the target field type or can be converted; determining a length matching result when the field length meets the target field length constraint; and determining a value range matching result when the field value range falls within the allowed value range of the target field; establishing the field correspondence between the source fields and the target fields based on the name matching result, type matching result, length matching result, and value range matching result, and determining the position of the target field corresponding to each source field according to the order of the target fields in the target field set, thus obtaining a field matching result set; Based on the field matching result set, perform data type conversion and encoding rule conversion on the source business data to generate a field conversion result set; The generation of the field conversion result set includes: extracting the field correspondence between the source fields and the target fields, as well as the position of the target fields, from the field matching result set; reading the data record values ​​corresponding to each source field from the source business data based on the field correspondence; performing data type conversion processing on the source field data record values ​​according to the target field data type requirements recorded in the field matching result set, by reorganizing the format and converting the values ​​of character data, numeric data, time data, and enumeration data to ensure that the converted data record values ​​meet the data type constraints of the target fields; simultaneously, performing encoding rule conversion processing on the converted data record values ​​according to the encoding rules of the target business system, by mapping character encoding methods, field value encoding methods, and status identifier encoding methods to ensure that the converted data record values ​​conform to the encoding specifications of the target business system; finally, arranging the converted data record values ​​in order according to the position of the target fields in the field matching result set, and aggregating the converted data record values ​​of each field according to the synchronization object identifier to generate the field conversion result set. Based on the field transformation result set and the field matching result set, perform primary key association processing and target structure reorganization processing to generate target synchronized data and write it to the corresponding target business system; The generation of target synchronization data includes: extracting the transformed data record values ​​corresponding to each source field from the field transformation result set, and performing field positioning processing on each transformed data record value according to the target field position of the record in the field matching result set; extracting the primary key field value according to the primary key field identifier in the field transformation result set, and matching the primary key field value with the target primary key field value of the corresponding record in the target business system, establishing the primary key association relationship between the source business data record and the target business system record by comparing the consistency of the primary key field values; after the primary key association relationship is determined, performing field rearrangement processing, field merging processing, and field nesting assembly processing on each transformed data record value according to the target field arrangement order, target table structure hierarchy, and target record organization rules of the target business system, so that each field data is organized according to the structure format of the target business system; and then performing record-level combination processing on the reorganized field data according to the synchronization object identifier to generate target synchronization data consistent with the data structure of the target business system.

[0025] In this embodiment, the output of the synchronization result identifier specifically includes: Extract the target synchronization data and the target structure data from the target business system, and collect and process the data of each field in the target synchronization data according to the synchronization object identifier to generate a synchronization verification data unit. Based on the synchronization verification data unit, perform integrity verification and consistency verification on the target synchronization data to obtain the integrity verification result and the consistency verification result. Integrity and consistency verification include: based on the target field set, target field constraint information, and target record organization rules in the target structure data, comparing the number of fields, field positions, field hierarchical relationships, and record organization structure in the target synchronized data item by item to determine whether there are missing fields, redundant fields, or structural omissions in the target synchronized data, and obtaining the integrity verification result; simultaneously, extracting the correspondence between source fields and target fields from the field matching result set, reading the corresponding source field values ​​from the source business data based on the correspondence between source fields and target fields, and reading the corresponding target field values ​​from the target synchronized data, performing consistency comparison processing on the field content, field order, and record correspondence before and after mapping, and determining whether there are inconsistencies in field content and record correspondence between the target synchronized data and the source business data, and obtaining the consistency verification result; Based on the synchronization verification data unit, perform primary key constraint verification and field value matching verification on the target synchronization data to obtain the primary key constraint verification result and the field value matching verification result; Primary key constraint verification and field value matching verification include: extracting primary key field values ​​from the target synchronized data and performing duplicate and correlation comparisons with the target primary key field values ​​of existing records in the target business system to determine whether there are duplicate primary keys, missing primary keys, or abnormal primary key associations in the target synchronized data, thus obtaining the primary key constraint verification result; simultaneously, according to the field matching result set, extracting the target field values ​​from the target synchronized data and the corresponding source field values ​​from the source business data, performing field value correspondence matching processing on the target field values ​​and source field values ​​to determine whether there are numerical deviations, character deviations, status identifier deviations, and enumeration value deviations between each target field value in the target synchronized data and the corresponding source field value in the source business data, thus obtaining the field value matching verification result; Based on the synchronization verification data unit, the time validity of the target synchronization data is verified, and the synchronization verification result is generated by combining the integrity verification result, consistency verification result, primary key constraint verification result, and field value matching verification result. The generation of time validity verification and synchronization verification results includes: extracting the time field value from the target synchronization data, and performing time correspondence comparison processing with the task start execution time, task end execution time in the task scheduling time interval, and time constraint rules in the target business system to determine whether the time field value in the target synchronization data falls within the allowed time range and to determine whether there are any expired time records, thus obtaining the time validity verification result; then, the integrity verification result, consistency verification result, primary key constraint verification result, field value matching verification result, and time validity verification result are aggregated to generate the synchronization verification result. Based on the synchronization verification results, the result judgment process is performed, and the synchronization result identifier is output. The synchronization result identifier includes a synchronization success identifier and a synchronization error identifier. The result determination process includes: when the integrity check result, consistency check result, primary key constraint check result, field value matching check result, and time validity check result all meet the preset passing conditions, a synchronization success flag is output; when any check result does not meet the preset passing conditions, a synchronization error flag is output.

[0026] In this embodiment, the adaptive optimization of the data synchronization strategy specifically includes: Collect synchronization status feedback data during the execution of tasks by the RPA robot, and process the data according to the synchronization object identifier to generate a synchronization status feedback data set. The synchronization status feedback data includes task status information, synchronization verification results, synchronization result identifier, execution duration information, resource consumption information, and abnormal event information. The synchronization state feedback data set is standardized and structured, and then used to generate model feedback data. The generation of model feedback data includes: performing state encoding processing on the task start status, task execution status, task interruption status, and task end status in the task status information to obtain task status feature data; performing result classification processing on the integrity verification results, consistency verification results, primary key constraint verification results, field value matching verification results, and time validity verification results in the synchronization verification results to obtain verification result feature data; performing identifier mapping processing on the synchronization success identifier and synchronization exception identifier in the synchronization result identifier to obtain result identifier feature data; and processing the task start execution time, task end execution time, and task duration in the execution duration information. The execution time feature data is obtained by extracting and processing the execution time feature data. The processor usage, memory usage, network transmission usage, and storage read / write usage in the resource consumption information are normalized to obtain resource consumption feature data. The abnormal event type, occurrence time, duration, and recovery status in the abnormal event information are encoded to obtain abnormal event feature data. Finally, the task status feature data, verification result feature data, result identifier feature data, execution time feature data, resource consumption feature data, and abnormal event feature data are concatenated and integrated according to the synchronization object identifier to generate model feedback data. The parameters of the improved TimeXer model are updated based on the model feedback data, and the data synchronization strategy parameter set is regenerated based on the updated model. The parameter update process includes: inputting model feedback data into the input embedding layer of the improved TimeXer model; performing numerical alignment, feature concatenation, and dimensionality unification mapping on task status information, synchronization verification results, synchronization result identifiers, execution duration information, resource consumption information, and abnormal event information in the model feedback data to generate a model update input feature sequence; inputting the model update input feature sequence into the multi-strategy semantic prototype label generation layer for feedback semantic aggregation; performing time-order aggregation on the model update input feature sequence according to the synchronization object identifier; performing statistical summary and semantic mapping on the aggregated feature data to generate a set of feedback strategy semantic features; and constructing corresponding update semantic prototype labels based on synchronization priority semantic information, synchronization time window semantic information, synchronization path semantic information, synchronization batch semantic information, and synchronization trigger condition semantic information; and inputting the model update input feature sequence into the endogenous time segment self-attention layer. The force encoding layer performs feedback time segment association encoding processing. By calculating the association weights based on the degree of feature differences between time segments, and performing weighted interaction processing on the features of each time segment based on the association weights, an updated time series representation result is generated. Then, the updated semantic prototype label and the updated time series representation result are input into the constraint-driven dynamic sparse cross-attention encoding layer for feedback interaction encoding processing. By establishing a dynamic sparse attention connection relationship between the updated semantic prototype label and the model update input feature sequence according to the preset constraint connection rules, and performing weighted interaction encoding processing on the model update input feature sequence based on the attention weights, an update strategy representation result is generated. Finally, the parameters in the improved TimeXer model are iteratively updated based on the update strategy representation result. By calculating the degree of difference between the updated strategy representation result and the historical strategy representation result, and performing weight adjustment processing on the internal parameters of the model based on the degree of difference, the updated model parameters are obtained. The set of tasks to be executed by the RPA robot is regenerated based on the new set of data synchronization strategy parameters, and the RPA robot is scheduled to perform data synchronization processing.

[0027] Example 1: To verify the feasibility of this invention in practice, it was applied to the data system environment of a large manufacturing enterprise. During its long-term IT development, this enterprise has established multiple business systems, including a production management system, a warehouse management system, a supply chain management system, and a financial management system. These systems are deployed in different server environments, with differences in data structure, field definitions, and coding rules. In its daily operations, the enterprise needs to synchronize production order data, inventory change data, and supply information across these multiple business systems to ensure data consistency in production scheduling, inventory control, and financial accounting. Existing data synchronization methods mainly rely on preset rules or fixed scheduling procedures. When the field structure of a business system changes or the synchronization load fluctuates, synchronization failures, duplicate data writes, or synchronization delays can easily occur, leading to short-term inconsistencies in data status between the production and warehouse systems, thus affecting the timeliness of the enterprise's business decisions.

[0028] After deploying the data synchronization method of this invention in this environment, the system first collects operational monitoring data from each business system, including changes in source business data, target business data structure information, and historical synchronization log data, and forms a standardized synchronization dataset through unified format processing. Based on this, data synchronization characteristics such as data change frequency characteristics, data update time interval characteristics, and system resource consumption characteristics are extracted, and a set of data synchronization strategy parameters is generated through an improved modeling and analysis process. The system automatically constructs a set of RPA robot execution tasks according to the strategy parameters, enabling the RPA robot to execute synchronization tasks according to priority order, time scheduling interval, and data transmission path. During execution, by establishing a data field mapping relationship between the source and target business systems, field matching, data type conversion, and encoding rule conversion are performed on the source business data, and target synchronization data is generated after primary key association and structure reorganization. Subsequently, the generated target synchronization data undergoes integrity verification, consistency verification, and primary key constraint verification, and a synchronization result identifier is output based on the verification results. Simultaneously, the system continuously collects task status information, resource consumption information, and abnormal event information of the RPA robot during execution, and inputs this synchronization status feedback data into the model for parameter updates, enabling the data synchronization strategy to be dynamically adjusted according to the business system's operational status.

[0029] In the actual operating environment of this enterprise, this invention was deployed on the enterprise's data center server and ran continuously during peak production periods. During system operation, the data synchronization task covered multiple types of business data between the production system, warehousing system, and supply chain system. During the synchronization process, a real-time feedback mechanism continuously updated the synchronization strategy, enabling the synchronization schedule to adaptively adjust according to changes in system load. Operation monitoring records showed that during long-term continuous production cycles, the data synchronization process between various business systems remained stable, with no structural inconsistencies or abnormal field mappings occurring during data writing, and the synchronization verification results remained consistent. The enterprise's information management department found in the data center's operation and maintenance records that the system could maintain stable execution of synchronization tasks even when multiple business systems were concurrently updating data, effectively reducing data synchronization anomalies caused by changes in field structure or system load fluctuations, thus verifying the feasibility and stability of this invention in complex business system environments.

[0030] Table 1. Performance Comparison of the Invention and Traditional RPA Robot Data Synchronization Methods

[0031] As can be clearly seen from Table 1, the method of the present invention is superior to the traditional method in many indicators.

[0032] Regarding the average data synchronization latency, the traditional method is 842ms, while the method of this invention reduces it to 691ms, a reduction of 151ms. This improvement mainly stems from the fact that this invention models and analyzes the data synchronization characteristics, generates a set of data synchronization strategy parameters, and dynamically schedules RPA robots to execute tasks based on these parameters. This allows the data synchronization path and task scheduling order to be adjusted according to the business system status, thereby reducing ineffective scheduling waiting time.

[0033] Regarding the average processing time for a single task, traditional methods take 17.6 seconds, while the method of this invention takes 14.8 seconds. The processing time is reduced by 2.8 seconds because this invention establishes a data field mapping relationship between the source business system and the target business system before the synchronous task is executed, and directly performs field matching and data type conversion according to the mapping rules during execution, thereby reducing the time overhead caused by repeatedly parsing the field structure during the process.

[0034] In terms of synchronization success rate and anomaly rate, the traditional method achieves a 95.2% success rate and a 4.6% anomaly rate, while the method of this invention improves the success rate to 97.9% and reduces the anomaly rate to 3.4%. This improvement stems from the fact that this invention adds multi-dimensional verification mechanisms, including integrity verification, consistency verification, primary key constraint verification, and field value matching verification, after data synchronization is completed. By performing multi-layered verification on the target synchronized data and outputting synchronization result identifiers, potential data anomalies can be detected in a timely manner during the synchronization process, preventing abnormal data from being written into the business system.

[0035] Regarding system resource utilization, the traditional method achieved an average CPU utilization rate of 63.4%, while the method of this invention achieved 56.7%; the average memory utilization rate decreased from 58.1% to 52.9%. The reason for the reduction in resource utilization is that this invention utilizes synchronization status feedback data to continuously update model parameters, enabling the data synchronization strategy to be dynamically optimized according to the system's operating status, thereby reducing unnecessary repetitive synchronization tasks and resource contention.

[0036] Regarding concurrent task processing capability, the traditional method achieves 46 tasks / minute, while the method of this invention reaches 54 tasks / minute, an improvement of 8 tasks / minute. This performance improvement is mainly due to the fact that this invention rationally schedules the RPA robot's tasks through task priority sequences, task scheduling time intervals, and task execution path sequences, enabling multiple synchronous tasks to be executed concurrently according to priority and time windows.

[0037] Regarding the data consistency verification pass rate, the traditional method achieves 96.1%, while the method of this invention reaches 98.4%. This improvement is mainly due to the fact that this invention performs structural reorganization after data mapping, field transformation, and primary key association processing, ensuring that the target synchronized data strictly conforms to the data structure requirements of the target business system, thereby effectively improving the data consistency level. In summary, the above indicators demonstrate that this invention exhibits superior performance in data synchronization efficiency, system stability, and resource utilization.

[0038] The above are merely preferred embodiments 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 data synchronization method for RPA robots based on big data analysis, characterized in that, Includes the following steps: Collect operational monitoring data from multiple business systems, preprocess it, and generate a standardized synchronous dataset; Data synchronization features are extracted from standardized synchronization datasets, and a set of data synchronization feature vectors is constructed. The set of data synchronization feature vectors is input into the improved TimeXer model to model and analyze the synchronization requirements of each synchronization object, and output a set of data synchronization strategy parameters. Construct a set of RPA robot execution tasks based on the data synchronization strategy parameter set; Based on the set of tasks executed by the RPA robot, establish the data field mapping relationship between the source business system and the target business system and perform data mapping processing to generate target synchronization data; Perform verification processing on the target synchronization data, generate synchronization verification results, and output synchronization result identifiers; The system collects synchronization status feedback data during the execution of tasks by the RPA robot, and updates the parameters of the improved TimeXer model based on the synchronization status feedback data to achieve adaptive optimization of the data synchronization strategy.

2. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The operational monitoring data includes source business data, target business data, historical synchronization log data, and system operation status data. The preprocessing includes timestamp alignment, field parsing, missing value handling, outlier removal, data format unification, and structured encoding processing.

3. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The construction of the data synchronization feature vector set specifically includes: The standardized synchronization dataset is divided into time windows and the data is collected according to a preset time interval to generate time window data units of the synchronization objects, and the time window data units of the synchronization objects are arranged in chronological order. Based on the time window data units of the synchronization object, extract the characteristics of data change frequency, data update time interval, and data volume fluctuation; Extract field change characteristics, system resource usage characteristics, and synchronization anomaly event characteristics from the time window data unit of the synchronization object; The data change frequency characteristics, data update time interval characteristics, data volume fluctuation characteristics, field change characteristics, system resource usage characteristics, and synchronization anomaly event characteristics are processed with unified dimensions, and the characteristics are sequentially concatenated according to the synchronization object identifier to generate the data synchronization feature vector corresponding to each synchronization object. The data synchronization feature vectors corresponding to all synchronization objects are summarized to construct a data synchronization feature vector set.

4. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The output of the data synchronization strategy parameter set specifically includes: The data synchronization feature vector set is grouped according to the synchronization object identifier, and further divided into endogenous variable sequences and exogenous variable sequences based on the feature source; The endogenous and exogenous variable sequences are input into the input embedding layer of the improved TimeXer model. The endogenous variable sequence is processed by piecewise encoding to generate endogenous time segment labels, and the exogenous variable sequence is processed by variable-level encoding to generate exogenous variable labels. The improvement of the improved TimeXer model compared with the original model is that the original global endogenous label generation layer is improved into a multi-strategy semantic prototype label generation layer, and the exogenous variable cross-attention encoding layer is improved into a constraint-driven dynamic sparse cross-attention encoding layer. The endogenous variable sequence is input into the multi-policy semantic prototype tag generation layer for semantic aggregation processing to generate multiple policy semantic prototype tags. The endogenous time segment tags are then input into the endogenous time segment self-attention encoding layer for time segment association encoding processing to generate endogenous time segment representation results. Constraint join rules are generated based on field change relationships, primary key relationships, system coupling relationships, and abnormal event relationships. The exogenous variable labels, multiple policy semantic prototype labels, and endogenous time segment representation results are input into the dynamic sparse cross-attention encoding layer. According to the constraint connection rules, dynamic sparse attention connection relationships are established between multiple policy semantic prototype labels and exogenous variable labels. The exogenous variable labels are weighted and interactively encoded through attention weights, and feature fusion processing is performed in combination with the endogenous time segment representation results to generate the synchronous demand association representation results. The output mapping layer maps the representation results of the synchronization requirements, outputting synchronization priority parameters, synchronization time window parameters, synchronization path parameters, synchronization batch parameters, and synchronization trigger condition parameters respectively. The above parameters are then summarized according to the synchronization object identifier to output a set of data synchronization strategy parameters.

5. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The construction of the RPA robot's task set specifically includes: The various strategy parameters in the data synchronization strategy parameter set are grouped according to the synchronization object identifier, and corresponding task parameter configuration units are generated for each synchronization object. The task priority sequence and task scheduling time interval corresponding to each synchronization object are determined based on the task parameter configuration unit. The task execution path sequence and task batch execution sequence corresponding to each synchronization object are determined based on the task parameter configuration unit. Based on the task parameter configuration unit, the task trigger control instructions corresponding to each synchronization object are determined, and the task trigger control instructions are associated and encapsulated with the task priority sequence, task scheduling time interval, task execution path sequence and task batch execution sequence to generate the task execution configuration unit. The task execution configuration units corresponding to each synchronization object are summarized to construct a set of RPA robot execution tasks.

6. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The generation of the target synchronization data specifically includes: Extract the source business system identifier and target business system identifier corresponding to each synchronous task in the RPA robot's task execution set. Read the source business data in the source business system based on the source business system identifier, read the target structure data in the target business system based on the target business system identifier, and establish a data field mapping relationship between the source business system and the target business system. Based on the data field mapping relationship, perform field matching processing on the source business data to obtain a set of field matching results; Based on the field matching result set, perform data type conversion and encoding rule conversion on the source business data to generate a field conversion result set; Based on the field transformation result set and the field matching result set, perform primary key association processing and target structure reorganization processing to generate target synchronized data and write it to the corresponding target business system.

7. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The output of the synchronization result identifier specifically includes: Extract the target synchronization data and the target structure data from the target business system, and collect and process the data of each field in the target synchronization data according to the synchronization object identifier to generate a synchronization verification data unit. Based on the synchronization verification data unit, perform integrity verification and consistency verification on the target synchronization data to obtain the integrity verification result and the consistency verification result. Based on the synchronization verification data unit, perform primary key constraint verification and field value matching verification on the target synchronization data to obtain the primary key constraint verification result and the field value matching verification result. Based on the synchronization verification data unit, the time validity of the target synchronization data is verified, and the synchronization verification result is generated by combining the integrity verification result, consistency verification result, primary key constraint verification result, and field value matching verification result. Based on the synchronization verification result, a result determination process is performed, and a synchronization result identifier is output. The synchronization result identifier includes a synchronization success identifier and a synchronization error identifier.

8. The RPA robot data synchronization method based on big data analysis according to claim 1, characterized in that, The adaptive optimization of the data synchronization strategy specifically includes: Synchronization status feedback data is collected during the execution of tasks by the RPA robot, and then aggregated and processed according to the synchronization object identifier to generate a synchronization status feedback data set. The synchronization status feedback data includes task status information, synchronization verification results, synchronization result identifiers, execution duration information, resource consumption information, and abnormal event information. The synchronization state feedback data set is standardized and structured, and then used to generate model feedback data. The parameters of the improved TimeXer model are updated based on the model feedback data, and the data synchronization strategy parameter set is regenerated based on the updated model. The set of tasks to be executed by the RPA robot is regenerated based on the new set of data synchronization strategy parameters, and the RPA robot is scheduled to perform data synchronization processing.

Images