A digital employee role management and orchestration method and system supporting dynamic plugging

Abstract

The application discloses a kind of digital staff role management and arrangement method and system supporting dynamic plugging, comprising: obtaining digital staff list, collecting the business execution ability of each staff and encapsulating as one-to-one business operator set;Response logic arrangement instruction, generate the virtual team execution graph for representing business process topology structure and data flow relationship;When receiving update release request, generate target virtual team graph, and control it with original virtual team execution graph parallel execution;Target business traffic is diverted to target virtual team graph, and after the inventory business instance in original execution graph is naturally exhausted, execute resource recovery.Compared with prior art, the application can significantly improve the reusability of capability components, the agility of business response and the continuity of version update, while overcoming the technical shortcomings of component reuse, slow business response, service interruption and abnormal termination of inventory tasks in traditional solutions.

Description

Technical Field

[0001] This invention relates to the fields of computer data processing and software engineering technology, and more specifically, to a method and system for managing and orchestrating digital employee roles that supports dynamic plug-and-play. Background Technology

[0002] With the large-scale development of enterprise digital transformation and intelligent automation technologies, digital employees, as the core productive force for refactoring business processes, have become crucial to enterprise operational efficiency due to the flexibility of their role management and the agility of their business orchestration. Modern complex business scenarios require digital employee systems to possess comprehensive capabilities such as 24 / 7 uninterrupted service, rapid iteration of business logic, and seamless upgrades in changing environments. Traditional RPA deployment methods based on monolithic scripts or rigidly fixed processes, limited by high code coupling, long maintenance windows, and strong dependencies on the runtime environment, can no longer meet the efficient management needs of large-scale, high-concurrency business flows. Microservice-based digital employees, supporting dynamic plug-and-play functionality, are gradually becoming an important carrier for business process automation due to their component-based reusability advantages. The smoothness of their version updates, the continuity of their operational status, and the compatibility between new and old logic directly determine the stability and operational security of the business system.

[0003] In existing technologies, although some solutions attempt to update business logic through downtime maintenance or direct overwrite deployment, existing technologies still have problems such as the inability to reuse capability components and slow business response. Furthermore, during system updates, technical defects such as service interruption, abnormal termination of existing tasks, and untimely recycling of abandoned resources occur.

[0004] Therefore, how to provide a digital employee role management and orchestration method that supports dynamic plug-and-play, which can significantly improve the reusability of capability components, the agility of business response, and the continuity of version updates, and overcome the technical shortcomings of traditional solutions such as non-reusable components, slow business response, service interruption, and abnormal termination of existing tasks, has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play functionality. This method can significantly improve the continuity of digital employee version updates, the completion rate of existing business tasks, and the efficiency of computing resource recycling, overcoming the technical shortcomings of traditional solutions, such as the need for system shutdown for updates, forced interruption of existing tasks, and the inability to automatically release abandoned resources.

[0006] The first technical solution provided by this invention is as follows: This invention provides a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play functionality, comprising the following steps: S1 Obtaining a list of digital employees, collecting the business execution capabilities of each digital employee based on the list, encapsulating each business execution capability, and obtaining a set of business operators, wherein each business operator in the set corresponds one-to-one with each digital employee; S2 Responding to the logical orchestration instructions of the set of business operators, generating a virtual team execution diagram, which is used to represent the topological structure and data flow relationships of the business process; S3 When receiving an update release request based on the virtual team execution diagram, generating a target virtual team diagram and controlling the target virtual team diagram and the virtual team execution diagram to execute in parallel; S4 Directing the target business traffic to the target virtual team diagram, and performing resource reclamation after the existing business instances in the virtual team execution diagram are naturally exhausted.

[0007] Furthermore, in a preferred embodiment of the present invention, the step of encapsulating each of the said business execution capabilities includes: Configure a standardized interface for each of the aforementioned business execution capabilities; the standardized interface is used to standardize the format of data interaction. Configure a state transition rule for each of the aforementioned business execution capabilities. The state transition rule is used to limit the legal state change path of the business execution capability from startup to termination. The program code of the business execution capability and the dependent files required for running are packaged together, and the standardized interface and the state transition rules are integrated to obtain an executable image file. The executable image file represents the complete environment required for the business execution capability to run independently. Configure version number identifier and service level indicators for the executable image file, and generate business operators.

[0008] Further, in a preferred embodiment of the present invention, the step of generating a virtual work group execution diagram in response to the logical orchestration instructions of the set of business operators includes: The logical orchestration instructions are parsed to identify several selected business operators within the logical orchestration instructions, and the data flow connection relationships between the multiple business operators are determined. Based on the standardized interface of each business operator, verify one by one whether the output data format of the upstream business operator and the input data format of the downstream business operator along the data flow connection meet the consistency matching requirements; When the data flow connection relationship meets the consistency matching requirement, a virtual team execution graph is constructed according to the logical orchestration instruction. The virtual team execution graph includes all the selected business operators and the data flow connection relationship.

[0009] Furthermore, in a preferred embodiment of the present invention, the state hot migration mechanism based on topology difference analysis during runtime is specifically implemented as follows: By comparing the node composition and connection relationship between the target virtual work group diagram and the virtual work group execution diagram, the business operators that have undergone logical changes are identified and marked as difference nodes; Based on the data flow connection relationship in the virtual team execution diagram, determine the upstream path set and downstream path set for each of the differential nodes; Detect the current running node of each of the existing service instances, and filter the existing service instances based on the upstream path set; If the currently running node is in the upstream path set, then the existing service instance is marked as a migrateable instance; The runtime context data of the migrateable instance is migrated across graphs to the target virtual work group graph, and the target virtual work group graph is controlled to take over the subsequent business processes of the migrateable instance.

[0010] Further, in a preferred embodiment of the present invention, the step of migrating the runtime context data of the migrateable instance across graphs to the target virtual class group graph includes: Freeze the running state of the portable instance in the virtual team execution graph, and extract the runtime context data containing the current variable values ​​and execution stack; Based on the standardized interface of the node corresponding to the target virtual class group diagram, the field structure of the runtime context data is mapped to generate adapted migration context data; The migration context data is injected into the corresponding node position in the target virtual work group diagram to activate the takeover of the subsequent business process.

[0011] Furthermore, in a preferred embodiment of the present invention, the step of directing the target business traffic to the target virtual work group diagram includes: Intercept newly incoming business request traffic, parse the header metadata of the business request traffic, and extract traffic coloring tags, which are used to distinguish business attribution characteristics; According to the preset gray-scale traffic splitting strategy, the service request traffic that matches the specific traffic coloring label is routed to the target virtual team diagram, and the remaining unmatched service request traffic continues to be routed to the virtual team execution diagram until all traffic is switched over.

[0012] Furthermore, in a preferred embodiment of the present invention, the version tracing and rollback mechanism specifically includes: In response to each update release request for the virtual work group execution diagram, a digital hash fingerprint containing the changes and a timestamp record are generated; Write the digital hash fingerprint and the timestamp record into an immutable operation audit log to construct a version traceability chain; When an abnormal operation of the target virtual work group diagram is detected, the operation audit log is used to locate the previous version of the virtual work group execution diagram, and a version rollback operation is performed to restore business operation.

[0013] Furthermore, in a preferred embodiment of the present invention, the step of performing resource recovery includes: Establish a reference counting mechanism based on the virtual work group execution graph, and track the current concurrent business volume in real time according to the reference counting mechanism; When the existing service instances are naturally exhausted, an atomic decrement operation is performed on the reference counting mechanism; When the reference count is detected to be zero, it is determined that the virtual team execution graph has entered a zero-load state, and the resource release and service offline process based on the underlying container environment is triggered to perform resource reclamation.

[0014] The present invention provides a second technical solution as follows: This invention also provides a digital employee role management and orchestration system that supports dynamic plug-in, including the digital employee role management and orchestration method that supports dynamic plug-in as described in any one of the above, comprising: The operator encapsulation generation module obtains a list of digital employees, collects the business execution capabilities of each digital employee based on the list of digital employees, encapsulates each business execution capability, and obtains a set of business operators, wherein each business operator in the set of business operators corresponds one-to-one with each digital employee. The virtual work group scheduling module, in response to the logical scheduling instructions of the business operator set, generates a virtual work group execution diagram, which is used to represent the topological structure and data flow relationship of the business process; The dual-graph parallel control module, upon receiving an update release request based on the virtual work group execution graph, generates a target virtual work group graph and controls the target virtual work group graph and the virtual work group execution graph to execute in parallel. The traffic diversion and resource recycling module diverts target business traffic to the target virtual work group diagram, and performs resource recycling after the existing business instances in the virtual work group execution diagram are naturally exhausted.

[0015] Furthermore, in a preferred embodiment of the present invention, the operator encapsulation generation module includes: A standard interface configuration unit configures a standardized interface for each of the aforementioned business execution capabilities. The standardized interface is used to standardize the format of data interaction. The state transition configuration unit configures state transition rules for each of the aforementioned service execution capabilities. The state transition rules are used to limit the legal state change paths of the service execution capability from startup to termination. The image packaging and generation unit packages the program code of the business execution capability and the dependency files required for running, and integrates the standardized interface and the state transition rules to obtain an executable image file. The executable image file represents the complete environment required for the business execution capability to run independently. The version indicator configuration unit configures the version number identifier and service level indicator for the executable image file and generates business operators.

[0016] This invention provides a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play functionality. This method can significantly improve the continuity of digital employee version updates, the completion rate of existing business, and the efficiency of computing resource recycling. It overcomes the technical shortcomings of traditional solutions, such as the need for system shutdown for updates, forced interruption of existing tasks, and the inability to automatically release abandoned resources. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play includes: S1 obtaining a list of digital employees, collecting the business execution capabilities of each digital employee based on the list, encapsulating each business execution capability, and obtaining a set of business operators, wherein each business operator in the set corresponds one-to-one with each digital employee; S2 responding to the logical orchestration instructions of the set of business operators, generating a virtual team execution diagram, which is used to represent the topological structure and data flow relationship of the business process; S3 when receiving an update release request based on the virtual team execution diagram, generating a target virtual team diagram and controlling the target virtual team diagram and the virtual team execution diagram to execute in parallel; S4 diverting the target business traffic to the target virtual team diagram, and performing resource reclamation after the existing business instances in the virtual team execution diagram are naturally exhausted. This involves standardizing and encapsulating the business execution capabilities of digital employees, transforming them into a set of independent and reusable business operators, thus decoupling functions from specific scenarios. Subsequently, it dynamically generates virtual team execution diagrams representing the process topology in response to logic orchestration instructions. This construction model based on operator reuse and graphical orchestration allows business logic adjustments to be made only by reorganizing operator connections without refactoring the code, thereby achieving agile response to business needs and cross-scenario reuse of core capabilities. When the system receives an update release request, it does not directly overwrite the current operating environment, but instead generates an independent target virtual team diagram and controls the relationship between this target virtual team diagram and the original virtual team execution diagram. Parallel execution within the runtime space, this topology-isolated parallel operation mode ensures that available service instances are always available during version transitions, thus eliminating service interruptions caused by deployment downtime or direct overwrite. The system accurately directs newly arriving target business traffic to the new version's target virtual task graph while maintaining the original virtual task execution graph until any remaining unfinished existing business instances complete their execution according to predetermined logic, thereby avoiding abnormal service termination due to forced deletion. Once existing instances are exhausted, the system immediately triggers a resource reclamation process, resolving the issue of long-term computational resource occupation after the old version is abandoned. Compared to existing technologies, this invention significantly improves the reusability of capability components, the agility of business response, and the continuity of version updates, while overcoming the technical shortcomings of traditional solutions such as non-reusable components, slow business response, service interruptions, and abnormal termination of existing tasks. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the steps of a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play, as provided in an embodiment of the present invention. Figure 2 A flowchart illustrating the business execution capability encapsulation steps provided in this embodiment of the invention; Figure 3 A logical framework diagram of the virtual team execution diagram provided in the embodiments of the present invention; Figure 4 A logical framework diagram of a digital employee role management and orchestration system supporting dynamic plug-and-play provided in an embodiment of the present invention; Figure 5 Experimental comparison diagram of the graph construction and verification mechanism provided in the embodiments of the present invention; Figure 6 This is a comparison diagram showing the implementation of the thermal migration mechanism provided in the embodiments of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0020] It should be noted that when a component is referred to as being "fixed to" or "set on" another component, it can be directly on or indirectly set on the other component; when a component is referred to as being "connected to" another component, it can be directly connected to or indirectly connected to the other component.

[0021] It should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "first", "second", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0022] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" or "several" means two or more, unless otherwise explicitly specified.

[0023] It should be noted that the structures, proportions, sizes, etc., shown in the accompanying drawings of this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0024] like Figures 1 to 6 As shown in the figure, the present invention provides a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play, which can significantly improve the reusability of capability components, the agility of business response, and the continuity of version updates. At the same time, it overcomes the technical shortcomings of traditional solutions, such as components that cannot be reused, slow business response, service interruption, and abnormal termination of existing tasks.

[0025] A method for managing and orchestrating digital employee roles that supports dynamic plug-and-play includes the following steps: S1 Obtain a list of digital employees, collect the business execution capabilities of each digital employee based on the list, encapsulate each business execution capability, and obtain a set of business operators, wherein each business operator in the set of business operators corresponds one-to-one with each digital employee; S2 Respond to the logical orchestration instructions of the set of business operators, generate a virtual team execution diagram, which is used to represent the topology and data flow relationship of the business process; S3 When an update release request based on the virtual team execution diagram is received, generate a target virtual team diagram and control the target virtual team diagram and the virtual team execution diagram to execute in parallel; S4 Direct the target business traffic to the target virtual team diagram, and perform resource reclamation after the existing business instances in the virtual team execution diagram are naturally exhausted. This involves standardizing and encapsulating the business execution capabilities of digital employees, transforming them into a set of independent and reusable business operators, thus decoupling functions from specific scenarios. Subsequently, it dynamically generates virtual team execution diagrams representing the process topology in response to logic orchestration instructions. This construction model based on operator reuse and graphical orchestration allows business logic adjustments to be made only by reorganizing operator connections without refactoring the code, thereby achieving agile response to business needs and cross-scenario reuse of core capabilities. When the system receives an update release request, it does not directly overwrite the current operating environment, but instead generates an independent target virtual team diagram and controls the relationship between this target virtual team diagram and the original virtual team execution diagram. Parallel execution within the runtime space, this topology-isolated parallel operation mode ensures that available service instances are always available during version transitions, thus eliminating service interruptions caused by deployment downtime or direct overwrite. The system accurately directs newly arriving target business traffic to the new version's target virtual task graph while maintaining the original virtual task execution graph until any remaining unfinished existing business instances complete their execution according to predetermined logic, thereby avoiding abnormal service termination due to forced deletion. Once existing instances are exhausted, the system immediately triggers a resource reclamation process, resolving the issue of long-term computational resource occupation after the old version is abandoned. Compared to existing technologies, this invention significantly improves the reusability of capability components, the agility of business response, and the continuity of version updates, while overcoming the technical shortcomings of traditional solutions such as non-reusable components, slow business response, service interruptions, and abnormal termination of existing tasks.

[0026] The following describes in detail the steps and flow of a method and system for managing and orchestrating digital employee roles that supports dynamic plug-and-play, using specific embodiments.

[0027] Specifically, in a specific embodiment of the present invention, the steps encapsulate each business execution capability, including: S11 configuring a standardized interface for each business execution capability, the standardized interface being used to standardize the format of data interaction; S12 configuring state transition rules for each business execution capability, the state transition rules being used to limit the legal state change path of the business execution capability from startup to termination; S13 packaging the program code of the business execution capability and the dependent files required for running, and integrating the standardized interface and state transition rules to obtain an executable image file, the executable image file representing the complete environment required for the independent operation of the business execution capability; S14 configuring version number identifier and service level indicators for the executable image file, and generating business operators.

[0028] Among them, such as Figure 2 As shown, in a specific embodiment of the present invention, firstly, a set of general interface definition language specifications needs to be established, such as using JSON Schema or Protobuf format as the data contract. For each business execution capability collected from the digital employee list, the data type of its input parameters, the validation logic of required fields, and the structured template of the output results are defined to construct a standardized interface and ensure that scripts written in different programming languages ​​can follow a unified data interaction format. Next, based on the finite state machine theory, state transition rules are designed for each business execution capability. Specifically, discrete state nodes such as "initialization," "running," "suspended waiting," "retry due to exception," and "execution termination" can be defined in the configuration document, and the legal transition conditions between each state are specified in the form of an adjacency matrix or directed graph. For example, it is specified that the transition to the "retry due to exception" state can only occur when the state is "running" and a specific error code occurs, thereby locking out illegal paths at the logical level. Subsequently... The process begins by initiating a containerized build process. Based on the source code language of the business execution capability, a corresponding base image is selected. A build script is written, and the program source code, a list of third-party dependency libraries, the previously generated standardized interface definition file, and the state transition rule configuration file are copied to the image file system. The build tool then performs compilation and layered packaging operations to generate a read-only container image file containing a complete runtime environment, ensuring consistent performance across any host machine. Finally, tag information is injected into the metadata layer of the generated executable image file, a version number identifier conforming to semantic versioning specifications is configured, and service level indicators, such as maximum response time threshold, concurrency limits, and memory usage limits, are set. This completes the generation of business operators, transforming discrete scripts into standardized, independently manageable assets.

[0029] Specifically, in a specific embodiment of the present invention, the step of generating a virtual work group execution graph in response to the logical orchestration instructions of the set of business operators includes: parsing the logical orchestration instructions, identifying a plurality of the selected business operators in the logical orchestration instructions, and determining the data flow connection relationship between the plurality of business operators; according to the standardized interface of each business operator, verifying one by one whether the output data format of the upstream business operator and the input data format of the downstream business operator along the data flow connection relationship meet the consistency matching requirement; when the data flow connection relationship meets the consistency matching requirement, constructing a virtual work group execution graph according to the logical orchestration instructions, wherein the virtual work group execution graph includes all the selected business operators and the data flow connection relationship.

[0030] Among them, such as Figure 3 As shown, in an embodiment of the present invention, a data structure based on a directed acyclic graph is established to parse and carry logical orchestration instructions. These instructions are typically in the form of a configuration file in JSON or YAML format, which defines in detail the list of node IDs involved in the business process and the connection relationships between nodes. The system first loads the configuration file, parses out all selected business operator identifiers using a traversal algorithm, and maps abstract identifiers to specific operator instance objects based on the previously generated business operator library index. Simultaneously, it parses out the edge set list defining the data flow direction, constructing a preliminary adjacency list structure. Subsequently, the system enters the core interface consistency verification stage. It traverses each connection relationship in the edge set list, locates the upstream business operator as the starting point and the downstream business operator as the ending point, and retrieves the standardized output interface definition configured by the upstream operator. The standardized input interface definition of downstream operators is based on a preset field mapping protocol. It compares the names, data types, and constraints of data fields layer by layer. For example, if the "amount" field output by the upstream operator is a string type, while the "amount" field input by the downstream operator requires a double-precision floating-point type, the system will immediately trigger a verification failure exception and block the generation of the graph to prevent runtime crashes caused by type mismatch. The system instantiates a directed acyclic graph object according to the logical orchestration instructions only when the interface definitions on all connection paths meet the strict consistency matching requirements. It encapsulates business operators as graph nodes, instantiates data flow relationships as graph edges with metadata, and performs full-graph loop detection to ensure the finality of the process. Finally, it generates a virtual team execution graph that can be recognized by the scheduling engine. This execution graph can accurately represent the topology and data flow path of the business process.

[0031] To verify the reliability of this graph construction and verification mechanism in complex business scenarios, such as Figure 5As shown, a simulation test environment with a high-concurrency scheduling engine was built. Four stages in the supply chain finance scenario—order capture, credit assessment, credit limit calculation, and loan approval—were selected as test cases. A complex mesh topology with fifty nodes was constructed. One hundred sets of logical orchestration instructions were prepared for the experiment, including eighty sets of instructions conforming to the interface specifications and twenty sets of artificially inserted abnormal instructions with interface type conflicts (such as mismatched date formats or mixed use of integer and floating-point types). The experimental group used the verification construction mechanism described in this application, while the control group used a traditional construction method that only verifies the existence of nodes without verifying data types. Test results showed that the experimental group successfully identified and intercepted all twenty sets of abnormal instructions during the graph construction phase. The experimental group achieved a 100% accuracy rate in intercepting abnormal commands. In contrast, while the control group successfully generated all execution graphs, all twenty task groups were interrupted due to "TypeError" or "DataParseError" during the subsequent simulation run, with an average fault recovery time of up to 30 minutes. Furthermore, the data flow latency of the compliant execution graphs generated by the experimental group was reduced by 15% compared to the control group. This is because the strict verification during the pre-compilation stage reduced the overhead of dynamic type checking at runtime. These experimental data demonstrate that the consistency verification mechanism based on standardized interfaces can effectively eliminate data compatibility risks in logical orchestration and significantly improve the operational stability and execution efficiency of the virtual team execution graph.

[0032] Specifically, in a specific embodiment of the present invention, it further includes executing a state hot migration mechanism based on topology difference analysis during runtime, specifically: comparing the node composition and connection relationship between the target virtual work group diagram and the virtual work group execution diagram, identifying business operators that have undergone logical changes and marking them as difference nodes; determining the upstream path set and downstream path set for each difference node according to the data flow connection relationship in the virtual work group execution diagram; detecting the current running node of each existing business instance, and filtering existing business instances based on the upstream path set; when the current running node is in the upstream path set, marking the existing business instance as a migrateable instance; migrating the runtime context data of the migrateable instance across the graph to the target virtual work group diagram, and controlling the target virtual work group diagram to take over the subsequent business processes of the migrateable instance.

[0033] In an embodiment of the present invention, the system serializes the target virtual work group graph and the currently running virtual work group execution graph into adjacency matrix structures containing node attribute hash values ​​and edge relationships, respectively. Specifically, the process of establishing the adjacency matrix structure is as follows: First, determine the total number N of all business operator nodes in the graph, and construct an N×N two-dimensional array as the matrix skeleton, where the row index and column index correspond to the starting node and ending node in the graph, respectively. In order to simultaneously represent connection relationships and node logical changes in a single matrix, each element in the matrix no longer simply stores binary connectivity identifiers, but is designed as a structure object containing a composite data structure. This object contains three key fields: connectivity identifier, source node attribute hash, and edge metadata. During initialization, the system traverses each node in the virtual work group execution graph, using the SHA-256 algorithm to hash the node's configuration parameters (including image version number, environment variables, resource constraints, etc.) to generate a unique node attribute fingerprint. This fingerprint, along with the edge connectivity status and data flow format constraints, is then filled into the corresponding structure element of the matrix. By comparing the hash fingerprints of two adjacency matrices row by row, nodes with changed attributes or edges with added or deleted connections are quickly located, and these changed logical units are marked as difference nodes, thus defining the... The minimum boundary of the impact of the change; then, based on the directed acyclic graph structure of the current virtual team execution graph, starting from each difference node, traversing all parent node paths pointing to that difference node, storing all traversed node IDs into the upstream path set, and simultaneously performing a forward traversal to obtain the downstream path set, thereby constructing a logical boundary line based on the flow direction; here, in order to balance the state consistency guarantee under long-link dependencies and system migration performance, the upstream path set is strictly defined as a reverse topological closure with the difference node as the endpoint, that is, including the direct upstream node and all indirect upstream nodes, to prevent state breakage caused by the lack of logical dependencies of indirect upstream nodes; query The system identifies the node ID currently residing in each existing business instance and performs a set inclusion determination between this node ID and the aforementioned upstream path set. If the current node ID of an instance exists in the upstream path set, it means that the instance has not yet executed the logic area that has changed, and the system marks it as a migrateable instance. Finally, the system performs an atomic state freeze operation on the migrateable instance, extracting runtime data including the current variable stack, execution pointer, and global context. Specifically, this includes serialized copies of business variable key-value pairs, session tokens of external interfaces, and database transaction IDs, but does not include the physical network connection handle, which is automatically reconnected by the target node based on the transaction ID.Subsequently, a low-latency transmission channel is established using the gRPC high-performance remote procedure call protocol based on the HTTP / 2 standard. This channel is then mapped and injected into the corresponding node with the same name in the target virtual work group diagram. The scheduler of the target diagram is then unfrozen and activated, allowing the instance to seamlessly switch to the new version of the business logic without being aware of it. Instances located at different nodes or downstream path sets remain in the original diagram and continue running until completion.

[0034] To verify the effectiveness of this hot migration mechanism in ensuring business continuity, such as Figure 6 As shown, a simulation experimental environment including a microservice orchestration engine was built. The "order fulfillment process" of an e-commerce platform was selected as the test scenario. This process includes four sequential nodes: "order creation," "inventory deduction," "logistics order placement," and "shipping notification." A version change scenario was designed, upgrading the internal logic of the "inventory deduction" node and generating a new target graph. During the test, 5000 order business instances at different stages of the process were generated. The experimental group used the topology difference analysis hot migration strategy described in this application, while the control group used the traditional "shutdown and full migration" strategy. Experimental data showed that the experimental group successfully identified 1200 instances in the "order creation" stage as migrateable. The system successfully migrated objects and hot-migrated them to the new graph with a 100% success rate. 3800 instances in the "inventory deduction" stage and later were correctly identified as non-migratable and remained in the original graph, completing the migration naturally. The average business interruption time for the entire process was only 150 milliseconds. In contrast, the control group, due to the forced migration of instances after the change node, experienced a mismatch between the context state and the new logic, resulting in 340 database transaction exceptions and state rollbacks, with a business interruption time of up to 5 seconds. These experimental results demonstrate that the hot migration mechanism based on topology difference analysis can accurately distinguish the migration eligibility of existing businesses, significantly improving business continuity during version updates while ensuring data consistency.

[0035] Specifically, in this embodiment of the invention, the step of migrating the runtime context data of the migrateable instance across graphs to the target virtual work group graph includes: freezing the running state of the migrateable instance in the virtual work group execution graph, extracting runtime context data containing the current variable values ​​and execution stack; mapping the field structure of the runtime context data according to the standardized interface of the corresponding node in the target virtual work group graph to generate adapted migration context data; and injecting the migration context data into the corresponding node position in the target virtual work group graph to activate the takeover of subsequent business processes.

[0036] In a specific embodiment of the present invention, when the monitoring system determines that a certain existing business instance meets the migration conditions, the scheduler sends an application-level suspension instruction to the container running the instance. This instruction triggers the current worker thread to enter a spinlock state or a sleep state, thereby freezing the movement of the execution pointer and the changes of variables. Subsequently, through reflection or memory introspection, all local variables, global object references, and the current position of the program counter in the current execution stack are traversed and serialized into an intermediate general data format, such as JSON or Protocol. The system generates a source context snapshot in Buffers format, containing a triplet of "variable name-variable value-data type". Next, it enters the crucial field structure mapping stage. The system retrieves the interface definition file of the corresponding node in the target virtual work group diagram and instantiates a field mapping engine. This engine first establishes the correspondence between source and target fields based on fuzzy matching of field names or a preset alias mapping table. For fields with changed names, it performs redirection assignment (e.g., changing "userId" to "account_id"). For fields with changed data types (e.g., converting integer "1" to boolean "true"), it executes a type casting function and fills in default values ​​for newly added required fields in the target node, thus generating adapted migration context data. Finally, through the state injection RESTful or RPC interface exposed by the target node, the migration context data is sent to the target container. The target container's deserializer restores the data to the memory stack, resets the instruction pointer to the corresponding logical breakpoint, sends a resumption execution signal, and completes the seamless takeover of the business process.

[0037] To verify the accuracy and stability of the aforementioned cross-graph migration mechanism in handling complex data structure changes, a full-link stress testing environment including heterogeneous microservices was built. The "e-commerce promotion calculation" scenario was selected as the test case. In this scenario, the "price calculation" node used "single-precision floating-point numbers" to store prices in the old version, while in the new version, the target graph used the "BigDecimal" type to improve precision, and some field names were changed. The experiment generated 10,000 business instances in the intermediate calculation state. The experimental group used the field structure mapping mechanism described in this application for migration, while the control group used direct memory binary copying. The experimental results showed that the experimental group achieved a migration success rate of 99.98%, with only two failures due to network timeouts. After migration, all "BigDecimal" type variables were correctly initialized without precision loss. The average execution time of the business logic in the target graph until completion only increased by 20 milliseconds (due to serialization and deserialization overhead). In contrast, due to the inconsistent memory layout between the old and new versions, 100% of the instances in the control group experienced "Segmentation" upon recovery. The "Fault" or "ClassCastException" error caused a complete business interruption. This experimental data strongly demonstrates that the context migration method based on standardized interface mapping can effectively shield the differences in the underlying code implementation and ensure a smooth transition of business instances and data integrity in scenarios of logical changes.

[0038] Specifically, in a specific embodiment of the present invention, the step of diverting target business traffic to the target virtual work group diagram includes: intercepting newly incoming business request traffic, parsing the header metadata of the business request traffic, extracting traffic coloring tags, which are used to distinguish business affiliation characteristics; and, according to a preset gray-scale diversion strategy, routing business request traffic that matches a specific traffic coloring tag to the target virtual work group diagram, and continuing to route the remaining unmatched business request traffic to the virtual work group execution diagram until all traffic has been switched.

[0039] In this embodiment of the invention, a high-performance gateway component with Layer 7 load balancing capabilities, such as a cloud-native gateway built on Nginx or Envoy, needs to be deployed at the system's traffic entry layer first. Custom traffic interception and parsing logic is then embedded in the gateway's Lua script layer or filter chain. Before this, the gray-scale traffic splitting strategy needs to be preset and constructed. Specifically, the preset process involves the system first analyzing user attribute distribution and business call peak and valley characteristics, combining this with a risk level assessment model for current business changes, and automatically constructing a routing rule chain containing multi-level priorities through a strategy generation engine. This rule chain is then serialized into a JSON or YAML policy configuration file and pushed to a distributed configuration center such as Etcd or Consul. The policy configuration file strictly defines three progressive stages of traffic filtering logic: the first stage is a whitelist verification strategy, configured to only allow HTTP traffic. Internal test traffic carrying a specific debug token in the header is directly routed to the target graph; the second level is a feature-based grayscale strategy, configured to identify the region or member attribute tags of user IDs based on Bloom filters, and import traffic matching specific profiles into the target graph; the third level is a dynamic stepping strategy, configured to bind to the real-time health monitoring indicators of the target graph, such as an error rate of less than 0.05%, and set an automatically increasing traffic weight step size according to time windows: Specifically, this strategy clearly defines the execution timing logic: in terms of startup timing, the first round of traffic splitting can only be started after the target virtual class group graph is online and running for 60 seconds without any abnormal crashes; during the splitting process, a judgment is performed every 5 minutes, and if the current error rate continues to be lower than the preset threshold (0.05%), the traffic splitting will be initiated. The traffic weight will be increased by 20% sequentially. For the full switchover standard, when the traffic weight reaches 100% and the system runs continuously under full load for 10 minutes without any new fault alarms, the full switchover is considered complete, and the default routing rules are updated. When business request traffic reaches the gateway, the gateway temporarily suspends the request, reads the protocol header information, extracts coloring tags such as "X-Tenant-Tag" or calculates the hash value based on the user ID, and then pulls the aforementioned policy file from the configuration center for real-time matching. If the request matches a routing rule at a certain level, the gateway immediately modifies the upstream context of the request, accurately distributing it to the container instance corresponding to the target virtual class group diagram, while unmatched requests are forwarded along their original paths until the full traffic switchover is completed smoothly.

[0040] It should be noted that the specific construction method of the canary traffic splitting strategy is as follows: The system first analyzes the distribution of user attributes and the peak and valley characteristics of business calls, and combines this with the risk level assessment model of the current business changes. Then, through the strategy generation engine, it automatically constructs a routing rule chain containing multiple levels of priority, and serializes this rule chain into a strategy configuration file in JSON or YAML format, which is then pushed to a distributed configuration center such as Etcd or Consul. This strategy configuration file strictly defines the three progressive stages of traffic filtering logic: the first level is a whitelist verification strategy, configured to only allow HTTP traffic. Internal test traffic carrying a specific debug token in the header is directly routed to the target graph; the second level is a feature-based grayscale strategy, configured to identify the region or member attribute tags of user IDs based on Bloom filters, and import traffic that matches a specific profile into the target graph; the third level is a dynamic step-by-step strategy, configured to bind to the real-time health monitoring indicators of the target graph, such as an error rate of less than 0.05%, and set an automatically increasing traffic weight step size according to time windows: Specifically, this strategy clearly defines the execution timing logic: in terms of startup timing, the first round of traffic splitting can only be started after the target virtual class group graph is online and running for 60 seconds without any abnormal crashes; during the splitting process, a judgment is performed every 5 minutes, and if the current error rate continues to be lower than the preset threshold (0.05%), the traffic weight is increased by 20% sequentially; in terms of the full switch standard, when the traffic weight reaches 100% and the system runs continuously under full load for 10 minutes without any new fault alarms, the full switch is judged to be completed, and the default routing rules are updated; Specifically, in the embodiments of the present invention, a version traceability and rollback mechanism is also included, which specifically involves: in response to each update release request for the virtual work group execution diagram, generating a digital hash fingerprint containing the changed content and a timestamp record; writing the digital hash fingerprint and the timestamp record into an immutable operation audit log to construct a version traceability chain; when an abnormal operation of the target virtual work group diagram is detected, locating the previous version of the virtual work group execution diagram based on the operation audit log, and performing a version rollback operation to restore business operation.

[0041] In a specific embodiment of the present invention, when the system receives an update release request for the virtual team execution diagram, the configuration management engine first standardizes the newly submitted logical orchestration instruction set and operator configuration parameters. For example, it rearranges the JSON configuration file according to the lexicographical order of the key names to eliminate the interference of format differences on hash calculation. Then, it uses SHA-256 or a higher-level cryptographic hash algorithm to calculate the standardized configuration data and generate a unique digital hash fingerprint. At the same time, it obtains the current high-precision network timestamp and combines the timestamp, digital hash fingerprint, initiator ID of the change operation, and hash fingerprint (Parent) of the previous version of the execution diagram. The hash is encapsulated into a standardized "version block" object. By including the fingerprint of the previous version, a unidirectionally growing cryptographic chain, i.e., a version traceability chain, is logically constructed. The system writes this version block to an operation audit log storage medium that supports Write-once-Read-Many (WORM) functionality, or to the write-ahead log of a distributed consistent key-value storage system based on the Raft protocol (such as Etcd). This ensures that any tampering with the historical record will cause the chain verification to fail, thereby guaranteeing the immutability of the audit data. During the runtime phase, the system deploys real-time monitoring probes to continuously collect key performance indicators of the target virtual work group diagram and sets composite circuit breaker thresholds, such as when HTTP... If the 5xx error rate exceeds 3% within 1 minute or the average response latency exceeds 2 seconds, an anomaly alarm will be triggered immediately. Once an operational anomaly is detected, the rollback controller will quickly trace back to the version block of the previous version based on the chain structure in the audit log, extract the historical configuration snapshot stored therein, send instructions to the orchestration engine to reload the old version execution graph, and simultaneously adjust the traffic routing rules to switch business traffic back to the stable old version environment, thereby achieving rapid recovery of business operations.

[0042] Specifically, in a specific embodiment of the present invention, the step of performing resource reclamation includes: establishing a reference counting mechanism based on the virtual team execution graph, and tracking the current concurrent business volume in real time according to the reference counting mechanism; when the existing business instances are naturally exhausted, performing an atomic decrement operation on the reference counting mechanism; when the reference count is detected to be zero, determining that the virtual team execution graph has entered a zero-load state, and triggering a resource release and service offline process based on the underlying container environment to perform resource reclamation.

[0043] In a specific embodiment of this invention, a highly available, atomically supported distributed state storage component, such as a Redis cluster or Etcd, is first introduced to maintain an active task counter for each virtual team execution graph instance. The system assigns a unique session identifier to each running execution graph and creates a corresponding key-value pair in Redis, with the key name being the session identifier and an initial value of 0. Whenever new business request traffic enters the execution graph through the gateway route, the gateway layer interceptor performs an INCR atomic increment operation on the key-value pair via a Lua script, ensuring the accuracy of the count under high concurrency scenarios. When an existing business instance completes the flow of all logical nodes within the execution graph and is about to exit the system, a post-hook deployed in the egress gateway or Sidecar proxy is triggered. The function sends an atomic DECR decrement command to Redis. Simultaneously, a dedicated resource lifecycle daemon is deployed, using a publish / subscribe pattern or keyspace notification mechanism to monitor the counter's value changes in real time. Once the daemon detects that the reference count of an older version of the execution graph has changed to 0, it immediately initiates zero-load judgment logic. To prevent false judgments due to instantaneous zeroing caused by network jitter, a very short silent observation window can be set, such as 500 milliseconds. If the count is still 0 after the window ends, the execution graph is officially determined to have entered a zero-load state. Subsequently, the daemon calls the interface of the underlying container orchestration platform to send a smooth termination signal to the Pod set carrying the execution graph. After the containers complete log writing to disk and connection closure, it executes the DeleteDeployment or Scale to 0 operation, thereby physically releasing CPU, memory, and network handle resources, completing the service's shutdown and recycling.

[0044] This invention also provides a digital employee role management and orchestration system that supports dynamic plug-and-play, such as... Figure 4 As shown, it includes: an operator encapsulation generation module, which obtains a list of digital employees, collects the business execution capabilities of each digital employee based on the list, encapsulates each business execution capability, and obtains a set of business operators, wherein each business operator in the set of business operators corresponds one-to-one with each digital employee; a virtual team orchestration module, which generates a virtual team execution graph in response to the logical orchestration instructions of the set of business operators, wherein the virtual team execution graph is used to represent the topological structure and data flow relationship of the business process; a dual-graph parallel control module, which generates a target virtual team graph when it receives an update release request based on the virtual team execution graph, and controls the target virtual team graph and the virtual team execution graph to execute in parallel; and a traffic diversion and resource recovery module, which diverts the target business traffic to the target virtual team graph, and performs resource recovery after the existing business instances in the virtual team execution graph are naturally exhausted.

[0045] Specifically, in a specific embodiment of the present invention, the operator encapsulation generation module includes: a standard interface configuration unit, which configures a standardized interface for each business execution capability, the standardized interface being used to standardize the format of data interaction; a state transition configuration unit, which configures state transition rules for each business execution capability, the state transition rules being used to limit the legal state change path of the business execution capability from startup to termination; an image packaging generation unit, which packages the program code of the business execution capability and the dependent files required for running, and integrates the standardized interface and the state transition rules to obtain an executable image file, the executable image file representing the complete environment required for the independent operation of the business execution capability; and a version indicator configuration unit, which configures a version number identifier and service level indicator for the executable image file to generate a business operator.

[0046] As described above, the present invention relates to a method for managing and orchestrating digital employee roles that supports dynamic plug-and-play, which can significantly improve the continuity of digital employee version updates, the completion rate of existing business, and the efficiency of computing resource recycling, overcoming the technical shortcomings of traditional solutions such as the need for system shutdown for updates, forced interruption of existing tasks, and the inability to automatically release abandoned resources. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play includes: S1 obtaining a list of digital employees, collecting the business execution capabilities of each digital employee based on the list, encapsulating each business execution capability, and obtaining a set of business operators, wherein each business operator in the set corresponds one-to-one with each digital employee; S2 responding to the logical orchestration instructions of the set of business operators, generating a virtual team execution diagram, which is used to represent the topological structure and data flow relationship of the business process; S3 when receiving an update release request based on the virtual team execution diagram, generating a target virtual team diagram and controlling the target virtual team diagram and the virtual team execution diagram to execute in parallel; S4 diverting the target business traffic to the target virtual team diagram, and performing resource reclamation after the existing business instances in the virtual team execution diagram are naturally exhausted. This involves standardizing and encapsulating the business execution capabilities of digital employees, transforming them into a set of independent and reusable business operators, thus decoupling functions from specific scenarios. Subsequently, it dynamically generates virtual team execution diagrams representing the process topology in response to logic orchestration instructions. This construction model based on operator reuse and graphical orchestration allows business logic adjustments to be made only by reorganizing operator connections without refactoring the code, thereby achieving agile response to business needs and cross-scenario reuse of core capabilities. When the system receives an update release request, it does not directly overwrite the current operating environment, but instead generates an independent target virtual team diagram and controls the relationship between this target virtual team diagram and the original virtual team execution diagram. Parallel execution within the runtime space, this topology-isolated parallel operation mode ensures that available service instances are always available during version transitions, thus eliminating service interruptions caused by deployment downtime or direct overwrite. The system accurately directs newly arriving target business traffic to the new version's target virtual task graph while maintaining the original virtual task execution graph until any remaining unfinished existing business instances complete their execution according to predetermined logic, thereby avoiding abnormal service termination due to forced deletion. Once existing instances are exhausted, the system immediately triggers a resource reclamation process, resolving the issue of long-term computational resource occupation after the old version is abandoned. Compared to existing technologies, this invention significantly improves the reusability of capability components, the agility of business response, and the continuity of version updates, while overcoming the technical shortcomings of traditional solutions such as non-reusable components, slow business response, service interruptions, and abnormal termination of existing tasks.

[0047] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for managing and orchestrating digital employee roles that supports dynamic plug-and-play, characterized in that, include: S1 obtains a list of digital employees, collects the business execution capabilities of each digital employee based on the list of digital employees, encapsulates each business execution capability, and obtains a set of business operators, wherein each business operator in the set of business operators corresponds one-to-one with each digital employee. S2 responds to the logical orchestration instructions of the business operator set and generates a virtual team execution diagram, which is used to represent the topological structure and data flow relationship of the business process; When S3 receives an update release request based on the virtual team execution diagram, it generates a target virtual team diagram and controls the target virtual team diagram and the virtual team execution diagram to be executed in parallel. S4 redirects the target business traffic to the target virtual work group diagram, and performs resource reclamation after the existing business instances in the virtual work group execution diagram are naturally exhausted.

2. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 1, characterized in that, The steps encapsulate each of the aforementioned business execution capabilities, including: Configure a standardized interface for each of the aforementioned business execution capabilities; the standardized interface is used to standardize the format of data interaction. Configure a state transition rule for each of the aforementioned business execution capabilities. The state transition rule is used to limit the legal state change path of the business execution capability from startup to termination. The program code of the business execution capability and the dependent files required for running are packaged together, and the standardized interface and the state transition rules are integrated to obtain an executable image file. The executable image file represents the complete environment required for the business execution capability to run independently. Configure version number identifier and service level indicators for the executable image file, and generate business operators.

3. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 2, characterized in that, The step, responding to the logical orchestration instructions of the set of business operators, generates a virtual work group execution diagram, including: The logical orchestration instructions are parsed to identify several selected business operators within the logical orchestration instructions, and the data flow connection relationships between the multiple business operators are determined. Based on the standardized interface of each business operator, verify one by one whether the output data format of the upstream business operator and the input data format of the downstream business operator along the data flow connection meet the consistency matching requirements; When the data flow connection relationship meets the consistency matching requirement, a virtual team execution graph is constructed according to the logical orchestration instruction. The virtual team execution graph includes all the selected business operators and the data flow connection relationship.

4. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 3, characterized in that, It also includes a state hot migration mechanism for performing topology difference analysis at runtime, specifically: By comparing the node composition and connection relationship between the target virtual work group diagram and the virtual work group execution diagram, the business operators that have undergone logical changes are identified and marked as difference nodes; Based on the data flow connection relationship in the virtual team execution diagram, determine the upstream path set and downstream path set for each of the differential nodes; Detect the current running node of each of the existing service instances, and filter the existing service instances based on the upstream path set; If the currently running node is in the upstream path set, then the existing service instance is marked as a migrateable instance; The runtime context data of the migrateable instance is migrated across graphs to the target virtual work group graph, and the target virtual work group graph is controlled to take over the subsequent business processes of the migrateable instance.

5. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 4, characterized in that, The step of migrating the runtime context data of the migrateable instance across graphs to the target virtual class group graph includes: Freeze the running state of the portable instance in the virtual team execution graph, and extract the runtime context data containing the current variable values ​​and execution stack; Based on the standardized interface of the node corresponding to the target virtual class group diagram, the field structure of the runtime context data is mapped to generate adapted migration context data; The migration context data is injected into the corresponding node position in the target virtual work group diagram to activate the takeover of the subsequent business process.

6. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 1, characterized in that, The steps to redirect target business traffic to the target virtual work group diagram include: Intercept newly incoming business request traffic, parse the header metadata of the business request traffic, and extract traffic coloring tags, which are used to distinguish business attribution characteristics; According to the preset gray-scale traffic splitting strategy, the service request traffic that matches the specific traffic coloring label is routed to the target virtual team diagram, and the remaining unmatched service request traffic continues to be routed to the virtual team execution diagram until all traffic is switched over.

7. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 1, characterized in that, It also includes version tracking and rollback mechanisms, specifically: In response to each update release request for the virtual work group execution diagram, a digital hash fingerprint containing the changes and a timestamp record are generated; Write the digital hash fingerprint and the timestamp record into an immutable operation audit log to construct a version traceability chain; When an abnormal operation of the target virtual work group diagram is detected, the operation audit log is used to locate the previous version of the virtual work group execution diagram, and a version rollback operation is performed to restore business operation.

8. The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play according to claim 1, characterized in that, The steps involved in resource recovery include: Establish a reference counting mechanism based on the virtual work group execution graph, and track the current concurrent business volume in real time according to the reference counting mechanism; When the existing service instances are naturally exhausted, an atomic decrement operation is performed on the reference counting mechanism; When the reference count is detected to be zero, it is determined that the virtual team execution graph has entered a zero-load state, and the resource release and service offline process based on the underlying container environment is triggered to perform resource reclamation.

9. A digital employee role management and orchestration system supporting dynamic plug-and-play, characterized in that, The method for managing and orchestrating digital employee roles that supports dynamic plug-and-play, as described in any one of claims 1 to 8, includes: The operator encapsulation generation module obtains a list of digital employees, collects the business execution capabilities of each digital employee based on the list of digital employees, encapsulates each business execution capability, and obtains a set of business operators, wherein each business operator in the set of business operators corresponds one-to-one with each digital employee. The virtual work group scheduling module, in response to the logical scheduling instructions of the business operator set, generates a virtual work group execution diagram, which is used to represent the topological structure and data flow relationship of the business process; The dual-graph parallel control module, upon receiving an update release request based on the virtual work group execution graph, generates a target virtual work group graph and controls the target virtual work group graph and the virtual work group execution graph to execute in parallel. The traffic diversion and resource recycling module diverts target business traffic to the target virtual work group diagram, and performs resource recycling after the existing business instances in the virtual work group execution diagram are naturally exhausted.

10. The digital employee role management and orchestration system supporting dynamic plug-and-play as described in claim 9, characterized in that, The operator encapsulation generation module includes: A standard interface configuration unit configures a standardized interface for each of the aforementioned business execution capabilities. The standardized interface is used to standardize the format of data interaction. The state transition configuration unit configures state transition rules for each of the aforementioned service execution capabilities. The state transition rules are used to limit the legal state change paths of the service execution capability from startup to termination. The image packaging and generation unit packages the program code of the business execution capability and the dependency files required for running, and integrates the standardized interface and the state transition rules to obtain an executable image file. The executable image file represents the complete environment required for the business execution capability to run independently. The version indicator configuration unit configures the version number identifier and service level indicator for the executable image file and generates business operators.

Images