Intelligent agent scheduling method and device, storage medium and electronic equipment

Abstract

The invention discloses an intelligent agent scheduling method and device, a storage medium and electronic equipment. The scheduling method comprises the following steps: in response to a user instruction, processing the user instruction according to a preset situation analysis rule to obtain situation data corresponding to the user instruction; according to the situation data and the user instruction, judging whether the user instruction needs to call an intelligent agent; if yes, determining a target agent according to the user instruction and a preset scheduling strategy; determining an input parameter of the target agent according to the situation data, the user instruction and a preset parameter determination rule; and calling the target agent to enable the target agent to execute the user instruction according to the input parameter.

Description

Technical Field

[0001] This application relates to the technical field of agent scheduling. Specifically, it relates to a scheduling method, device, storage medium, and electronic device for an agent. Background Art

[0002] An agent system is an intelligent technology that can automatically complete specific task instructions according to preset rules and algorithms.

[0003] The inventors of this application found that currently, during the agent task allocation process, the agent system often requires a large amount of manual intervention to accurately select a suitable agent, which not only reduces the response speed of the agent system but also increases the workload of the operator.

[0004] The content of the background art section is only the technology known to the applicant and does not necessarily represent the prior art in this field. Summary of the Invention

[0005] According to one aspect of this application, this application provides a scheduling method for an agent. The scheduling method includes: in response to a user instruction, processing the user instruction according to a preset situation analysis rule to obtain situation data corresponding to the user instruction; determining whether the user instruction needs to invoke an agent based on the situation data and the user instruction; if so, determining a target agent according to the user instruction and a preset scheduling strategy; determining input parameters of the target agent according to the situation data, the user instruction, and a preset parameter determination rule; and invoking the target agent so that the target agent executes the user instruction according to the input parameters.

[0006] According to some embodiments of this application, in response to a user instruction, processing the user instruction according to a preset situation analysis rule to obtain situation data of the user instruction includes: in response to the user instruction, extracting key information of the user instruction; generating a query condition based on the key information to determine situation data that meets the user instruction according to the query condition.

[0007] According to some embodiments of this application, if so, determining a target agent according to the user instruction and a preset scheduling strategy includes: if so, determining the matching degree of the user instruction according to a preset matching degree rule; and determining the target agent according to the matching degree of the user instruction and the preset scheduling strategy.

[0008] According to some embodiments of the present application, rules are determined based on situation data, user instructions, and preset parameters to determine the input parameters of the target agent, including: generating the first input parameter of the target agent according to the situation data, user instructions, and the first preset parameter determination rule; when it is determined that the first input parameter meets the preset parameter condition, determining the first input parameter as the input parameter; when it is determined that the first input parameter does not meet the preset parameter condition, sending an input instruction to the user interface so that the user inputs the second input parameter according to the input instruction; determining the input parameter according to the first input parameter and the second input parameter from the user interface.

[0009] According to some embodiments of the present application, after invoking the target agent to enable the target agent to execute the target task according to the input parameters, the scheduling method further includes: collecting the task status parameters of the target agent executing the user instruction; determining the task result parameters according to the task status parameters; sending the task result parameters to the user interface.

[0010] According to another aspect of the present application, the present application provides a scheduling device for an agent. The scheduling device includes a situation analysis module, an agent scheduling module, a parameter parsing module, and an execution control module. The situation analysis module, in response to a user instruction, processes the key information of the user instruction according to a preset situation analysis rule to obtain the situation data of the user instruction; the agent scheduling module determines whether the user instruction needs to invoke an agent in the agent library according to the situation data and the user instruction; when the agent scheduling module determines that an agent needs to be invoked, it determines the target agent according to the user instruction and a preset scheduling strategy; the parameter parsing module determines the input parameters of the target agent according to the situation data, user instructions, and preset parameter determination rules; the execution control module invokes the target agent to enable the target agent to execute the target task according to the input parameters, where the target task is determined according to the user instruction.

[0011] According to another aspect of the present application, the present application further provides a non-volatile computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it can implement the agent scheduling method as described above.

[0012] According to another aspect of the present application, the present application further provides an electronic device, including: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors can implement the agent scheduling method as described above.

[0013] According to another aspect of the present application, the present application also provides a computer program product, including: a computer program stored on a computer-readable storage medium; the computer program includes program instructions, when the program instructions are executed by a computer, enabling the computer to execute the scheduling method of the agent as described above.

[0014] The scheduling method provided by the present application can process user instructions through preset situation analysis rules to obtain situation data corresponding to the user instructions. The present application can determine whether a user instruction needs to invoke an agent based on the situation data and the user instruction. The present application can determine a target agent through the user instruction and a preset scheduling strategy in the case of determining that an agent needs to be invoked. The present application can determine the input parameters of the target agent through the situation data, the user instruction, and a preset parameter determination rule. The present application can invoke the target agent so that the target agent executes the user instruction according to the input parameters.

[0015] The scheduling method provided by the present application realizes the automatic selection and task assignment of agents through the collaborative work of a situation analysis module and an agent scheduling module, significantly reducing the need for manual intervention. It not only improves the response speed of the scheduling device to user instructions but also reduces the workload of operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 A flowchart showing the scheduling method 1000 according to an embodiment of the present application;

[0018] Figure 2 A flowchart showing step S110 according to an embodiment of the present application;

[0019] Figure 3 A flowchart showing step S130 according to an embodiment of the present application;

[0020] Figure 4 A flowchart showing step S140 according to an embodiment of the present application;

[0021] Figure 5 Another flowchart showing the scheduling method 1000 according to an embodiment of the present application;

[0022] Figure 6 A structural diagram showing the scheduling device according to an embodiment of the present application;

[0023] Figure 7 Another structural schematic diagram of a scheduling device according to an embodiment of the present application is shown.

[0024] Reference numerals:

[0025] Scheduling device 200.

[0026] Situation analysis module 210; agent scheduling module 220; parameter parsing module 230; execution control module 240; agent management module 250. Detailed implementation manners

[0027] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art. Identical reference numerals in the figures denote identical or similar parts, and thus their repetitive description will be omitted.

[0028] The features, structures, or characteristics described may be combined in one or more embodiments in any suitable manner. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present disclosure. However, those skilled in the art will realize that the technical solutions of the present disclosure may be practiced without one or more of these specific details, or may be implemented in other ways, components, materials, devices, etc. In such cases, well-known structures, methods, devices, implementations, materials, or operations will not be shown or described in detail.

[0029] In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0030] The terms "first", "second", etc. in the description and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order.

[0031] Next, the technical solutions of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative efforts fall within the scope of protection of this application.

[0032] As Figure 6As shown, the present application provides a scheduling device 200 for an intelligent agent. The scheduling device 200 may include a situation analysis module 210, an intelligent agent scheduling module 220, a parameter parsing module 230, and an execution control module 240. As Figure 7 shown, the scheduling device 200 may include a situation analysis module 210, an intelligent agent scheduling module 220, a parameter parsing module 230, an execution control module 240, and an intelligent agent management module 250. The scheduling device 200 may schedule the intelligent agent according to the scheduling method of the intelligent agent.

[0033] Next, in combination with Figure 6 and Figure 7 to describe the scheduling method of the intelligent agent provided by the present application.

[0034] Referring to Figure 1 , the scheduling method 1000 of the intelligent agent may include steps S110 - S150.

[0035] In step S110, the situation analysis module 210 responds to the user instruction, processes the user instruction according to the preset situation analysis rule, to obtain the situation data corresponding to the user instruction.

[0036] According to the exemplary embodiment, the intelligent agent may be a program that autonomously executes a certain instruction or task.

[0037] The situation analysis module 210 may be a server or a host with data processing capabilities.

[0038] The user instruction may be task instruction information from the user. The user may input the user instruction through the user interface.

[0039] The preset situation analysis rule may be a rule for performing situation analysis on the user instruction. The preset situation analysis rule may include extracting key information in the user group instruction, determining the key information query condition according to the key information, and filtering and screening the situation data that meets the user instruction according to the key information query condition.

[0040] The situation data may be parameter data describing the state of the current scheduling device 200. For example, the situation data may be data such as the price, trading volume, capital inflow, and outflow of a certain stock at present; for another example, in intelligent transportation, the situation data may be data such as traffic congestion situation data and vehicle flow data on the traffic road.

[0041] In step S120, the intelligent agent scheduling module 220 determines whether the user instruction needs to call the intelligent agent according to the situation data and the user instruction.

[0042] According to an exemplary embodiment, the agent library may include at least one agent. The agent scheduling module 220 may be a server or a host with data processing capabilities. The agent scheduling module 220 may determine whether the user instruction needs to call an agent through large model analysis.

[0043] If so, step S130 is executed. In step S130, the agent scheduling module 220 determines the target agent according to the user instruction and the preset scheduling policy.

[0044] According to an exemplary embodiment, the preset scheduling policy may be a preset policy rule for calling an agent. The preset scheduling policy may include determining the matching degree of the user instruction and selecting the target agent.

[0045] The agent scheduling module 220 may determine the matching degree between the user instruction and the agent according to the user instruction. The agent scheduling module 220 may determine the target agent according to the matching degree, the historical execution effect of the agent, and the agent resource consumption information.

[0046] If not, the process ends.

[0047] After step S130, step S140 is executed. In step S140, the parameter parsing module 230 determines the input parameters of the target agent according to the situation data, the user instruction, and the preset parameter determination rule.

[0048] According to an exemplary embodiment, the input parameters of the target agent may be the input parameters required for the target agent to execute the instruction. The input parameters of the target value agent may include the task type, the target object, and the constraint conditions, etc.

[0049] The preset parameter determination rule may be the input parameters of the target agent determined according to the situation data and the user instruction.

[0050] In step S150, the execution control module 240 calls the target agent so that the target agent executes the user instruction according to the input parameters.

[0051] According to an exemplary embodiment, the parameter parsing module 230 may transmit the input parameters of the target agent to the execution control module 240. The execution control module 240 may call the target agent according to the selected target agent. The execution module may transmit the input parameters of the target agent to the target agent.

[0052] The execution control module 240 may call the target agent through a unified call interface. The interface definition includes: the input parameter set, the execution timeout time, and the return value format, etc.

[0053] Through the above embodiments, the scheduling method provided by the present application can process user instructions through preset situation analysis rules to obtain situation data corresponding to the user instructions. The present application can determine whether a user instruction needs to invoke an agent based on the situation data and the user instruction. When it is determined that an agent needs to be invoked, the present application can determine the target agent through the user instruction and a preset scheduling strategy. The present application can determine the input parameters of the target agent by using the situation data, the user instruction, and a preset parameter determination rule. The present application can invoke the target agent so that the target agent executes the user instruction according to the input parameters.

[0054] The scheduling method provided by the present application realizes the automatic selection and task allocation of agents through the collaborative work of the situation analysis module 210 and the agent scheduling module 220, significantly reducing the need for manual intervention. This not only improves the response speed of the scheduling device 200 to user instructions but also reduces the workload of the operator.

[0055] Optionally, referring to Figure 2 , step S110 may include step S111 and step S112.

[0056] The situation analysis module 210 includes a semantic parsing unit, a data filtering unit, and a large model summarization unit.

[0057] In step S111, the semantic parsing unit extracts the key information of the user instruction in response to the user instruction.

[0058] According to the exemplary embodiment, the semantic parsing unit can parse the user instruction through natural language processing technology to extract the key information of the user instruction. The key information may include the task type, the target object, the constraint conditions, etc. The semantic parsing unit can extract the key information in the user instruction through a preset keyword dictionary, a preset semantic rule, and deep learning, etc.

[0059] For example, if the user instruction is "Track the target with the number A001 and keep the tracking time not less than 10 minutes", the semantic parsing unit can extract through the preset keyword dictionary and the preset semantic rule that the task type in the user instruction is "target tracking", the target object is "A001", and the constraint condition is "tracking time ≥ 10 minutes".

[0060] In step S112, the data filtering unit generates a query condition based on the key information to determine the situation data that meets the user instruction according to the query condition.

[0061] According to the exemplary embodiment, the key information query condition can be the condition for querying the situation database determined according to the key information. The data filtering unit can construct the key information query condition according to the extracted key information. The data filtering unit filters the information in the situation database according to the key information query condition to filter out the situation data that meets the user's instruction, so as to determine the situation data of the user's instruction. The situation data determined by the data filtering unit is structured data.

[0062] For example, the data filtering unit constructs a key information query condition based on the task type being "target tracking", the target object being "A001", and the constraint condition being "tracking time ≥ 10 minutes". The key information query condition can be to find the situation data where the task type is "target tracking", the target object is "A001", and the constraint condition is "tracking time ≥ 10 minutes". The data filtering unit can query the situation database according to the key information query condition, so as to filter out the situation data such as the current position of the target object A001, whether the target object A001 is aggressive, and which of our units are in the idle state and can execute tasks.

[0063] Optionally, after step S112, the large model summarization unit can also summarize and generalize the situation data through a large language model, so as to propose suggestions for action decisions according to the current situation.

[0064] For example, the suggestion is: A001 is currently in the southern region, A001 is not aggressive, A001 has a slow speed, and it is recommended that our side set the speed to low speed and track at B003 in the southern region.

[0065] Optionally, referring to Figure 3 , step S130 can include step S131 and step S132.

[0066] If so, execute step S131. In step S131, the agent scheduling module 220 determines the matching degree of the user instruction according to the preset matching degree rule.

[0067] According to the exemplary embodiment, the preset matching degree rule can be the rule for determining the matching degree between the user instruction and the agent.

[0068] The preset matching degree rule can include the task type, the situation data, and the agent capability description information. The task type can be the task type in the key information. The agent description information can be the information describing the functions and performance of the agent.

[0069] The agent scheduling module 220 can calculate the matching degree according to the task type, the situation data, and the agent capability description information. The matching degree calculation can adopt the weighted scoring method.

[0070] In step S132, the agent scheduling module 220 determines a target agent according to the matching degree of the user instruction and a preset scheduling policy.

[0071] According to the exemplary embodiment, the target agent can be an agent that can execute the user instruction and has a good execution effect. The preset scheduling policy can be a policy for determining the target agent. For example, the preset scheduling policy can include the matching degree of the user instruction, the historical execution effect of the agent, and the agent resource consumption information.

[0072] For example, there are three agents in the agent library, namely agent A, agent B, and agent C. Agent A is suitable for processing simple target tracking tasks, has a fast execution speed, and low resource occupancy. Agent B is suitable for processing complex image recognition tasks, has a slower execution speed, and higher resource occupancy. Agent C is suitable for processing natural language processing tasks, has a medium execution speed, and medium resource occupancy.

[0073] When the user instruction is "recognize the target type in the picture", the agent scheduling module 220 determines that an agent needs to be called and selects agent B to process the user instruction according to the preset scheduling policy because the function of agent B best matches the user instruction.

[0074] Optionally, referring to Figure 4 , step S140 may include steps S141 - S144.

[0075] In step S141, the parameter parsing module 230 determines a rule based on the situation data, the user instruction, and a first preset parameter to generate a first input parameter for the target agent.

[0076] The first input parameter of the target agent can be the input parameter of the target agent determined according to the situation data and the user instruction.

[0077] The first preset parameter determination rule can be a rule for determining the first input parameter. The first preset parameter determination rule can include parameter extraction, parameter verification, and parameter conversion.

[0078] The parameter parsing module 230 can extract the first input parameter from the situation data and the user instruction. For example, the parameter parsing module 230 can extract the first input parameter of the target agent from the situation data and the user instruction through a preset input parameter list of the target agent.

[0079] The parameter parsing module 230 verifies the extracted first input parameter to ensure that the type of the first input parameter is correct and satisfies constraint conditions such as the value range of the first input parameter of the target agent.

[0080] The parameter parsing module 230 can convert the data type of the first input parameter into a data type that the target agent can accept.

[0081] In step S142, when the parameter parsing module 230 determines that the first input parameter meets the preset parameter conditions, the first input parameter is determined as the input parameter.

[0082] Optionally, the preset parameter conditions can be that the first input parameter does not lack the necessary input parameters for the target agent to execute the user instruction.

[0083] The parameter parsing module 230 can determine whether the first input parameter meets the preset parameter conditions according to the preset parameter conditions.

[0084] When the parameter parsing module 230 determines that the first input parameter meets the preset parameter conditions, the first input parameter is determined as the input parameter.

[0085] For example, when the parameter parsing module 230 determines that the first input parameter does not lack the necessary input parameters for the target agent to execute the instruction, the first input parameter is determined as the input parameter.

[0086] In step S143, when the parameter parsing module 230 determines that the first input parameter does not meet the preset parameter conditions, an input instruction is sent to the user interface so that the user can input a second input parameter according to the input instruction.

[0087] According to the exemplary embodiment, the second input parameter can be the input parameter of the target agent input by the user. When the parameter parsing module 230 determines that the first input parameter lacks the necessary input parameters for the target agent to execute the instruction, the parameter parsing module 230 sends an input instruction to the user interface, and the user can input the second input parameter in the user interface according to the input instruction. Or the parameter parsing module 230 can also take the default value of the missing parameter from the preset input parameter list.

[0088] For example, if the agent requires the coordinate information of the target as an input parameter, but the user instruction does not provide this information, the parameter parsing module 230 can send an input instruction. A dialog box can automatically pop up on the user interface, requesting the user to select the target area on the map. Or the parameter parsing module 230 uses the default coordinate value.

[0089] In step S144, the parameter parsing module 230 determines the input parameter according to the first input parameter and the second input parameter from the user interface.

[0090] According to an exemplary embodiment, after the user inputs a second input parameter in the user interface, the parameter parsing module 230 can receive the second input parameter. The parameter parsing module 230 determines the first input parameter and the second input parameter as the input parameter.

[0091] Through the above embodiments, the parameter parsing module 230 provided by the present application determines the input parameter of the target agent, solving the problems of complex agent parameter configuration and high usage threshold. When the input parameter is missing or incomplete, the parameter parsing module 230 can effectively process it through human-computer interaction, improving the fault tolerance of the scheduling device 200 and the success rate of executing user instructions.

[0092] Optionally, referring to Figure 5 , the scheduling method 1000 may further include steps S160 - S180.

[0093] In step S160, the execution control module 240 collects the task status parameters in the target agent's execution of the user instruction.

[0094] According to an exemplary embodiment, the task status parameter may be the running status parameter of the target agent during the execution of the user instruction by the target agent. The task status parameter may include execution progress, resource occupancy, and exception information, etc.

[0095] For example, the execution control module 240 may collect the task status parameters of the target agent in real time during the execution of the user instruction by the target agent, so as to monitor the situation of the target agent's execution of the user instruction.

[0096] In step S170, the execution control module 240 determines the task result parameter according to the task status parameter.

[0097] According to an exemplary embodiment, the task result parameter may be the result parameter after the target agent completes the execution of the user instruction. The considered result parameter may include an execution status code, result data, error information, and execution time consumption, etc. The execution control module 240 may record the task status parameter and the result status parameter in the execution log.

[0098] In step S180, the execution control module 240 sends the task result parameter to the user interface.

[0099] According to an exemplary embodiment, the execution control module 240 may perform standardization processing on the task result parameter and display it to the user interface in forms such as text, table, and image.

[0100] For example, the execution control module 240 may send the task result parameter to the user interface in the form of a pop-up window, so as to feedback the situation of executing the user instruction to the user.

[0101] The execution control module 240 can also send task status parameters to the user interface. For example, if an error occurs during the execution of the target agent, the execution control module 240 will capture the exception information and feedback the error information to the user interface, such as prompting the user in the form of a pop-up window "The execution of the target agent fails, reason: xxxx".

[0102] Through the above embodiments, the scheduling method provided by the present application can establish a feedback mechanism during the execution of the user instruction by the target agent, enabling the scheduling device 200 to timely discover and handle problems during the execution process and efficiently perform interactive feedback, improving the reliability of the scheduling device 200 and the user interaction experience.

[0103] Optionally, before step S110, the agent management module 250 determines the agents in the agent library according to the information of each agent.

[0104] According to the exemplary embodiment, the agent management module 250 can build and manage agents. The agent management module 250 can use a document-type database to store the information of agents.

[0105] Each agent information can include fields such as a unique identifier (Identifier, id), agent name (name), function description (description), version number (version), creation time (created_at), last update time (updated_at), input parameter list (input_params), output parameter definition (output_params), execution conditions (conditions), applicable scenario description (applicable_scenarios), and execution priority (priority).

[0106] The agent id can be the unique identifier of the agent. The name of the agent, such as "Target Tracking Agent". The function description can be a brief description of the agent's function, such as "Used to track the movement trajectory of the target in real time". The version number of the agent, such as "v1.0", is used to trace the version changes of the agent.

[0107] The creation time of the agent can record the creation timestamp. The last update time can be the time when the agent was last updated, recording the update timestamp. The input parameter list can be stored in JSON format. Each input parameter in the input parameter list can include attributes such as parameter name, data type, whether it is required, and default value.

[0108] The output parameter definition can be the definition of the return value of the agent. The output parameter definition can be a list of agent return values stored in JSON format. Each return value contains attributes such as the return value name, type, description, etc., and is used for interactive feedback presentation.

[0109] The applicable scenario description can be a description of the applicable scenario of the agent, such as "applicable to radar target tracking". The execution priority can be the execution priority of the agent, such as "high", "medium", "low".

[0110] For example, the agent management module 250 can describe the information of a target tracking agent as follows:

[0111]

[0112]

[0113] The agent management module 250 can also update and manage agents, such as adding, deleting, modifying, and querying agents.

[0114] Through the above embodiments, the calling method provided by the present application can uniformly describe and standardize the management of agents through the agent management module 250, effectively solving the problem of chaotic agent management in the existing agent system, and greatly improving the efficiency of organizing and calling agents in the agent calling system.

[0115] According to another aspect of the present application, the present application also provides a non-volatile computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it can implement the scheduling method of the agent as described above.

[0116] According to another aspect of the present application, the present application also provides an electronic device, including: one or more processors; a storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors can implement the scheduling method of the agent as described above.

[0117] According to another aspect of the present application, the present application also provides a computer program product, including: a computer program stored on a computer-readable storage medium; the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the scheduling method of the agent as described above.

[0118] Finally, it should be noted that the above are only the preferred embodiments of the present application and are not used to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions of the foregoing embodiments or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A scheduling method for an agent, characterized in that, Including: In response to a user instruction, process the user instruction according to a preset situation analysis rule to obtain situation data corresponding to the user instruction; Judge whether the user instruction needs to call an agent according to the situation data and the user instruction; If so, determine a target agent according to the user instruction and a preset scheduling policy; Determine input parameters of the target agent according to the situation data, the user instruction, and a preset parameter determination rule; Call the target agent so that the target agent executes the user instruction according to the input parameters.

2. The scheduling method according to claim 1, wherein The step of, in response to a user instruction, processing the user instruction according to a preset situation analysis rule to obtain situation data of the user instruction includes: In response to the user instruction, extract key information of the user instruction; Generate a query condition according to the key information to determine the situation data that meets the user instruction according to the query condition.

3. The scheduling method according to claim 1, wherein The step of, if so, determining a target agent according to the user instruction and a preset scheduling policy includes: If so, determine a matching degree of the user instruction according to a preset matching degree rule; Determine the target agent according to the matching degree of the user instruction and the preset scheduling policy.

4. The scheduling method according to claim 1, wherein The step of determining input parameters of the target agent according to the situation data, the user instruction, and a preset parameter determination rule includes: Generate first input parameters of the target agent according to the situation data, the user instruction, and a first preset parameter determination rule; When it is determined that the first input parameters meet the preset parameter conditions, determine the first input parameters as the input parameters; When it is determined that the first input parameters do not meet the preset parameter conditions, send an input instruction to a user interface so that a user inputs second input parameters according to the input instruction; Determine the input parameters according to the first input parameters and the second input parameters from the user interface.

5. The scheduling method according to claim 1, wherein After calling the target agent so that the target agent executes a target task according to the input parameters, the scheduling method further includes: Collect task status parameters of the target agent when executing the task in the user instruction; Determine task result parameters according to the task status parameters; Send the task result parameters to the user interface.

6. The scheduling method according to claim 4, wherein The preset parameter condition is that the first input parameters do not lack necessary input parameters for the target agent to execute the user instruction.

7. A scheduling device for an intelligent agent, characterized in that, Including: A situation analysis module, which, in response to a user instruction, processes key information of the user instruction according to a preset situation analysis rule to obtain situation data of the user instruction; An agent scheduling module, which judges whether the user instruction needs to call an agent in an agent library according to the situation data and the user instruction; When the agent scheduling module judges that an agent needs to be called, determine a target agent according to the user instruction and a preset scheduling policy; A parameter parsing module, which determines input parameters of the target agent according to the situation data, the user instruction, and a preset parameter determination rule; The execution control module invokes the target agent to enable the target agent to execute the target task according to the input parameters, where the target task is determined according to the user instruction.

8. A non-volatile computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the scheduling method of the agent as described in any one of claims 1-6.

9. An electronic device, characterized in that, Comprising: One or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the scheduling method of the agent as described in any one of claims 1-6.

10. A computer program product, characterized in that, It includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the scheduling method of the agent as described in any one of claims 1-6.

Images