Task scheduling method and device, electronic equipment and storage medium

Abstract

This invention discloses a task scheduling method, apparatus, electronic device, and storage medium. The method includes: decomposing and parsing a first task to obtain at least one second task and first information; matching at least one first intelligent construct based on the first information; and processing at least one second task based on the at least one first intelligent construct to obtain a first result. This method, by decomposing and parsing a first task to obtain at least one second task and first information, matching intelligent constructs based on the first information, and processing the corresponding second task based on the obtained intelligent constructs, enables collaborative processing of multiple intelligent constructs. Furthermore, each intelligent construct can process different tasks without interfering with each other's information, thereby improving the accuracy of the reasoning results and the stability of the system.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a task scheduling method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the development of generative artificial intelligence technology, AI agent systems built on AI models are gradually evolving from a single-agent execution mode to a multi-agent collaborative execution mode. In this mode, multiple agents can each undertake different responsibilities such as information retrieval, task planning, tool invocation, and result verification, thereby improving the automation capabilities of complex tasks.

[0003] In existing technologies, multi-agent systems typically rely on shared context or weakly isolated state management methods. This means that multiple agents share the same context space or distinguish context data only through simple labeling when performing different tasks. This approach works in scenarios with small task scale or low concurrency, but it has gradually revealed many problems in scenarios involving multi-task parallelism, high-frequency calls, and complex collaboration.

[0004] On the one hand, when different tasks or agents write intermediate states, inference results, or tool call information into a shared context, it can easily lead to the overlapping or confusion of context information, thereby affecting the accuracy of subsequent inference and even causing erroneous decisions. On the other hand, existing technologies generally lack effective management mechanisms for the lifecycle of context data. Context data is retained for a long time after the task is completed, consuming system resources and reducing the overall stability and scalability of the system.

[0005] In addition, some existing systems attempt to solve the above problems by assigning an independent context to each agent. However, this approach often completely severs the information sharing capabilities between different agents, making it difficult to meet the needs of multiple agents to reuse intermediate information and perform collaborative reasoning when performing tasks together, resulting in a decrease in system efficiency. Summary of the Invention

[0006] This invention provides a task scheduling method, apparatus, electronic device, and storage medium to solve the problems of low data sharing capability and chaotic context information sharing in intelligent constructs.

[0007] According to one aspect of the present invention, a task scheduling method is provided, comprising: The first task is broken down and analyzed to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task; the second task is a sub-task obtained by breaking down the first task according to its completion steps. At least one first intelligent construct is matched based on the first information; the first intelligent construct is a processing unit capable of processing the first task, and the first intelligent construct is configured with a first space; the first space is used to record data information generated by the intelligent construct when processing the first task; The at least one first intelligent construct processes the at least one second task to obtain a first result; the first result is the completion result of the first task.

[0008] According to another aspect of the present invention, a task scheduling apparatus is provided, comprising: The first task processing module is used to decompose and parse the first task to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task; the second task is a sub-task obtained by decomposing the first task according to its completion steps. The first intelligent construct determination module is used to match at least one first intelligent construct based on the first information; the first intelligent construct is a processing unit capable of processing the first task, and the first intelligent construct is configured with a first space; the first space is used to record data information generated by the intelligent construct when processing the first task. The task processing module is used to process the at least one second task according to the at least one first intelligent construct to obtain a first result; the first result is the completion result of the first task.

[0009] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the task scheduling method according to any embodiment of the present invention.

[0010] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the task scheduling method of any embodiment of the present invention.

[0011] The technical solution of this invention involves decomposing and parsing a first task to obtain at least one second task and first information; matching at least one first intelligent construct based on the first information; and processing at least one second task based on the at least one first intelligent construct to obtain a first result. This method, by decomposing and parsing a first task to obtain at least one second task and first information, matching intelligent constructs based on the first information, and processing the corresponding second task based on the obtained intelligent constructs, enables collaborative processing of multiple intelligent constructs. Furthermore, each intelligent construct can ensure that information does not interfere with each other when processing different tasks, thereby improving the accuracy of the reasoning results and the stability of the system.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0014] Figure 1 A flowchart of a task scheduling method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a task scheduling device provided in an embodiment of the present invention; Figure 3 A schematic diagram of the structure of an electronic device for implementing the task scheduling method of this embodiment of the invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] Figure 1 This is a flowchart illustrating a task scheduling method provided in an embodiment of the present invention. This embodiment is applicable to situations where tasks are scheduled in isolation. The method can be executed by a task scheduling device, which can be implemented in hardware and / or software. This task scheduling device can be configured in any electronic device with network communication capabilities. Figure 1 As shown, the method includes: S110. The first task is decomposed and analyzed to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task; the second task is a sub-task obtained by decomposing the first task according to its completion steps.

[0018] The first task is to obtain natural language information that can represent the content of the task from the display interface.

[0019] The first piece of information includes: the task information of the second task and the sorting information of the second task.

[0020] The task information includes at least: the task type, task priority, number of intelligent constructs required, and the capability characteristics of the intelligent constructs for the second task.

[0021] The sorting information refers to the processing order of the second task.

[0022] Specifically, semantic parsing and keyword extraction are performed on the first task to obtain third information. Based on the obtained third information, the first task is broken down to obtain at least one second task. The at least one second task is parsed to obtain task information for each second task. The obtained task information is matched with the corresponding second tasks and sorted. The sorted information is then structured and transformed to obtain the first information.

[0023] Among them, the capability characteristics of an intelligent construct are the algorithmic framework that the intelligent construct is equipped with to complete the task.

[0024] Among them, the AI ​​Agent is an algorithmic structural framework capable of handling different tasks.

[0025] Furthermore, the above steps are implemented through a task receiving and parsing module, which is configured within the task scheduling system. The task scheduling system also includes: an intelligent constructor scheduling module, a context management module, a context lifecycle control module, and an execution and result aggregation module.

[0026] The intelligent construct scheduling module is used to select one or more qualified AI Agents from the intelligent construct pool based on the first information, and assign corresponding execution tasks to them.

[0027] Among them, the intelligent agent construction pool is a pre-established database used to store different types of intelligent agents.

[0028] The context management module is used to create and maintain a context space for each intelligent construct and its tasks. It controls access permissions for different context spaces to prevent unauthorized access to information.

[0029] The context lifecycle control module manages the creation, updating, freezing, and release of context spaces. Management includes, but is not limited to: context expiration control; context state version management; and context recycling or archiving upon task completion. Through lifecycle control, the context lifecycle control module prevents contexts from growing indefinitely or occupying system resources for extended periods.

[0030] The execution and result aggregation module is used to coordinate the execution of specific tasks by intelligent constructs, and to aggregate, integrate and output the execution results of each intelligent construct after the task is completed.

[0031] S120. Match at least one first intelligent construct according to the first information; the first intelligent construct is a processing unit capable of processing the first task, and the first intelligent construct is configured with a first space; the first space is used to record data information generated when the intelligent construct processes the first task.

[0032] The context space includes at least two components: a private context space and a shared context space. The private context space stores the intermediate states and inference information of the intelligent constructs during the current task execution. The shared context space allows reusable information or intermediate results to be shared among multiple intelligent constructs.

[0033] Specifically, based on the capability characteristics of the intelligent construct and the number of intelligent constructs in the first information, a corresponding number of intelligent constructs are matched from the intelligent construct pool to obtain at least one intelligent construct.

[0034] Furthermore, after obtaining the smart construct, a context space is created for each smart construct according to the context management module.

[0035] The context management module is used to create and maintain a context space for each intelligent construct and its tasks. The context management module controls access permissions for different context spaces to prevent unauthorized access to information.

[0036] The above steps demonstrate that setting the context space can ensure multi-agent collaboration capabilities while achieving effective isolation and controllable sharing of the context, thereby improving the system's stability, scalability, and inference consistency.

[0037] S130. Process at least one second task according to at least one first intelligent construct to obtain a first result; the first result is the completion result of the first task.

[0038] Specifically, at least one second task is matched with at least one first intelligent construct, and the matched intelligent construct is designated as the second intelligent construct. The corresponding context space is determined based on the second intelligent construct, resulting in a first space. The execution information required by the second intelligent construct for task processing is obtained from the first space, resulting in fourth information. The task information of the second task corresponding to the second intelligent construct is obtained, resulting in fifth information. The fourth information and the fifth information are compared. If the fourth information contains the fifth information, the second intelligent construct is considered capable of processing the second task, and the second task is input into the second intelligent construct for processing. If the fourth information does not contain the fifth information, the second intelligent construct is matched with a third intelligent construct based on the second result.

[0039] Furthermore, after the second intelligent construct matches the third intelligent construct based on the second result, the process includes: the third intelligent construct processing the second task, and after processing, sending the intermediate state and reasoning information generated during the processing, as well as the processing result, to the second intelligent construct.

[0040] The above steps, a multi-agent collaborative execution mode, involve multiple intelligent constructs processing different tasks in parallel, invoking different tools, and sharing some intermediate information. This solves the problem in most existing systems that use simple global contexts or weak isolation mechanisms, which easily lead to mutual interference between information from different tasks, thus affecting the accuracy of inference results and the stability of the system.

[0041] For example, the task scheduling method in this case can be adapted to the following application scenarios: intelligent question answering and generative search systems, business process automation systems for enterprises, intelligent construct systems that need to call multiple external tools or services, AI agent service platforms with multiple concurrent users, intelligent construct systems with audit and backtracking requirements, and intelligent construct systems with multi-model or multi-version model collaboration.

[0042] Intelligent question answering and generative search systems often employ multiple AI agents to handle different stages of processing, such as query parsing, information retrieval, result verification, and content generation. During these processes, different intelligent constructs need to share some intermediate information, which can lead to contextual confusion. The method presented in this paper effectively avoids contextual confusion between different tasks by allocating independent private context spaces to different intelligent constructs and introducing a shared context for storing reusable information. This improves the efficiency and accuracy of collaborative reasoning among multiple intelligent constructs.

[0043] Enterprise-oriented business process automation systems require complex operations such as process decomposition, task planning, tool invocation, and result verification. The method presented in this case completes complex business processes through parallel or serial collaboration of multiple intelligent constructs, ensuring context isolation between different business process instances and preventing the erroneous reuse of business states. Furthermore, a context lifecycle management mechanism allows for timely reclamation of context resources after the process ends, improving overall system efficiency.

[0044] In intelligent construct systems that require the invocation of multiple external tools or services, the invocation of different types of tools within the same task can generate numerous intermediate results, leading to errors in the intermediate result invocation. The method presented in this case controls context access permissions, ensuring that tool invocation results are only written to the private context of the corresponding intelligent construct or an authorized shared context. This prevents the tool invocation results from being misused by other tasks or intelligent constructs, thereby improving the security and reliability of system execution.

[0045] In multi-user concurrent AI agent service platforms, the system typically needs to process task requests from different users simultaneously. Each user task may require multiple processing modules to complete, potentially leading to cross-referencing or data leakage of user information. The method presented in this case binds the context to specific task instances, achieving context isolation between different user tasks. This prevents cross-referencing or leakage of user data in concurrent scenarios, meeting the requirements for data isolation and security in multi-user concurrent environments.

[0046] Among them, intelligent construct systems with auditing and backtracking requirements need to be able to backtrack the intermediate processing and inference information of tasks. The method in this case can manage the context during task execution in an orderly manner through context lifecycle management and state management mechanisms, so that the execution path and intermediate results of each task are traceable, which facilitates subsequent auditing, debugging or analysis.

[0047] Intelligent construct systems that utilize multiple models or multiple versions of models often require tasks to be performed based on different models or different versions of the same model. Intermediate information generated during inference by different models needs to be managed in isolation. The method presented in this case can achieve context isolation between different models or model versions by binding the context to intelligent construct instances. It also supports information synchronization through shared context when necessary, improving the overall flexibility and controllability of the system.

[0048] Optionally, the first task is broken down and analyzed to obtain at least one second task and the first information, including steps A1-A3: Step A1: Perform semantic parsing and data cleaning on the first task to obtain the second information.

[0049] Specifically, the first task is broken down into multiple word units. Each word unit is then cleaned to remove conjunctions and modifiers, resulting in a selection of words. Semantic analysis is then performed on these selection words to determine the relationships between their meanings, thus obtaining the second piece of information.

[0050] For example, suppose the task is to retrieve all operation information from day A to day B from the database and categorize it. The filter keywords would be: database, day A, day B, all, operation information, and category. The second set of information would be: day A to day B is the time interval; the database represents the location from which the operation information is retrieved; all represents the range of operation information retrieved; and the category represents the processing actions performed on the retrieved operation information.

[0051] Step A2: Extract keywords from the second information to obtain the third information; keywords are words that can represent the content of the first task.

[0052] Specifically, words related to the task content are extracted from the acquired second information, and these words are used as the third information.

[0053] For example, suppose the second information is the time interval from day A to day B; the database is the location where the operation information is obtained; all is the range of the obtained operation information; and the category is the processing action for the obtained operation information. Then the third information obtained is: location: database, time interval: day A to day B, range of obtained information: all operation information, and processing action: category.

[0054] Step A3: Based on the third information, the first task is decomposed and structured to obtain at least one second task and the first information.

[0055] Specifically, based on the task content contained in the third information, the steps required to complete the first task are determined. The first task is then broken down according to the obtained steps to obtain at least one second task. The at least one second task is then parsed, and the obtained parsed information and the corresponding second task are structurally transformed to obtain the first information.

[0056] For example, assuming the third piece of information is: location: database, time interval: day A - day B, acquisition range: operation information, and processing action: classification, then the processing steps of the first task are to first acquire the operation information and then classify the operation information. For step 1, an intelligent constructor for acquiring information and at least one intelligent constructor for segmenting operation information are needed; for step 2, at least one intelligent constructor for classification is needed, and each intelligent constructor can acquire at least one type of information.

[0057] Optionally, at least one second task is processed according to at least one first intelligent construct to obtain a first result, including steps B1-B2: Step B1: Match at least one second task with at least one first intelligent construct and process them to obtain at least one second result.

[0058] Specifically, based on the acquired first information, the second task is matched with the corresponding first intelligent construct. The matched first intelligent construct processes the corresponding second task, and the processing result of each first intelligent construct is obtained to obtain at least one second result.

[0059] Furthermore, after obtaining at least one second result, each first intelligent construct reports its obtained second result to the execution and result aggregation module. A combination of "real-time reporting + batch aggregation" is used to aggregate results. That is, after the task is completed, each first intelligent construct automatically reports its own execution result to the execution and result aggregation module; if multiple first intelligent constructs execute in parallel or subtasks are completed in stages, a batch aggregation mechanism is used to periodically collect stage results to ensure aggregation efficiency and avoid data backlog.

[0060] Furthermore, during the reporting process, a second result verification is performed simultaneously. The list of first intelligent constructs is compared with the number of first intelligent constructs whose results have been collected to check for any unreported, delayed, or missing data. A reminder is triggered for unreported first intelligent constructs, and requirements are issued to complete missing data results, ensuring the completeness and validity of the aggregated results.

[0061] The first smart construct list is a list of first smart constructs generated based on at least one first smart construct.

[0062] Step B2: Integrate at least one second result to obtain a first result.

[0063] Specifically, at least one second result is obtained and subjected to data verification and standardization. The processed second result is then cleaned and classified. Finally, the results are fused to obtain the first result.

[0064] Furthermore, data verification involves performing integrity verification on at least one second result, which means identifying all first intelligent structures participating in this task and ensuring that the execution results of each intelligent structure are collected without omitting the result data of any subtask or execution node, including results of normal completion, results of partial completion, and abnormal feedback of execution failure.

[0065] Furthermore, standardization processing involves standardizing the second result after data verification to prevent difficulties in subsequent integration due to inconsistent formats.

[0066] Furthermore, data cleaning involves verifying each of the at least one secondary result obtained, removing problematic data, classifying and labeling abnormal results, clarifying the causes of abnormalities, and providing a basis for subsequent review.

[0067] Problem data includes at least two types: redundant data and erroneous data. Redundant data refers to data that is repeatedly reported, such as the same type of data reported by multiple intelligent constructs; erroneous data refers to data generated during processing, such as format errors, numerical deviations, and ambiguous descriptions.

[0068] Abnormal results include at least: task execution failure and task partial completion.

[0069] Among them, the cause of the anomaly is the reason for the problematic data and abnormal results. For example, it can be the insufficient capabilities of the intelligent construct, interference from the external environment, or parameter errors.

[0070] Furthermore, the data is categorized by associating the cleaned second result with the corresponding sub-tasks and task nodes based on the first information, clarifying the task division for each result; at the same time, it is classified according to the result type.

[0071] The result type can be at least: data type, operation type, or feedback type.

[0072] Furthermore, result fusion involves combining the categorized results to form a complete task execution chain. For example, the results of the data acquisition intelligent construct are fused with the results of the data processing intelligent construct to form a complete data link of "acquisition-processing-analysis"; missing auxiliary information and unclear execution details are supplemented by combining the perceived data and historical data of the intelligent construct to obtain the first information.

[0073] The above steps ensure that the integrated result set is complete and coherent, and can comprehensively reflect the entire task execution process.

[0074] Furthermore, after obtaining the initial information, the integrated results are quantitatively evaluated by combining the initial task objective with the preset evaluation criteria, clarifying core indicators such as overall task completion rate, execution accuracy rate, and time consumption compliance rate; the performance of each intelligent construct is compared to analyze its strengths and weaknesses, providing data support for subsequent intelligent construct capability optimization and task allocation adjustment.

[0075] Optionally, at least one second task is matched with at least one first intelligent construct and processed, including steps C1-C3: Step C1: Match at least one second task with at least one first intelligent construct to obtain at least one second intelligent construct; the second intelligent construct corresponds one-to-one with the second task.

[0076] Specifically, at least one second task is matched with at least one first intelligent construct, and the matched intelligent construct is used as the second intelligent construct.

[0077] Step C2: Determine the first space based on the second intelligent construct, and determine the fourth information from the first space; the fourth information is used to characterize the execution information required by the second intelligent construct to perform task processing.

[0078] The fourth information includes at least: information about the tasks that the second intelligent construct can handle and the auxiliary information needed in the task handling process.

[0079] Specifically, the context space corresponding to the second intelligent construct is determined to obtain the first space. The execution information required by the second intelligent construct for task processing is obtained from the first space to obtain the fourth information.

[0080] Step C3: The second intelligent construct processes the corresponding second task based on the fourth information.

[0081] Specifically, the task information of the second task corresponding to the second intelligent construct is obtained as the fifth information. The fourth information is compared with the fifth information. If the fourth information contains the fifth information, it is considered that the second intelligent construct can process the second task, and the second task is input into the second intelligent construct for processing. If the fourth information does not contain the fifth information, the second intelligent construct matches the third intelligent construct according to its processing result of the second task, i.e., the second result.

[0082] Optionally, the second intelligent construct processes the corresponding second task based on the fourth information, including steps D1-D4: Step D1: Determine the fifth piece of information; the fifth piece of information is the task information corresponding to the second task.

[0083] The fifth piece of information includes: task type, task processing auxiliary information, and the capability characteristics of the intelligent construct.

[0084] Among them, task processing auxiliary information refers to information that needs to be provided by the outside world during task processing.

[0085] Specifically, obtain the task processing information corresponding to the second task to obtain the fifth information.

[0086] Step D2: Compare the fourth information with the fifth information to obtain the second result; the second result is used to characterize whether the fourth information contains the fifth information.

[0087] Specifically, first, the task type in the fifth piece of information is compared with the task type recorded in the fourth piece of information. If the fourth piece of information contains the task type in the fifth piece of information, then the task processing auxiliary information in the fifth piece of information is compared with the auxiliary information in the fourth piece of information. If the fourth piece of information contains the task processing auxiliary information in the fifth piece of information, then the second result is "included information". If the fourth piece of information does not contain the task type in the fifth piece of information, or if the fifth piece of information contains auxiliary information that is not included in the fourth piece of information, then the second result is "not included information", indicating that the second intelligent construct cannot process the second task.

[0088] The information not included includes: the task type of the second task and missing auxiliary information. The absence of a task type indicates that the algorithmic framework of the second intelligent construct is incompatible with the second task and cannot process it. Missing auxiliary information indicates that the second intelligent construct lacks the necessary auxiliary information to process the second task, and therefore cannot process it.

[0089] Step D3: If the second result contains information, then the second task is processed through the second intelligent construct.

[0090] Specifically, if the second result contains information, it indicates that the second intelligent construct can handle the second task, and the second task is input into the second intelligent construct for processing.

[0091] Step D4: If the second result does not contain information, then match the third smart construct.

[0092] Specifically, if the second result is that it does not contain information, it indicates that the second intelligent construct cannot handle the second task, and the second intelligent construct matches the third intelligent construct according to the second result.

[0093] Further, the matching of the second intelligent construct with the third intelligent construct based on the second result is as follows: If the second result indicates that the second intelligent construct does not contain the task type of the second task, then it indicates that the second intelligent construct processes the second task. In this case, the second task is further split into sub-tasks, and the third intelligent construct is matched based on the parsing results after the splitting. If the second result indicates that the second intelligent construct does not contain auxiliary information, then it indicates that the second intelligent construct cannot process the second task due to the lack of auxiliary information. In this case, the intelligent construct corresponding to the one that does not contain auxiliary information is obtained from the auxiliary information management library and used as the third intelligent construct. The second intelligent construct packages its processing result for the second task, the second task, and the query result in the auxiliary information management library and sends them to the third intelligent construct.

[0094] The auxiliary information management library is used to record auxiliary information and its corresponding intelligent constructs.

[0095] Optionally, at least one second task is processed according to at least one first intelligent construct, including steps E1-E3: Step E1: Determine the status information of the first node; the status information of the first node is the status information of the processing node of each second task that is set in advance.

[0096] The first node can be at least: a completion node, a checkpoint, or a breakpoint.

[0097] The completion node is the node after the second task is completed.

[0098] Checkpoints are pre-defined nodes where the second task needs to undergo status checks during execution. Checkpoints are used to prevent the second task from entering an infinite loop or becoming unresponsive.

[0099] Breakpoints are nodes where the execution of the second task requires the processing results of other second intelligent constructs to continue.

[0100] When processing the corresponding second task through different second intelligent constructs, the processing progress of the task is detected in real time, and it is determined whether the processing of the task has reached the first node based on the processing progress. If the first node has been reached, the current status information of the first node is obtained; if the first node has not been reached, the processing progress detection continues.

[0101] Step E2: Determine the first action based on the first node status information and the preset status information.

[0102] Among them, the preset status information is the pre-defined status information of the first node under different tasks.

[0103] Specifically, the preset state information is filtered based on the obtained first node state information. That is, the state information of the corresponding task at the first node is matched from the preset state information based on the first node state information. The matched state information is compared with the first node state information. If the comparison result is that the first node state information meets the conditions, the first action is matched from the preset action library.

[0104] The preset action library is a pre-built database of actions for processing the context space of intelligent constructs. The preset action library contains at least the following actions: update, freeze, recycle, and clean up.

[0105] The update involves replacing, adding to, or modifying the context space corresponding to the smart construct based on information obtained by the smart construct from other smart constructs.

[0106] Freezing refers to freezing the context space of a smart construct, but retaining the data recorded within it.

[0107] Here, recycling involves deleting some information from the context space of the intelligent construct.

[0108] Cleaning involves clearing all information in the context space of the smart construct.

[0109] Step E3: Process the second space according to the first action; the second space is the context space of the first intelligent construct corresponding to the first node.

[0110] Specifically, the first intelligent construct corresponding to the first node is determined, and the corresponding context space is determined based on the obtained first intelligent construct, thus obtaining the second space. The second space is then processed according to the first action.

[0111] The technical solution of this embodiment involves decomposing and parsing a first task to obtain at least one second task and first information; matching at least one first intelligent construct based on the first information; and processing at least one second task based on the at least one first intelligent construct to obtain a first result. This method, by decomposing and parsing a first task to obtain at least one second task and first information, matching intelligent constructs based on the first information, and processing the corresponding second task based on the obtained intelligent constructs, enables collaborative processing of multiple intelligent constructs. Furthermore, each intelligent construct can ensure that information does not interfere with each other when processing different tasks, thereby improving the accuracy of the reasoning results and the stability of the system.

[0112] Figure 2 This is a schematic diagram of a task scheduling device provided in an embodiment of the present invention. This embodiment is applicable to situations requiring isolated task scheduling. The task scheduling device can be implemented in hardware and / or software, and can be configured in any electronic device with network communication capabilities. Figure 2 As shown, the device includes: a first task processing module 210, a first intelligent construct determination module 220, and a task processing module 230, wherein: First task processing module 210: used to decompose and parse the first task to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task; the second task is a sub-task obtained by decomposing the first task according to the completion steps of the first task. First intelligent construct determination module 220: used to match at least one first intelligent construct according to first information; the first intelligent construct is a processing unit capable of processing a first task, and the first intelligent construct is configured with a first space; the first space is used to record data information generated when the intelligent construct processes the first task; Task processing module 230: used to process at least one second task based on at least one first intelligent construct to obtain a first result; the first result is the completion result of the first task.

[0113] Optionally, the first task processing module 210 includes: The first task involves semantic parsing and data cleaning to obtain the second information. Keyword extraction is performed on the second information to obtain the third information; keywords are words that can represent the content of the first task. Based on the third information, the first task is decomposed and structured to obtain at least one second task and the first information.

[0114] Optionally, the task processing module 230 includes: Second result determination unit: used to match at least one second task with at least one first intelligent construct and process them to obtain at least one second result; First result determination unit: used to integrate at least one second result to obtain a first result.

[0115] Optionally, the second result determination unit includes: Second intelligent construct determination subunit: used to match at least one second task with at least one first intelligent construct to obtain at least one second intelligent construct; the second intelligent construct corresponds one-to-one with the second task; The fourth information determination subunit is used to determine the first space based on the second intelligent construct and to determine the fourth information from the first space; the fourth information is used to characterize the execution information required by the second intelligent construct to perform task processing. Processing subunit: Used by the second intelligent construct to process the corresponding second task based on the fourth information.

[0116] Optional, processing subunit, specifically used for: The fifth piece of information is identified; the fifth piece of information is the task information corresponding to the second task. The fourth piece of information is compared with the fifth piece of information to obtain the second result; the second result is used to characterize whether the fourth piece of information contains the fifth piece of information. If the second result contains information, then the second task is processed through the second intelligent construct; If the second result does not contain information, then the third smart construct is matched.

[0117] Optionally, the task processing module 230 includes: First node status information determination unit: used to determine the first node status information; the first node status information is the status information of each pre-set processing node of the second task; First Action Determination Unit: Used to determine the first action based on the first node status information and preset status information; Processing unit: used to process the second space according to the first action; the second space is the context space of the first intelligent construct corresponding to the first node.

[0118] The task scheduling device provided in the embodiments of the present invention can execute the task scheduling method provided in any of the embodiments of the present invention, and has the corresponding functions and beneficial effects of executing the task scheduling method. For details, please refer to the relevant operations of the task scheduling method in the foregoing embodiments.

[0119] Figure 3This is a schematic diagram of the structure of an electronic device for implementing the task scheduling method of an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0120] like Figure 3 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0121] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0122] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as task scheduling methods.

[0123] In some embodiments, the task scheduling method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the task scheduling method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the task scheduling method by any other suitable means (e.g., by means of firmware).

[0124] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0125] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0126] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0127] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0128] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0129] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0130] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0131] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A task scheduling method, characterized in that, include: The first task is broken down and analyzed to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task. The second task is a sub-task obtained by breaking down the first task according to its completion steps; At least one first intelligent construct is matched based on the first information; the first intelligent construct is a processing unit capable of processing the first task, and the first intelligent construct is configured with a first space; The first space is used to record the data information generated when the intelligent construct processes the first task; The at least one second task is processed according to the at least one first intelligent construct to obtain a first result; The first result is the completion result of the first task.

2. The method according to claim 1, characterized in that, The process of breaking down and analyzing the first task to obtain at least one second task and first information includes: The first task is semantically parsed and data cleaned to obtain the second information; The second information is used to extract keywords to obtain the third information; the keywords are words that can characterize the content of the first task. Based on the third information, the first task is decomposed and structured to obtain at least one second task and the first information.

3. The method according to claim 1, characterized in that, The step of processing the at least one second task according to the at least one first intelligent construct to obtain a first result includes: Match at least one second task with at least one first intelligent construct and process them to obtain at least one second result; The at least one second result is integrated to obtain a first result.

4. The method according to claim 3, characterized in that, The process of matching at least one second task with at least one first intelligent construct and performing the matching includes: At least one second task is matched with at least one first intelligent construct to obtain at least one second intelligent construct; the second intelligent construct corresponds one-to-one with the second task. A first space is determined based on the second intelligent construct, and fourth information is determined from the first space; the fourth information is used to characterize the execution information required by the second intelligent construct for task processing. The second intelligent construct processes the corresponding second task based on the fourth information.

5. The method according to claim 4, characterized in that, The second intelligent construct processes the corresponding second task based on the fourth information, including: The fifth piece of information is determined; the fifth piece of information is the task information corresponding to the second task. The fourth information is compared with the fifth information to obtain a second result; the second result is used to characterize whether the fourth information contains the fifth information. If the second result contains information, then the second task is processed through the second intelligent construct; If the second result does not contain information, then a third smart construct is matched.

6. The method according to claim 1, characterized in that, The process of processing the at least one second task based on the at least one first intelligent construct includes: Determine the status information of the first node; the status information of the first node is the status information of the processing node of each second task that is preset. The first action is determined based on the first node status information and the preset status information; The second space is processed according to the first action; the second space is the context space of the first intelligent construct corresponding to the first node.

7. A task scheduling device, characterized in that, include: The first task processing module is used to decompose and parse the first task to obtain at least one second task and first information; the first information is used to characterize the task requirements of the first task; the second task is a sub-task obtained by decomposing the first task according to its completion steps. The first intelligent construct determination module is used to match at least one first intelligent construct based on the first information; the first intelligent construct is a processing unit capable of processing the first task, and the first intelligent construct is configured with a first space; the first space is used to record data information generated by the intelligent construct when processing the first task. A task processing module is used to process the at least one second task according to the at least one first intelligent construct to obtain a first result; The first result is the completion result of the first task.

8. The apparatus according to claim 7, characterized in that, The first task processing module includes: The first task is semantically parsed and data cleaned to obtain the second information; The second information is used to extract keywords to obtain the third information; the keywords are words that can characterize the content of the first task. Based on the third information, the first task is decomposed and structured to obtain at least one second task and the first information.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the task scheduling method according to any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the task scheduling method of any one of claims 1-6.

Images