A method and system for generating management of cross-domain production processes based on big data

Abstract

The application provides a big data-based generation management method and system for cross-domain production processes, which comprises the following steps: prioritizing work units in a task load, identifying and evaluating an execution sequence adjustment scheme through a business importance level in combination with initialization configuration information, and determining a priority adjustment arrangement order; obtaining a sub-chain distribution in an execution chain structure, applying continuous execution verification to each sub-chain, determining a continuous execution compatibility evaluation result by comparing the coincidence degree of connection points and boundary activation points between sub-chains, and merging sub-chains into a complete production process path if the compatibility evaluation result meets the business requirement condition, otherwise, backtracking to the time sequence adjustment arrangement order and readjusting to obtain a final production process path; guiding a boundary activation controller to output an adjustment scheme through the final production process path, applying the execution time sequence and unit reorganization operation in the scheme to the production process, and obtaining a stable execution state after crossing the domain boundary.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a generation management method and system for cross-domain production processes based on big data. BACKGROUND

[0002] With the rapid development of cloud computing, edge computing and Internet of Things, production processes need to be seamlessly migrated between different execution domains. The generation management of cross-domain production processes based on big data determines the continuity and reliability of task execution in cross-domain business scenarios. The efficiency of cross-domain transmission of production processes directly affects the overall business response speed and user experience. However, existing methods often ignore the problem of sudden change of the running state of production processes when they approach the domain boundary, resulting in frequent suspension or restart of production processes when they enter a new domain, which seriously affects the execution efficiency. In addition, existing methods usually lack the ability to dynamically reorganize the internal working units of production processes, and cannot adjust the execution rhythm of production processes in real time according to the running environment of the new domain. This limitation is particularly evident in high-dynamic and high-load business scenarios, such as in intelligent manufacturing, where production processes on the production line need to be quickly switched between different device domains. Any delay or interruption can lead to a decrease in production efficiency. In the management of cross-domain production processes, the key technical difficulty lies in the adjustment of the execution rhythm of production processes when they cross the domain boundary. Execution rhythm refers to the speed and execution order of internal working units of production processes when they run in different execution domains. When production processes approach the domain boundary, the execution speed of production processes may change suddenly and the internal running state needs to be reinitialized due to the differences in the running environment (such as computing resources, bandwidth or protocol requirements) between the new domain and the current domain. The complexity of this rhythm adjustment stems from the dynamic adaptation problem between the content of the working load carried by the production process and the requirements of the new domain environment. For example, a production process is responsible for coordinating the cooperative work of multiple devices on a production line, including sensor data acquisition, computation and instruction issuance. When the production process migrates from a local device domain to a cloud domain, the execution speed of the production process may slow down suddenly or even be suspended due to the differences in computing resource allocation and network delay characteristics between the cloud domain and the local domain. More seriously, the order of the internal working units of the production process may not be able to adapt to the running environment of the cloud domain, resulting in the failure or repeated execution of some working units, which affects the continuous operation of the production line. Therefore, how to dynamically adjust the execution rhythm of production processes and reorganize the execution order of internal working units according to the requirements of the running environment of the new domain and the current working load content of production processes when they approach the domain boundary, to ensure the continuity and efficiency of production processes during cross-domain migration, has become a key problem in the generation management of cross-domain production processes based on big data. SUMMARY

[0003] The application provides a big data-based cross-domain production process generation management method, mainly comprising:

[0004] An identification of boundary division of a business domain to which the production process belongs is recognized, current position information and execution speed of the production process in a domain boundary monitoring area are extracted, position coordinates and progress percentage are determined, and an estimated arrival time point of the production process crossing a domain boundary line is calculated through the position coordinates and the progress percentage;

[0005] Task load content carried by the production process at the estimated arrival time point and target domain environment requirement information are obtained, matching degree of the task load content and the target domain environment requirement information is calculated, and adaptive configuration information is generated based on the matching degree being lower than a business standard value;

[0006] Work units in the task load are sorted according to priority, execution order is adjusted in combination with the configuration information and business importance level, and a priority-adjusted arrangement order is generated;

[0007] Execution order is changed and resources are re-allocated from the priority-adjusted arrangement order, work units are reorganized to generate a work unit chain, and the work unit chain is divided into sub-chains based on the length of the work unit chain exceeding a business standard length, and an optimized execution chain structure is generated;

[0008] Sub-chain distribution in the execution chain structure is obtained, continuity verification of the sub-chains is performed, and a compatibility evaluation result is generated by comparing the coincidence degree of connection points and boundary activation points between the sub-chains;

[0009] The sub-chains are fused to generate a complete production process path based on the compatibility evaluation result meeting business requirements, otherwise the priority-adjusted arrangement order is adjusted again, and a production process path is generated;

[0010] The production process path is used to control boundary adjustment execution timing and reorganize the work units, and an execution state after crossing the domain boundary is generated.

[0011] Further, the task load content carried by the production process at the estimated arrival time point and the target domain environment requirement information are obtained, the matching degree of the task load content and the target domain environment requirement information is calculated, and adaptive configuration information is generated based on the matching degree being lower than a business standard value, comprising:

[0012] The process involves parsing the header information of the production process data packet to extract a task load content list, which includes task type identifiers, data block format definitions, and execution priority values. Environmental requirement information, including the number of available computing cores and supported communication protocol versions, is obtained from the target domain controller. Feature vectors for the task load content and environmental requirement information are constructed, and a similarity score is calculated between them. A matching percentage is calculated based on the number of matches between the task type identifier and the types supported by the target domain. A comprehensive matching score is generated by combining the similarity score and the matching percentage. Since the comprehensive matching score is lower than the business standard value, the difference elements between the feature vectors of the task load content and the environmental requirement information are extracted to generate a protocol version mapping relationship and the number of task granularity divisions, and configuration information is generated.

[0013] Furthermore, based on the fact that the overall matching degree value is lower than the business standard value, the difference elements in the feature vector of the task load content and the feature vector of the environment requirement information are extracted to generate a protocol version mapping relationship and the number of task granularity divisions, and to generate configuration information, including:

[0014] Based on the fact that the overall matching degree value is lower than the business standard value, the difference elements in the feature vector of the task load content and the feature vector of the environment requirement information are extracted to generate a protocol conversion table; the resource quota is adjusted based on the difference elements to generate a resource quota allocation table; the task load is divided based on the task granularity to generate a task segmentation scheme; and the protocol conversion table, the resource quota allocation table and the task segmentation scheme are integrated to generate configuration information.

[0015] Furthermore, the step of changing the execution order and reallocating resources from the priority adjustment arrangement, reorganizing the work units to generate a work unit chain, and dividing the work unit chain into sub-chains based on the length exceeding the business standard length to generate an optimized execution chain structure includes:

[0016] The process involves: reading the new execution position of each work unit in the priority adjustment sequence; calculating the offset between the new execution position and the original position of each work unit; adjusting the processor time slice and memory space quota for work units whose offset exceeds a preset threshold; generating a resource reallocation record table; reconnecting the data transmission paths between work units based on the new execution positions in the resource reallocation record table; marking work units without dependencies as parallel groups and work units with dependencies as serial sequences; generating a work unit chain containing parallel group identifiers and serial sequence identifiers; calculating the number of work units and the total data transmission volume in the work unit chain; selecting the connection position with the smallest data transmission volume in the work unit chain to divide it into sub-chains based on the fact that the number of work units and the total data transmission volume exceed the business standard length; adding a start unit number and an end unit number to the sub-chains; recording the data flow mapping relationship between the sub-chains; and integrating the sub-chains and the data flow mapping relationship to generate an execution chain structure.

[0017] Furthermore, the calculation of the number of work units and the total data transmission volume in the work unit chain, based on the fact that the number of work units and the total data transmission volume exceed the service standard length, involves selecting the connection position with the smallest data transmission volume in the work unit chain and dividing it into sub-chains, including:

[0018] Calculate the total number of work units and the total amount of data transmission between units in the work unit chain, and identify the connection position with the smallest data transmission volume in the work unit chain; divide the work unit chain based on the connection position to generate multiple sub-chains; assign sequence numbers to the sub-chains and record the data flow mapping relationship between the sub-chains.

[0019] Furthermore, the process of identifying the boundary delineation identifier of the business domain to which the production process belongs, extracting the current location information and execution speed of the production process within the domain boundary monitoring area, determining the location coordinates and progress percentage, and calculating the estimated arrival time of the production process crossing the domain boundary line using the location coordinates and the progress percentage includes:

[0020] Obtain the geometric coordinates and real-time location coordinates of the domain boundary line of the production process within the current business domain, read the execution speed of the production process, calculate the distance between the real-time location coordinates and the nearest point of the domain boundary line, and generate the estimated arrival time based on the distance and the execution speed; extract the historical trajectory data of the production process within the monitoring area, calculate the speed change rate based on the historical trajectory data, and adjust the estimated arrival time based on the speed change rate.

[0021] Furthermore, the step of prioritizing the work units in the task load, adjusting the execution order based on the configuration information and business importance level, and generating a priority adjustment order includes:

[0022] Extract the attribute records of the work units in the task load, read the business type identifier and expected execution time of the work unit, query the importance coefficient corresponding to the business type identifier based on the business importance mapping table in the configuration information, generate a priority score based on the importance coefficient and the initial priority value, sort the work units based on the priority score, adjust the dependencies, and generate an execution plan table as the priority adjustment order.

[0023] Furthermore, the step of obtaining the distribution of sub-chains in the execution chain structure, performing continuity checks on the sub-chains, and generating compatibility evaluation results by comparing the overlap between connection points and boundary activation points of the sub-chains includes:

[0024] Read the position index and connection point information of the sub-chains in the execution chain structure, and construct a distribution mapping table containing the sub-chain sequence number and connection relationship; verify the continuity of the execution sequence number of the working unit within the sub-chain, calculate the distance between the connection point and the boundary activation point, generate an overlap degree mark based on the distance and a preset threshold, calculate the compatibility level based on the overlap degree mark, and generate a compatibility evaluation result.

[0025] Furthermore, the step of adjusting the execution timing and reorganizing the work units through the production process path control boundary to generate the execution state after crossing the domain boundary includes:

[0026] The sequence of work units and execution time points in the production process path are read and transmitted to the boundary activation controller to generate an adjustment scheme containing timing parameter adjustment values ​​and unit reorganization instructions. The execution order of work units is updated based on the adjustment scheme, the running status parameters of the production process when crossing the domain boundary are monitored, the resource quota is adjusted based on the running status parameters, and the execution status after crossing the domain boundary is generated.

[0027] A production management system for cross-domain production processes based on big data, the system comprising:

[0028] The boundary demarcation identifier identification module is used to identify the boundary demarcation identifier of the business domain to which the production process belongs, extract the current location information and execution speed of the production process in the domain boundary monitoring area, determine the location coordinates and progress percentage, and calculate the expected arrival time of the production process crossing the domain boundary line using the location coordinates and the progress percentage.

[0029] The location and speed extraction module is used to obtain the task load content and target domain environment requirement information carried by the production process at the expected arrival time, calculate the matching degree between the task load content and the target domain environment requirement information, and generate adaptive configuration information based on the matching degree being lower than the business standard value.

[0030] The matching degree evaluation module is used to sort the work units in the task load by priority, adjust the execution order in combination with the configuration information and business importance level, and generate a priority adjustment and arrangement order;

[0031] The priority sorting module is used to change the execution order and reallocate resources from the priority adjustment and arrangement order, reorganize the work units to generate a work unit chain, divide the work unit chain into sub-chains based on the length of the work unit chain exceeding the business standard length, and generate an optimized execution chain structure.

[0032] The execution chain reorganization module is used to obtain the distribution of sub-chains in the execution chain structure, perform continuity verification on the sub-chains, and generate compatibility evaluation results by comparing the overlap between connection points and boundary activation points between sub-chains.

[0033] The compatibility assessment module is used to integrate the sub-chains to generate a complete production process path based on the compatibility assessment results, if the business requirements are met; otherwise, it backtracks to the priority adjustment and rearrangement order to re-adjust and generate a production process path.

[0034] The path fusion module is used to adjust the execution timing and reorganize the work units according to the production process path control boundary, generating the execution state after crossing the domain boundary. The technical solution provided by this embodiment of the invention can include the following beneficial effects:

[0035] This invention discloses a method and system for managing cross-domain production processes based on big data. Addressing the issues of low execution efficiency and process interruptions caused by mismatches between business content and new domain environment requirements when a production process crosses a business domain boundary, the invention dynamically identifies the current position and execution speed of the production process, accurately assesses the time required to cross the boundary, and extracts the task load content for matching analysis with the new domain requirements. If the matching degree is below the standard, the invention automatically initializes the adaptation configuration, prioritizes work units based on business importance, reorganizes the execution chain, and optimizes resource allocation. Through continuity verification and sub-chain fusion, this invention ensures that the production process forms a stable and coherent execution path after crossing domains, significantly improving the adaptability and execution efficiency of cross-domain production processes, reducing the risk of process interruptions, and providing an efficient and reliable solution for complex business scenarios. Attached Figure Description

[0036] Fig. 1 This is a flowchart of a cross-domain production process generation management method based on big data according to the present invention.

[0037] Fig. 2 This is a schematic diagram of the structure of a cross-domain production process generation management system based on big data according to the present invention. Detailed Implementation

[0038] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0039] like Figs. 1-2 This embodiment of a method and system for managing cross-domain production processes based on big data may specifically include:

[0040] S101. Identify the boundary demarcation marker of the business domain to which the production process belongs, extract the current location information and execution speed of the production process within the domain boundary monitoring area, extract the location coordinates and progress percentage, and evaluate the estimated arrival time point when the process is about to cross the boundary line using the location coordinates and progress percentage.

[0041] The process acquires the set of geometric coordinates of the domain boundary line within the current business domain and the real-time position coordinates of the production process actuators. It reads the current execution speed V and three-dimensional coordinates (X, Y, Z) of the production process, calculates the Euclidean distance D between the current position of the production process and the nearest point on the domain boundary line, and obtains the preliminary time T for the production process to reach the domain boundary and the speed V of the production process reaching the domain boundary using the formula T=D / V. It extracts the historical trajectory data and corresponding speed values ​​of the production process from the ten most recent sampling points within the monitoring area, calculates the rate of change of speed between adjacent sampling points, and if the rate of change of speed exceeds a preset threshold, it uses the least squares method to fit the speed sequence. Based on the fitted curve, it predicts the speed V' of the production process when it reaches the domain boundary, and updates the arrival time using D / V' to obtain the corrected estimated arrival time. Based on the revised estimated arrival time and the current progress percentage P of the production process, calculate the ratio of the remaining workload W of the production process to the current execution speed to obtain the time required to complete the remaining work. If the time required to complete the remaining work is less than the revised estimated arrival time, output the revised estimated arrival time as the moment when the production process crosses the domain boundary line; otherwise, output the time required to complete the remaining work as the crossing time.

[0042] In one embodiment, the set of domain boundary geometric coordinates is constructed by pre-deploying domain boundary markers in the production process execution environment.

[0043] Specifically, each domain boundary marker contains three-dimensional spatial coordinate information, and adjacent markers form a continuous interface through linear interpolation. The production process actuators have built-in position sensor modules that collect real-time position data of the production process within the execution domain at a fixed frequency. The Euclidean distance D is calculated using the standard three-dimensional spatial distance formula, i.e. , where (Xb,Yb,Zb) are the coordinates of the point on the domain boundary line that is closest to the current position in the production process.

[0044] It should be noted that the monitoring area refers to the buffer zone formed by extending a certain range on both sides of the domain boundary line. The operation status of the production process within this area will be closely monitored.

[0045] For example, in a smart manufacturing scenario, when the production process migrates from the local computing domain in the workshop to the edge server domain, the monitoring area covers a physical space of 50 meters before and after the workshop gateway. Historical trajectory data includes multi-dimensional information such as the location coordinates, execution speed, and resource utilization of the production process at each sampling time.

[0046] In one possible implementation, the rate of change of speed is calculated by dividing the speed difference between two adjacent sampling points by the time interval. When the rate of change of speed exceeds a preset threshold, it indicates that the production process may encounter network congestion or computing resource limitations. At this time, the least squares method is used to perform a quadratic polynomial fitting on the historical speed sequence, with the fitting function being v(t) = at² + bt + c. By solving the system of equations, the values ​​of coefficients a, b, and c are obtained, and the speed V' of the production process at the domain boundary is predicted.

[0047] Preferably, the remaining workload W is calculated by the number of unfinished task units carried by the production process and the estimated processing time of each task unit.

[0048] For example, a data processing workflow involves processing 1000 data blocks. 600 have been completed, so the progress percentage P is 60%, and the remaining workload W is the total processing time for the remaining 400 data blocks. By comparing the time required to complete the remaining work with the estimated time to reach the domain boundary, the system can determine whether the workflow can complete the current batch of tasks before crossing the boundary, thus deciding whether to save the task status or adjust the workload in advance.

[0049] S102. Obtain the task load content and new domain environment requirements information carried by the production process at the expected arrival time, evaluate the matching degree between the task load content and the new domain environment requirements, and determine if the matching degree is lower than the business standard value, then initialize the configuration information for new domain environment adaptation.

[0050] The system reads the header information of the production process data packet at the expected arrival time, parses the task load content list, which includes task type identifiers, data block format definitions, execution priority values, and estimated resource consumption. It obtains environmental requirement information from the target domain controller, including the number of available computing cores, network transmission bandwidth limit, supported communication protocol versions, and maximum task queue depth, and constructs task feature vectors and environmental feature vectors accordingly. The cosine similarity between the task feature vector and the environmental feature vector is calculated as a similarity score. The number of matches between the task type identifier and the types supported by the target domain is counted, and the percentage of matches is calculated as a percentage of the total number of tasks. The overall matching degree is obtained using the formula: Overall Matching Degree = 0.6 × Similarity Score + 0.4 × Percentage. If the overall matching degree is lower than a preset business standard threshold, the system extracts the discrepancies between the task feature vector and the environmental feature vector. For each discrepancy element, corresponding adaptation parameters are generated, including protocol version mapping relationships, resource allocation adjustment coefficients, and the number of task granularity divisions. These adaptation parameters are combined to form a configuration parameter set. A protocol conversion table is generated based on the protocol version mapping relationship in the configuration parameter set. The resource quota of each task is recalculated based on the resource allocation adjustment coefficient. Large tasks are divided according to the number of task granularity divisions. The protocol conversion table, resource quota allocation table and task division scheme are integrated to output the new domain environment adaptation configuration information.

[0051] Specifically, the header information of the production process data packet is stored in a fixed-length structured format, containing basic metadata such as a unique production process identifier, creation timestamp, source domain identifier, and total number of tasks. The task payload list is stored in the payload segment of the data packet. Each task entry occupies a fixed number of bytes and contains four core fields: a task type identifier encoded as a 32-bit integer to distinguish different task categories such as data processing, control commands, and state synchronization; a data block format definition using 8-bit encoding to represent the data serialization method, such as JSON, Protocol Buffer, or a custom binary format; an execution priority value ranging from 0 to 255, with higher values ​​indicating higher priority; and an estimated resource consumption measured in standard computing units, recording the number of CPU cycles and memory usage required for task execution.

[0052] In one embodiment, the target domain controller maintains a real-time updated resource status table that records the current domain's operating environment parameters. The number of available computing cores is obtained by reading processor configuration information, including the number of physical cores and virtual cores; the maximum network transmission bandwidth is read from the configuration register of the network interface controller; the supported communication protocol version numbers are stored in the protocol stack configuration file, covering various protocol types such as TCP / IP, UDP, and custom real-time transport protocols; the maximum task queue depth is dynamically calculated based on memory capacity and the average size of a single task.

[0053] It should be noted that the construction process of the task feature vector and the environment feature vector adopts a normalization method.

[0054] For example, in a smart factory production line control scenario, a typical task feature vector includes the following dimensions: task type distribution ratio, average data block size, priority distribution variance, and total resource requirement. The corresponding environmental feature vector includes: processing capacity for each type of task, maximum data throughput, priority scheduling strategy parameters, and total available resources. The values ​​of each dimension are normalized to the [0,1] interval for easier subsequent similarity calculation. The formula for calculating cosine similarity is... The calculation results reflect the degree of matching between task requirements and environmental capabilities.

[0055] Preferably, the overall matching degree is calculated using a weighted average method, with the weight coefficients dynamically adjusted according to different business scenarios. In scenarios with high real-time requirements, the weight of the similarity score can be increased to 0.7, while in batch processing scenarios, the weight of the type matching percentage can be increased to 0.6. The business standard threshold is typically set between 0.65 and 0.85, with the specific value determined based on the business's fault tolerance and performance requirements.

[0056] In one possible implementation, the extraction of difference elements is achieved through dimension-by-dimensional comparison. When the difference in a certain dimension exceeds a preset tolerance range, that dimension is marked as a difference element.

[0057] For example, when a task requires the use of the HTTP / 2 protocol but the target domain only supports HTTP / 1.1, the protocol version becomes a differentiating element. To address this protocol version difference, the adaptation parameter generator creates a protocol version mapping relationship and defines the conversion rules from HTTP / 2 requests to HTTP / 1.1 requests, including specific conversion logic such as header field mapping, multiplexing splitting, and flow control parameter adjustment.

[0058] Understandably, the resource allocation adjustment coefficient is calculated based on the ratio of the actual task demand to the available resources in the target domain. When the task demand exceeds the available resources, the adjustment coefficient is less than 1, indicating that resource allocation needs to be reduced; conversely, the adjustment coefficient can be greater than 1, indicating that resource allocation can be increased to improve execution efficiency. The number of task granularity divisions is determined based on the complexity of a single task and the parallel processing capability of the target domain. Complex tasks can be divided into 2 to 10 subtasks for parallel execution. Furthermore, the protocol conversion table uses a key-value pair structure for storage, where the key is the message type identifier of the source protocol, and the value is the corresponding message format template of the target protocol. The resource quota allocation table records resource allocation information such as CPU time slices, memory quotas, and network bandwidth quotas for each task or subtask. The task partitioning scheme includes information such as the partitioning point location, subtask dependency graph, and merging strategy.

[0059] For example, in a production process containing 1000 data processing tasks, if the target domain is detected to support only 500 concurrent tasks, the task partitioning scheme automatically divides it into two batches, each containing 500 tasks, and marks the execution order and data dependencies between batches in the configuration information. Through this fine-grained adaptation configuration, the production process can run smoothly in a resource-constrained new domain environment, avoiding execution interruptions or performance degradation caused by resource contention.

[0060] In one embodiment, the complete new domain environment adaptation configuration information is output in the form of a structured document, including configuration version number, effective time, applicable domain identifier, detailed configuration item list, etc. Each configuration item has a corresponding check code to ensure the integrity and correctness of the configuration information during transmission and application.

[0061] Read the list of business content and execution requirements carried by the production process that is about to cross the boundary, obtain the target business domain's acceptance standards and processing capability boundaries for task types, compare the business types in the task list with the target domain's business acceptance range, check whether the task execution requirements are consistent with the target domain's processing standards, identify the business content items that differ, count the proportion of the number of differing items to the total task content, and determine the degree of adaptation between the production process and the new domain environment based on the proportion.

[0062] Extract a list of business content from the production process data structure that is about to cross the boundary. Read the business type codes, business processing identifiers, data block sizes, and computational complexity coefficients. Parse the execution requirement configuration file to obtain the response time limit, maximum concurrent requests, transaction integrity level, and failure retry policy. Read the supported business type code set, maximum single-task processing time, maximum concurrent processing capacity, and memory usage threshold from the target business domain registry. Use set difference operations to identify business type codes in the business content list that are not supported by the target domain, forming a type incompatibility list. Compare the response time limit in the execution requirement configuration file with the maximum single-task processing time of the target domain, and compare the maximum concurrent requests with the maximum concurrent processing capacity. Record items exceeding the target domain's processing capacity to form a capacity over-limit list. Merge the type incompatibility list and the capacity over-limit list to obtain a total difference table. Calculate the percentage R of the number of entries in the total difference table to the total number of entries in the business content list. If R is less than 0.2, the fit is set to high; if R is between 0.2 and 0.6, the fit is set to medium; and if R is greater than 0.6, the fit is set to low. Based on the compatibility value and the specific content of the difference item summary table, targeted processing suggestions are generated using preset mapping rules. The processing suggestions include business types that require protocol conversion, high-concurrency requests that require downgrade processing, and large data block tasks that require batch execution. The output is a complete judgment result containing the compatibility value and processing suggestions.

[0063] In one embodiment, the business content list is stored in the description segment of the production process using a structured metadata description method.

[0064] Specifically, each business entry contains four core attributes: the business type encoding adopts a hierarchical encoding system, such as "01.02.03" representing streaming aggregation business in the real-time computing subclass under the data processing category; the business processing identifier is used to distinguish different processing instances under the same type; the data block size records the amount of data involved in a single business operation, in bytes; the computational complexity coefficient reflects the computational intensity of the business, with a value ranging from 1 to 10, where a larger value indicates higher computational complexity. The execution requirement configuration file is stored in key-value pair format, the upper limit of response time is in milliseconds, the maximum number of concurrent requests represents the upper limit of the number of requests processed simultaneously, the transaction integrity level is divided into three levels: strong consistency, eventual consistency, and weak consistency, and the failure retry strategy includes three modes: exponential backoff, fixed interval, and immediate retry.

[0065] It should be noted that the target business domain registry is a dynamically maintained configuration database that reflects the processing capacity status of the target domain in real time. Supported business type encoding sets are stored in hash table format for easy searching and matching; the maximum processing time for a single task is dynamically updated based on statistical analysis of historical execution data; the upper limit of concurrent processing capacity is calculated by comprehensively considering CPU utilization, memory usage, and network bandwidth usage collected in real time by the monitoring system; the memory usage threshold is set at 80% of the total physical memory, with 20% reserved as a system buffer.

[0066] For example, in a cross-domain collaborative scenario of intelligent manufacturing, when the production control process migrates from the workshop edge node to the regional computing center, the execution process of the set difference operation is as follows: First, a source domain business type set S={01.01.01,01.01.02,01.02.01,01.02.03} is constructed, and a target domain supported type set T={01.01.01,01.02.01,01.03.01} is constructed. The ST operation yields a type incompatibility list {01.01.02,01.02.03}. This indicates that the target domain does not support real-time monitoring business 01.01.02 and streaming aggregation business 01.02.03, requiring business downgrading or replacement. During the capability comparison process, if a task requires a 200-millisecond response time while the target domain's maximum single-task processing time is 500 milliseconds, this item is recorded in the capability over-limit list.

[0067] In one possible implementation, the total difference table is organized using a linked list data structure. Each node contains five fields: difference type, difference item identifier, original requirement value, target domain capability value, and difference degree. Difference types are divided into two categories: business type incompatibility and processing capacity exceeding limits. The difference degree is calculated by determining the relative deviation between the original requirement value and the target domain capability value. The percentage R is calculated by counting the number of nodes (diff) in the total difference table, obtaining the total number of items in the business content list (total), and then calculating R = diff / total. The suitability value is determined using a piecewise function: when R ∈ [0, 0.2), the suitability value is set to high, indicating that most businesses can be directly migrated; when R ∈ [0.2, 0.6), the suitability value is set to medium, indicating that some adjustments are needed; and when R ∈ [0.6, 1], the suitability value is set to low, indicating that significant reconstruction is required.

[0068] Preferably, the preset mapping rule base includes three types of processing rules: protocol conversion rules, degradation processing rules, and task splitting rules. Protocol conversion rules define the conversion relationship between different business type codes, such as mapping real-time monitoring business 01.01.02 to batch collection business 01.03.02, achieving near real-time performance through a caching mechanism. Degradation processing rules target high-concurrency requests, reducing instantaneous load through rate limiting, queuing, and batch processing. Task splitting rules divide large data block tasks into multiple smaller tasks, with the data volume of each smaller task not exceeding the single-processing limit of the target domain.

[0069] Understandably, the process of generating processing recommendations needs to take into account both the specific characteristics of the differences and the actual capabilities of the target domain.

[0070] For example, for streaming aggregation business type 01.02.03, if the target domain does not support but does support batch aggregation business 01.03.03, the processing suggestions include caching the streaming data to a certain amount and then converting it to batch processing; adjusting the aggregation time window from seconds to minutes; and using micro-batch processing to simulate the effect of streaming processing. Furthermore, for cases of exceeding capacity limits, the processing suggestions will formulate different strategies based on the degree of exceedance. Mild exceedance (less than 20% of capacity) uses optimization strategies, such as compressing data transmission and simplifying calculation logic; moderate exceedance (20%-50% of capacity) uses degradation strategies, such as reducing processing precision and extending response time; severe exceedance (more than 50% of capacity) uses a decomposition strategy, breaking the task down into multiple sub-tasks for batch execution.

[0071] In one embodiment, the complete judgment result is output in the form of a structured report, comprising four parts: an overall assessment of the compatibility level, displaying high, medium, and low compatibility values; a detailed list of discrepancies, listing all incompatible and out-of-limit items; targeted handling suggestions, providing specific adaptation solutions for each discrepancy; and execution priority ranking, prioritizing the handling suggestions according to business importance. Through this refined discrepancy analysis and handling suggestion mechanism, the production process can fully assess the compatibility of the target domain before cross-domain migration, formulate corresponding adaptation strategies in advance, and avoid business interruptions or severe performance degradation during the migration process.

[0072] S103. Prioritize the work units in the task load, and combine the initialization configuration information with the business importance level to identify and evaluate the execution order adjustment plan, and determine the priority adjustment order.

[0073] Extract attribute records for all work units from the task load data structure, read the unit identifier, initial priority value, business type identifier, and estimated execution time. Obtain the business importance mapping table from the initialization configuration information. This mapping table records the importance coefficient corresponding to each business type identifier. Query the importance coefficient corresponding to the business type of each work unit. Calculate the priority score as the initial priority value plus the importance coefficient multiplied by a preset weighting factor. Arrange all work units in descending order according to the priority score to form an initial sequence. Scan the pre-dependency table between work units. If work unit A depends on the output of work unit B, adjust B's position before A. Check for dependency loops during the adjustment process. If a loop is detected, remove the dependency edge with the lowest priority score in the loop. Generate an execution plan table based on the adjusted work unit sequence. The execution plan table includes the execution order number of each work unit, the delay time relative to the production process start time, and the resource allocation priority flag. Output the execution plan table as the priority adjustment and arrangement order.

[0074] Specifically, the attribute records of work units are stored in the metadata area of ​​the production process, and each work unit contains four basic attribute fields. The unit identifier uses a unique 32-bit integer encoding, with an initial priority value ranging from 0 to 100. The business type identifier uses a three-level classification encoding system, such as "A01B02C03" representing the aggregation operation type in the batch processing subclass under the data processing category, and the expected execution time is recorded in milliseconds.

[0075] In one embodiment, the business importance mapping table is implemented using a hash table data structure, where the key is the business type identifier and the value is the importance coefficient. The importance coefficient ranges from 0.5 to 2.0, with the coefficient for critical business types greater than 1.5, the coefficient for ordinary business types between 0.8 and 1.2, and the coefficient for auxiliary business types less than 0.8. The priority score is calculated using a linear combination method, specifically as: Priority Score = Initial Priority Value + Importance Coefficient × Weighting Factor. The weighting factor is dynamically adjusted based on the overall load of the production process and is typically set between 20 and 50.

[0076] It should be noted that the prerequisite dependency table records the execution dependency constraints between work units.

[0077] For example, in the production process of intelligent manufacturing, the data acquisition unit must be executed before the data processing unit, and the data processing unit must be executed before the control command issuance unit. These dependencies form a directed graph structure. Dependency loop detection uses a depth-first search algorithm, traversing its dependency path starting from each work unit. If the same node is visited repeatedly during the traversal, it indicates the existence of a loop.

[0078] Preferably, when a dependency loop is detected, the system identifies the priority score of all working units in the loop, selects the unit with the lowest score, and removes one of its outgoing or incoming edges to break the loop.

[0079] For example, if unit A depends on B, B depends on C, and C depends on A to form a loop, and C has the lowest priority score, then remove the dependency edge from C to A so that C can be executed in parallel with A.

[0080] In one possible implementation, the execution plan table is stored as a structured array, with each array element corresponding to the execution information of a work unit. Execution order numbers are assigned incrementally starting from 1, and relative latency is calculated based on the cumulative execution time of preceding units. Resource allocation priority is categorized into high, medium, and low levels, with high-priority units receiving the resources they need first during resource contention. Through this mechanism, the execution order of work units can be reorganized according to the characteristics of the new domain when the production process migrates across domains.

[0081] S104. Select execution order changes and resource reallocation from the priority adjustment and sorting order, reorganize the execution order of work units, generate the reorganized work unit chain, and determine if the chain length exceeds the business standard length, then divide it into sub-chains to obtain the optimized execution chain structure.

[0082] The new execution position and original position of each work unit are read from the priority adjustment and sorting order. The position offset is calculated, and work units whose position changes exceed a preset offset threshold are identified. For these work units, processor time slices, memory space quotas, and network bandwidth ratios are reallocated, forming a resource reallocation record table containing unit identifiers and new resource quotas. Based on the new execution position information in the resource reallocation record table, the data transmission paths between work units are reconnected. Work units without dependencies are marked as parallelizable groups, and work units with dependencies are marked as serial sequences. All work units are arranged from front to back according to their execution positions, constructing a reorganized work unit chain containing parallel group identifiers and serial sequence identifiers. The total number of work units N and the total data transmission volume D between units in the reorganized work unit chain are calculated. If N+D / preset transmission rate exceeds the service standard length threshold, the connection position with the minimum data transmission volume is found in the chain as a split point. At the split point, the chain is broken to form multiple sub-chains, each containing a continuous sequence of work units. Add a start unit number, an end unit number, and a sub-chain sequence number to each sub-chain. Record the mapping relationship between the output data identifier of the last unit of the previous sub-chain and the input data identifier of the first unit of the next sub-chain. Integrate all sub-chain information and mapping relationships to obtain an optimized execution chain structure that includes the sub-chain set and the data flow direction between chains.

[0083] Specifically, the position offset is calculated based on the difference in the execution order of the work unit before and after the priority adjustment.

[0084] In one embodiment, if a work unit was originally executed at position 10 and is moved to position 3 after adjustment, the position offset is 7. A preset offset threshold is typically set to 20% of the total number of work units. When the offset exceeds this threshold, it indicates a significant change in the execution urgency of the unit, requiring a reassessment of its resource requirements. The resource reallocation record table is stored in triples, containing a unique unit identifier, a resource type code, and a new quota value. Processor time slices are in milliseconds, memory space is in megabytes, and network bandwidth is in megabits per second.

[0085] It's important to note that reconnecting data transmission paths involves replanning the data flow between existing work units. In cross-domain production processes within smart manufacturing, some data acquisition units that were originally executed serially can now execute multiple acquisition tasks simultaneously in the new domain due to their increased parallel processing capabilities. The method for identifying no-dependency relationships is to check if there is a direct correlation between the data input and output of two work units. If the output of unit A is not the input of unit B, and the output of B is not the input of A, then they are considered to have no dependency and can be executed in parallel. Parallel group identifiers are represented using bitmasks, and work units within the same parallel group share the same group identifier code. The serial sequence is maintained using a linked list structure, with each node recording pointers to the current unit and its direct successor units.

[0086] In one possible implementation, the process of constructing the reorganized work unit chain first sorts all work units according to their new execution positions, and then analyzes the relationships between adjacent units one by one. If two adjacent units belong to the same parallel group, the start and end positions of the parallel segment are marked in the chain; if they belong to a serial sequence, the direction and data type of data transmission are recorded. The reorganized work unit chain is stored in the form of an adjacency list, with each node containing unit information, resource quota, parallel / serial identifier, and a pointer to the next node. This data structure facilitates the rapid location of the predecessor and successor relationships of any work unit during execution, enabling dynamic scheduling.

[0087] For example, the chain length calculation considers not only the number of work units but also the data transmission overhead between units. The specific calculation method is to first count the total number N of all work units in the chain, and then sum the data transmission volume D between all adjacent units. The data transmission volume includes the size of the transmitted data block and the overhead of the transmission protocol header. The preset transmission rate is dynamically determined based on the network environment of the target domain, typically set to 100MB / s in a Gigabit Ethernet environment. The service standard length threshold is set considering the processing capacity and memory capacity of the target domain, generally set to the maximum number of work units that can be processed in a single operation, such as 1000 units. When the calculated N+D / transmission rate exceeds this threshold, it indicates that a single chain is too long, which may cause execution delays or memory overflow.

[0088] Preferably, the process of finding the split point employs a sliding window mechanism to traverse the entire chain of work units. The window size is set to 5 adjacent work units, and the total data transfer volume between the middle unit and the units before and after it is calculated at each window position. The data transfer volume is obtained by accumulating the input data size and the output data size, and the position with the minimum transfer volume is selected as the candidate split point. Simultaneously, it is also necessary to check whether there are strong dependencies at this position, such as transaction consistency requirements or atomic operation constraints. If so, this position is skipped, and the search continues for the next candidate point. The selected split point must satisfy two conditions: the data transfer volume is locally minimized, and it does not compromise the integrity of the business logic.

[0089] Understandably, each subchain needs to maintain relatively independent execution capabilities after being split. The starting unit number records the global number of the first working unit in the subchain, and the ending unit number records the global number of the last working unit. Subchain numbers are assigned incrementally starting from 1. This identification information is stored in the metadata header of the subchain, facilitating rapid location and scheduling during execution. Furthermore, establishing inter-chain mapping relationships requires precise recording of data flow. The output data identifier of the last unit of the previous subchain includes the data type, data format, and data version number, while the input data identifier of the first unit of the next subchain includes the expected data type, receive buffer address, and data checksum. The mapping relationship table is stored using a hash table, with the output data identifier as the key and the input data identifier as the value, ensuring that data can be accurately transferred from one subchain to the next.

[0090] In one embodiment, the optimized execution chain structure is organized in the form of a hierarchical graph. The top layer is an index table of the sub-chain set, the middle layer contains the internal structure of each sub-chain, and the bottom layer contains specific work unit information. This hierarchical structure allows for flexible scheduling of the production process when executing across domains. It enables batch migration by sub-chain and dynamic adjustment of the execution order of sub-chains based on resource availability. Through this mechanism, even in resource-constrained new domain environments, the production process can maintain high execution performance and avoid execution bottlenecks caused by excessively long chains.

[0091] Based on the importance of the business, select the work units that need to be adjusted from the sorting results, move the high-priority units to the front of the execution queue, rearrange the allocation order of processing resources to ensure that key business units have priority processing rights, adjust the dependencies and execution timing between units, and combine related business units into work groups for continuous processing to form a new execution chain structure, ensuring the continuity of business processes and processing efficiency.

[0092] The importance level and current execution position of each work unit are read from the sorting results. Units with importance levels greater than a preset importance threshold are extracted as key business units. The distance between the current position of each key business unit and the head of the queue is calculated as the move-forward amount, and the units are arranged in descending order of the move-forward amount to form an adjustment sequence. Key business units are moved to the front of the execution queue one by one according to the adjustment sequence. The execution index number of each unit after the move is updated. The number of processor cores, memory capacity, and network bandwidth resources are reallocated according to the new index number. The correspondence between unit identifiers and new resource quotas is recorded to form a resource configuration table. The input-output dependencies between work units in the resource configuration table are scanned. Dependency pairs whose execution order is reversed due to position adjustments are identified. The start time of dependent units is adjusted so that it is later than the completion time of the dependent unit. A circular dependency exists where unit A depends on B and B also depends on A. If such a circular dependency exists, the lower priority dependency edge is disconnected. Based on the adjusted dependencies, identify work units with the same business type encoding or shared data source, combine the identified related units into work groups and assign them continuous execution time periods, connect the input and output interfaces of each work group to form a complete execution chain structure, and output chain configuration information containing the unit sequence within the group and the connection relationship between the groups.

[0093] In one embodiment, the business importance level value is represented by an integer range from 0 to 100, where 80 and above are critical businesses, 50 to 80 are ordinary businesses, and below 50 are auxiliary businesses. The preset importance threshold is typically set to 75, which can be dynamically adjusted according to the overall load of the production process. The forward movement amount is calculated by subtracting the current position from the total queue length and adding 1; the resulting value represents the number of positions the unit needs to move forward. The arrangement of the adjustment sequence follows a forward movement priority principle, with units having larger forward movement amounts being adjusted first.

[0094] It should be noted that the execution queue is implemented using a doubly linked list structure, where each node contains complete information about the work unit and pointers to its predecessor and successor. Move operations are performed by modifying the pointer relationships between nodes in the linked list; the target node is first disconnected from its original position and then inserted into the head of the queue. Execution index numbers are assigned consecutively starting from 1, and each position adjustment requires traversing the entire queue and renumbering the nodes. Resource allocation follows a decreasing priority principle, with units at the front of the queue receiving more resource allocation.

[0095] Specifically, the number of processor cores is allocated using a proportional allocation method, with the total number of cores weighted according to the importance level of each unit.

[0096] For example, in an 8-core processor environment, a unit with an importance level of 90 can be allocated 3 cores, a level of 70 can be allocated 2 cores, and a level of 50 can be allocated 1 core. Memory capacity allocation takes into account the data processing volume of the work unit, with a base memory quota of 512MB, increasing by 256MB for every 10 levels. Network bandwidth resources are allocated according to the data transmission needs of the unit: real-time data transmission units receive 40% of the total bandwidth, batch data transmission units receive 30%, control command transmission units receive 20%, and the remaining 10% is used as a buffer. The resource allocation table is stored using a hash mapping structure, with the unit identifier as the key and the resource quota triplet as the value.

[0097] For example, in the cross-domain production process scenario of intelligent manufacturing, the input-output dependency relationship is manifested as a data flow relationship. The output of the sensor data acquisition unit is the input of the data preprocessing unit, and the output of the preprocessing unit is the input of the feature extraction unit. This serial dependency must maintain the execution order. When a key business unit is moved forward, the feature extraction unit may appear before the data preprocessing unit. In this case, the start time of the feature extraction unit needs to be adjusted. The start time is calculated based on the expected completion time of the dependent unit, and the specific formula is: Start Time = Start Time of Dependent Unit + Execution Duration of Dependent Unit + Data Transmission Delay. The data transmission delay is calculated based on the data volume and network bandwidth, and is usually between 10 and 100 milliseconds. Circular dependency detection uses a directed graph loop detection algorithm, performing a depth-first traversal starting from each node, and using an access marker array to record the traversal path. If the same node is visited repeatedly on a path, a loop exists.

[0098] Preferably, a priority comparison strategy is used to handle circular dependencies. When a circular dependency is detected between unit A and unit B, their business importance level values ​​are compared. The dependency relationship of the unit with the higher level value is retained, and the dependency edge from the unit with the lower level value to the other party is disconnected.

[0099] For example, the data validation unit depends on the output of the data correction unit for validation, while the data correction unit also needs the validation results to determine the correction strategy, forming a loop. If the validation unit level is 85 and the correction unit level is 60, then the dependency between the correction unit and the validation unit is broken, allowing the correction unit to execute using the default strategy.

[0100] In one possible implementation, the business type coding adopts a three-level classification system. The first level represents the major business category, such as "DAT" for data processing, "CTL" for control instructions, and "MON" for monitoring and data acquisition. The second level represents the specific business type, and the third level represents the business version. Identification of shared data sources is achieved by comparing the input data identifiers of work units. If multiple units have the same input data identifier or an inclusion relationship, they are considered to share a data source.

[0101] Understandably, the process of grouping workgroups needs to consider the execution characteristics of the units within the group. Units with the same business type coding typically have similar resource requirements and execution patterns. Grouping them together allows them to share resource pools and execution contexts, reducing switching overhead. The allocation of continuous execution time periods is based on the cumulative execution duration of the units within the group, with a 20% buffer time reserved to handle execution latency. Inter-group connections are implemented by defining input / output interfaces, with the output of the last unit in the previous workgroup serving as the input of the first unit in the next workgroup. Furthermore, the chain configuration information includes three layers: the workgroup layer records the group identifier, the number of units within the group, and the estimated execution duration; the unit layer records the relative position of each unit within the group, resource quotas, and dependencies; and the interface layer defines the inter-group data transfer protocol, buffer size, and timeout handling strategy. This layered structure makes the execution chain highly maintainable and scalable; when business requirements change, only the configuration of the corresponding layer needs to be adjusted.

[0102] S105. Obtain the distribution of sub-chains in the execution chain structure, apply continuous execution verification to each sub-chain, and determine the continuous execution compatibility evaluation result by comparing the overlap between the connection points and boundary activation points of the sub-chains.

[0103] The system reads the position indices and quantity statistics of all sub-chains in the execution chain structure, obtains the starting unit number, ending unit number, and number of working units contained in each sub-chain, extracts the connection point information between sub-chains, and records the output interface identifier of the last unit of the previous sub-chain and the input interface identifier of the first unit of the next sub-chain. A distribution mapping table containing sub-chain sequence numbers, position information, and connection relationships is constructed. A continuity check is performed on each sub-chain recorded in the distribution mapping table to verify whether the execution sequence numbers of the working units within the sub-chain are continuously increasing, and to detect whether there are any missing or duplicate sequence numbers between units. The system reads the position coordinates of activation points from the domain boundary configuration; the activation points are preset access points on the domain boundary that allow data transmission. The Euclidean distance between the connection point coordinates and the coordinates of each activation point is calculated. The minimum value of the Euclidean distance is compared with a preset overlap threshold. If the minimum distance is less than one-third of the threshold, it is marked as completely overlapped. If it is between one-third and two-thirds, it is marked as partially overlapped. If it is greater than two-thirds, it is marked as not overlapped. The percentage of each type of overlap is counted. The compatibility level is determined based on the percentage of completely overlapped exceeding a preset percentage. The continuous execution compatibility evaluation result, which includes the compatibility level and the overlap category of each connection point, is output.

[0104] In one embodiment, the position index of the sub-chain uses a globally unique number, numbered consecutively starting from the beginning of the execution chain. The start and end unit numbers of each sub-chain form a closed interval, representing the range of working units contained in that sub-chain. The connection point information includes two key parts: the output interface identifier is represented by a four-tuple, including the sub-chain number, unit number, data type code, and port number; the input interface identifier also uses a four-tuple format, recording the corresponding information of the receiver.

[0105] It should be noted that the distribution mapping table is stored using a two-dimensional matrix structure. The row index corresponds to the sub-chain sequence number, and the column index contains a position information field and a connection relationship field. The position information field records the relative percentage of the sub-chain's position in the execution chain, and the connection relationship field stores the coordinates of the connection points with other sub-chains. The core of the continuity check lies in verifying the integrity of the execution sequence number. This is done by traversing the sequence numbers of all work units within the sub-chain and checking whether they satisfy the n+1 increment rule.

[0106] Specifically, activation points are pre-configured data transmission access points on domain boundaries, each with fixed two-dimensional coordinates. In cross-domain production process scenarios, activation points are typically deployed at gateway nodes, router interfaces, or dedicated data channel entry points. Euclidean distance is calculated using the standard two-dimensional distance formula. Let (x1, y1) be the coordinates of the connection point and (x2, y2) be the coordinates of the active point. Calculate the distance from each connection point to all active points, and select the minimum value as the matching distance for that connection point.

[0107] Preferably, the three-level overlap classification standard is set based on actual business needs. Complete overlap indicates that the distance between the connection point and the active point is less than one-third of the threshold, typically corresponding to directly connected physical ports; partial overlap indicates that a small number of routing forwards are required to reach the point; no overlap indicates that multiple hops or protocol conversions are required. The preset overlap threshold is generally set to 100 coordinate units, which can be adjusted according to the density of the network topology.

[0108] For example, in a smart manufacturing scenario, if there are 5 sub-chains, with 3 connection points completely overlapping, 1 partially overlapping, and 1 not overlapping, then the percentage of complete overlap is 60%. When the preset percentage is set to 50%, the compatibility level of this execution chain is determined to be high. The evaluation results are output in the form of structured data, including the overall compatibility level, a list of overlap categories for each connection point, and an overlap statistics report, providing a quantitative basis for cross-domain migration decisions in the production process.

[0109] S106. If the compatibility assessment results meet the business requirements, the sub-chains are merged into a complete production process path; otherwise, the time sequence is backtracked to adjust the arrangement order and readjusted to obtain the final production process path.

[0110] The system reads the compatibility level value and overlap statistics of each connection point from the compatibility assessment results. It then obtains preset business requirement conditions, including a minimum compatibility level threshold and an overlap percentage threshold. The system compares the compatibility level value with the minimum compatibility level threshold. If the compatibility level value reaches the threshold and the overlap percentage exceeds the preset value, the business requirement is deemed met. If the business requirement is met, the system extracts the start and end positions of all sub-chains and the inter-chain connection mapping relationships. It connects the output interfaces of each sub-chain to the input interfaces of the next sub-chain according to the sub-chain sequence number, merges the internal work units of the sub-chains to form a continuous sequence, removes the sub-chain splitting markers, and constructs a complete production process path from the start unit to the end unit. If the business requirement is not met, the system saves the position and resource configuration of each work unit in the current execution chain as a status record. It reads the initial arrangement of work units before priority adjustment, selects the unadjusted key business units, reassigns their execution positions, modifies the resource quotas and dependencies of the units, and then performs sub-chain splitting and compatibility assessment again. The regenerated production process path is subjected to an integrity check to verify that all work units are included in the path without omission, to check whether the data transmission interfaces between units are matched, and to confirm that the start and end points of the path conform to the production process specifications. If the check passes, the final production process path is output; otherwise, the process returns to adjust the position of the work units and sets a maximum number of backtracking attempts to prevent infinite loops.

[0111] In one embodiment, the compatibility level value is represented by a numerical range from 0 to 100, where 80 and above is high compatibility, 50 to 80 is medium compatibility, and below 50 is low compatibility. The overlap percentage statistics include the proportion of connection points in three categories: completely overlapping, partially overlapping, and non-overlapping. Business requirements typically set a minimum compatibility level threshold of 60 and an overlap percentage threshold of 70%, and these thresholds can be dynamically adjusted according to different cross-domain scenarios. The comparison process employs a dual-judgment mechanism; only when both the compatibility level value and the overlap percentage simultaneously meet the threshold requirements is the business requirement deemed met.

[0112] Specifically, the subchain merging process needs to handle multiple levels of connectivity. The start and end positions of a subchain are identified by a global index number, and the inter-chain connection mapping records the output data format, data volume, and transmission protocol of the previous subchain, as well as the input interface specification of the next subchain. The connection process proceeds in ascending order of subchain sequence number. For each pair of subchains connected, the system verifies whether the data type of the output interface matches the expected type of the input interface. If the types do not match, a data format conversion module is inserted for adaptation. When merging work units, the global execution sequence number of each unit needs to be updated to ensure sequence number continuity. The process of removing subchain split markers includes deleting special identifiers at the head and tail of the subchain and restoring the direct connection relationship between units.

[0113] It's important to note that the state recording mechanism is crucial for enabling backtracking and adjustments. State records contain three layers of information: the work unit layer records the current execution position, priority value, and resource quota for each unit; the connection layer stores the dependencies and data flow between units; and the global configuration layer stores the overall parameters of the execution chain, such as chain length, total resources, and time constraints. When saving the state, the system assigns a timestamp and version number to the current configuration, facilitating accurate restoration to a specific state during subsequent backtracking. The initial arrangement refers to the original order of work units when the production process is created. This order is based on the natural flow of business logic and has not undergone any optimization or adjustment.

[0114] For example, in the cross-domain production process scenario of intelligent manufacturing, the selection of key business units that have not been adjusted needs to follow specific rules. The system maintains a list of adjusted units, recording the adjustment history of each unit, including the number of adjustments, the magnitude of the adjustments, and the effects of the adjustments. When selecting a new adjustment target, priority is given to units with high business importance but which have not yet been adjusted, followed by units with fewer adjustments. When reassigning execution positions, not only the priority of the unit itself must be considered, but also the impact of the position change on adjacent units must be evaluated. Modification of resource quotas follows the principle of total conservation; increasing the resources of one unit must correspondingly reduce the resources of other units. Modification of dependencies mainly involves loosening tightly coupled dependencies, that is, adjusting dependencies that must be executed serially to those that can be executed in parallel, while ensuring that the correctness of the business logic is not compromised.

[0115] Preferably, the integrity check employs a multi-layered verification mechanism. The first layer of verification is quantity integrity, which involves counting the total number of work units in the path and comparing it with the number of units in the original production process to ensure that no units are missing or duplicated. The second layer of verification is connectivity integrity, which checks whether each unit's input has a corresponding upstream output and its output has a corresponding downstream input, disallowing broken links or isolated nodes. The third layer of verification is semantic integrity, which verifies whether data maintains type consistency and value range rationality when transferred between units.

[0116] In one possible implementation, the data transfer interface matching check includes two phases: static checking and dynamic checking. Static checking is performed immediately after path construction, verifying the compatibility of data types, data structures, and transmission protocols by analyzing the interface definition file. Dynamic checking is performed in a simulated execution environment, verifying the normality of actual data flow by injecting test data. The path start point must be a self-starting work unit, independent of external input or relying solely on system initialization parameters. The path end point must produce final output conforming to production process specifications, including result data, execution reports, and status codes.

[0117] Understandably, setting a maximum number of backtracks is to prevent the system from getting stuck in an unsolvable cycle of adjustments. Typically, the maximum number of backtracks is set to twice the total number of work units; this empirical value covers most adjustable combinations. Each backtracking session records the reason for the backtracking, the adjustments made, and the evaluation results, forming a backtracking log. When the maximum number of backtracks is reached, the system analyzes the backtracking log to identify the root cause of the inability to meet compatibility requirements, which could be insufficient total resources, overly complex dependencies, or significant differences between domains. At this point, the system outputs the relatively better path solution found, along with optimization suggestions, such as increasing resource allocation, simplifying the production process structure, or selecting a more suitable target domain. Furthermore, the final output format of the production process path includes a complete execution plan, detailing the execution sequence, resource allocation, data flow, and exception handling strategies for each work unit, providing comprehensive guidance for the actual execution of the production process.

[0118] S107. By guiding the boundary activation controller to output the adjustment scheme through the final production process path, the execution timing and unit reorganization operations in the scheme are applied to the production process to obtain a stable execution state after crossing the domain boundary.

[0119] The system reads detailed execution plan information from the final production process path, including the work unit sequence, execution time of each unit, and resource requirements. This path data is then transmitted to the boundary activation controller. The controller parses the path data and outputs an adjustment scheme containing timing parameter adjustment values, a unit reorganization instruction sequence, and a resource reallocation ratio. Based on the timing parameter adjustment values ​​in the scheme, the start time of each work unit is reset. The execution order and parallel relationship of the work units are changed according to the unit reorganization instruction sequence. New execution parameters are written through the configuration interface of the production process executor, completing the update of the internal execution logic of the production process. The system monitors the updated production process's running status parameters when crossing the domain boundary, including the percentage of progress of each unit, data transmission rate, and processor utilization. If the change in any parameter exceeds a preset fluctuation threshold, the execution speed is reduced or resource allocation is increased until the change in all parameters returns to within the threshold, confirming that the production process has reached a stable execution state after crossing the domain boundary.

[0120] Specifically, detailed execution plan information is stored in a structured document format, including a unique identifier for each work unit, the estimated start time, duration, CPU core requirements, memory usage, and input / output data volume. The boundary activation controller is a dedicated control device deployed at the domain boundary, possessing protocol conversion, timing adjustment, and resource scheduling capabilities. Path data is transmitted via a dedicated control channel, employing compression encoding to reduce transmission latency.

[0121] In one embodiment, the controller's process of parsing path data includes three stages: data format verification, content extraction, and parameter calculation. Timing parameter adjustment values ​​are calculated based on the ratio of the clock frequency of the new domain to the clock frequency of the original domain, used to synchronize the execution rhythm of different domains. The unit reorganization instruction sequence defines the new connection relationships of work units, including four basic operations: serial-to-parallel conversion, parallel-to-serial conversion, unit merging, and unit splitting. The resource reallocation ratio is dynamically calculated based on the total resources of the new domain and the resource requirements of each unit.

[0122] It should be noted that the configuration interface of the production process executor adopts a standardized parameter writing protocol, supporting batch parameter updates and atomic operations. When writing new execution parameters, the system first pauses the current execution, saves the intermediate state, updates the configuration parameters, and then resumes execution from the save point. This mechanism avoids data loss and execution errors during parameter updates.

[0123] For example, the monitoring of operational status parameters adopts a periodic sampling method with a sampling period of 100 milliseconds. The completion percentage of each unit is calculated by dividing the amount of processed data by the total amount of data; the data transmission rate is obtained by statistically analyzing the number of bytes transmitted per unit time; and the processor utilization rate is read from the operating system's performance counter. The preset fluctuation threshold is usually set to ±20% of the normal value and can be adjusted according to business stability requirements.

[0124] Preferably, a gradual adjustment strategy is adopted when parameter fluctuations exceed the threshold. The execution speed is adjusted by modifying the task scheduling frequency, decreasing it by 10% each time until the parameters stabilize. Resource quota increases are preferentially allocated to the work units with the largest fluctuations, increasing the quota by 15% each time. Parameter changes are continuously monitored during the adjustment process; if all parameters remain within the threshold range for five consecutive sampling periods, a stable execution state is determined.

[0125] In one possible implementation, confirming a stable execution state also includes data integrity verification and execution continuity checks to ensure that the production process can continue to operate normally after crossing domain boundaries without being interrupted or corrupted by environmental differences.

[0126] This invention provides a big data-based cross-domain production process generation and management system, mainly comprising: a boundary demarcation identifier identification module, used to identify the boundary demarcation identifier of the business domain to which the production process belongs, extract the current location information and execution speed of the production process within the domain boundary monitoring area, extract the location coordinates and progress percentage, and evaluate the estimated arrival time point before crossing the boundary line using the location coordinates and progress percentage; a location speed extraction module, used to obtain the task load content and new domain environment requirements information carried by the production process at the estimated arrival time point, evaluate the matching degree between the task load content and the new domain environment requirements, and if the matching degree is lower than the business standard value, initialize the configuration information for new domain environment adaptation; a matching degree evaluation module, used to prioritize the work units in the task load, and evaluate the execution order adjustment scheme by identifying the business importance level in combination with the initialization configuration information, and determine the priority adjustment order; priority sorting; The module is used to select execution order changes and resource reallocation from the priority adjustment and sorting order, reorganize the execution order of work units, generate a reorganized work unit chain, and determine if the chain length exceeds the business standard length, then divide it into sub-chains to obtain an optimized execution chain structure. The execution chain reorganization module is used to obtain the distribution of sub-chains in the execution chain structure, apply continuous execution verification to each sub-chain, and determine the continuous execution compatibility assessment result by comparing the overlap between the connection points between sub-chains and the boundary activation points. The compatibility assessment module is used to merge the sub-chains into a complete production process path if the compatibility assessment result meets the business requirements; otherwise, it backtracks to the timing adjustment and sorting order and readjusts it to obtain the final production process path. The path fusion module is used to guide the boundary activation controller to output an adjustment plan through the final production process path, apply the execution timing and unit reorganization operations in the plan to the production process, and obtain a stable execution state after crossing the domain boundary. Based on the above embodiments of the present invention, and through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the contents of the specification, but must be determined by the scope of the claims.

Claims

1. A generation management method for cross-domain production processes based on big data, characterized in that, include: Identify the boundary delineation marker of the business domain to which the production process belongs, extract the current location information and execution speed of the production process within the domain boundary monitoring area, determine the location coordinates and progress percentage, and calculate the expected arrival time of the production process crossing the domain boundary line using the location coordinates and the progress percentage. Obtain the task load content and target domain environment requirement information carried by the production process at the expected arrival time, calculate the matching degree between the task load content and the target domain environment requirement information, and generate adaptive configuration information based on the matching degree being lower than the business standard value; The work units in the task load are sorted by priority, and the execution order is adjusted in combination with the configuration information and business importance level to generate a priority adjustment and arrangement order; The execution order is changed and resources are reallocated from the priority adjustment and arrangement order. The work units are reorganized to generate a work unit chain. The work unit chain is divided into sub-chains based on the length of the work unit chain exceeding the business standard length, and an optimized execution chain structure is generated. Obtain the distribution of sub-chains in the execution chain structure, perform continuity verification on the sub-chains, and generate compatibility evaluation results by comparing the overlap between connection points and boundary activation points between sub-chains. If the compatibility assessment results meet the business requirements, the sub-chains are integrated to generate a complete production process path; otherwise, the priority adjustment and arrangement order are readjusted to generate a production process path. By adjusting the execution timing and reorganizing the work units through the production process path control boundary, the execution state after crossing the domain boundary is generated.

2. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The process of obtaining the task load content and target domain environment requirements information carried by the production process at the expected arrival time, calculating the matching degree between the task load content and the target domain environment requirements information, and generating adaptive configuration information based on the matching degree being lower than the business standard value includes: Parse the header information of the data packets in the production process to extract a task load content list, which includes task type identifiers, data block format definitions, and execution priority values. Obtain environmental requirement information from the target domain controller, which includes the number of available computing cores and supported communication protocol version numbers. Construct feature vectors for the task load content and environmental requirement information, calculate the similarity score between the feature vectors of the task load content and the environmental requirement information, calculate the matching percentage based on the number of matches between the task type identifier and the types supported by the target domain, and generate a comprehensive matching degree value by combining the similarity score and the matching percentage. If the comprehensive matching degree value is lower than the business standard value, extract the difference elements in the feature vectors of the task load content and the environmental requirement information, generate a protocol version mapping relationship and the number of task granularity divisions, and generate configuration information.

3. The generation management method for cross-domain production processes based on big data according to claim 2, characterized in that, Based on the fact that the comprehensive matching degree value is lower than the business standard value, the difference elements in the feature vector of the task load content and the feature vector of the environment requirement information are extracted to generate a protocol version mapping relationship and the number of task granularity divisions, and to generate configuration information, including: Based on the fact that the overall matching degree value is lower than the business standard value, the difference elements in the feature vector of the task load content and the feature vector of the environment requirement information are extracted to generate a protocol conversion table; the resource quota is adjusted based on the difference elements to generate a resource quota allocation table; the task load is divided based on the task granularity to generate a task segmentation scheme; and the protocol conversion table, the resource quota allocation table and the task segmentation scheme are integrated to generate configuration information.

4. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The process of changing the execution order and reallocating resources from the priority adjustment order, reorganizing the work units to generate a work unit chain, and dividing the work unit chain into sub-chains based on the length exceeding the business standard length, to generate an optimized execution chain structure includes: The process involves: reading the new execution position of each work unit in the priority adjustment sequence; calculating the offset between the new execution position and the original position of each work unit; adjusting the processor time slice and memory space quota for work units whose offset exceeds a preset threshold; generating a resource reallocation record table; reconnecting the data transmission paths between work units based on the new execution positions in the resource reallocation record table; marking work units without dependencies as parallel groups and work units with dependencies as serial sequences; generating a work unit chain containing parallel group identifiers and serial sequence identifiers; calculating the number of work units and the total data transmission volume in the work unit chain; selecting the connection position with the smallest data transmission volume in the work unit chain to divide it into sub-chains based on the fact that the number of work units and the total data transmission volume exceed the business standard length; adding a start unit number and an end unit number to the sub-chains; recording the data flow mapping relationship between the sub-chains; and integrating the sub-chains and the data flow mapping relationship to generate an execution chain structure.

5. The generation management method for cross-domain production processes based on big data according to claim 4, characterized in that, The calculation of the number of work units and the total data transmission volume in the work unit chain, based on the fact that the number of work units and the total data transmission volume exceed the service standard length, involves selecting the connection position with the smallest data transmission volume in the work unit chain and dividing it into sub-chains, including: Calculate the total number of work units and the total amount of data transmission between units in the work unit chain, and identify the connection position with the smallest data transmission volume in the work unit chain; divide the work unit chain based on the connection position to generate multiple sub-chains; assign sequence numbers to the sub-chains and record the data flow mapping relationship between the sub-chains.

6. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The process involves identifying the boundary delineation marker of the business domain to which the production process belongs, extracting the current location information and execution speed of the production process within the domain boundary monitoring area, determining the location coordinates and progress percentage, and calculating the estimated arrival time of the production process crossing the domain boundary line using the location coordinates and the progress percentage. Obtain the geometric coordinates and real-time location coordinates of the domain boundary line of the production process within the current business domain, read the execution speed of the production process, calculate the distance between the real-time location coordinates and the nearest point of the domain boundary line, and generate the estimated arrival time based on the distance and the execution speed; extract the historical trajectory data of the production process within the monitoring area, calculate the speed change rate based on the historical trajectory data, and adjust the estimated arrival time based on the speed change rate.

7. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The step of prioritizing the work units in the task load, adjusting the execution order based on the configuration information and business importance level, and generating a priority adjustment order includes: Extract the attribute records of the work units in the task load, read the business type identifier and expected execution time of the work unit, query the importance coefficient corresponding to the business type identifier based on the business importance mapping table in the configuration information, generate a priority score based on the importance coefficient and the initial priority value, sort the work units based on the priority score, adjust the dependencies, and generate an execution plan table as the priority adjustment order.

8. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The process of obtaining the distribution of sub-chains in the execution chain structure, performing continuity checks on the sub-chains, and generating compatibility evaluation results by comparing the overlap between connection points and boundary activation points between sub-chains includes: Read the position index and connection point information of the sub-chains in the execution chain structure, and construct a distribution mapping table containing the sub-chain sequence number and connection relationship; verify the continuity of the execution sequence number of the working unit within the sub-chain, calculate the distance between the connection point and the boundary activation point, generate an overlap degree mark based on the distance and a preset threshold, calculate the compatibility level based on the overlap degree mark, and generate a compatibility evaluation result.

9. The generation management method for cross-domain production processes based on big data according to claim 1, characterized in that, The step of adjusting the execution timing and reorganizing the work units through the production process path control boundary to generate the execution state after crossing the domain boundary includes: The sequence of work units and execution time points in the production process path are read and transmitted to the boundary activation controller to generate an adjustment scheme containing timing parameter adjustment values ​​and unit reorganization instructions. The execution order of work units is updated based on the adjustment scheme, the running status parameters of the production process when crossing the domain boundary are monitored, the resource quota is adjusted based on the running status parameters, and the execution status after crossing the domain boundary is generated.

10. A generation management system for cross-domain production processes based on big data, characterized in that, The system includes: a boundary demarcation identifier identification module, used to identify the boundary demarcation identifier of the business domain to which the production process belongs, extract the current location information and execution speed of the production process in the domain boundary monitoring area, determine the location coordinates and progress percentage, and calculate the expected arrival time of the production process crossing the domain boundary line through the location coordinates and the progress percentage; The location and speed extraction module is used to obtain the task load content and target domain environment requirement information carried by the production process at the expected arrival time, calculate the matching degree between the task load content and the target domain environment requirement information, and generate adaptive configuration information based on the matching degree being lower than the business standard value. The matching degree evaluation module is used to sort the work units in the task load by priority, adjust the execution order in combination with the configuration information and business importance level, and generate a priority adjustment and arrangement order; The priority sorting module is used to change the execution order and reallocate resources from the priority adjustment and arrangement order, reorganize the work units to generate a work unit chain, divide the work unit chain into sub-chains based on the length of the work unit chain exceeding the business standard length, and generate an optimized execution chain structure. The execution chain reorganization module is used to obtain the distribution of sub-chains in the execution chain structure, perform continuity verification on the sub-chains, and generate compatibility evaluation results by comparing the overlap between connection points and boundary activation points between sub-chains. The compatibility assessment module is used to integrate the sub-chains to generate a complete production process path based on the compatibility assessment results, if the business requirements are met; otherwise, it backtracks to the priority adjustment and rearrangement order to re-adjust and generate a production process path. The path fusion module is used to adjust the execution timing and reorganize the work units through the production process path control boundary to generate the execution state after crossing the domain boundary.

Images