Multi-artificial intelligence agent management method and device, equipment and storage medium

Abstract

The invention discloses a multi-artificial intelligence agent management method and device, equipment and a storage medium, according to the scheme, life cycle and state management is performed on a plurality of current artificial intelligence agents, so that survival and change management of each current artificial intelligence agent is more refined, and subsequent state monitoring and scheduling are more efficient; besides, the target life cycle and the target state are synchronized to other current artificial intelligence agents related to the current artificial intelligence agent, so that the communication capability between the current artificial intelligence agents is enhanced, the multiple current artificial intelligence agents can cooperate more effectively, and the user experience is improved. And the coordination and efficiency of multi-current artificial intelligence agent cooperation are improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and storage medium for managing multiple artificial intelligence agents. Background Art

[0002] With the continuous in-depth research and exploration of artificial intelligence applications, the capabilities of artificial intelligence agents have become increasingly powerful, and the application scenarios have become more and more diversified. Multiple artificial intelligence agents have been able to collaborate to complete the ability of a complex event.

[0003] Currently, the application industry for multiple artificial intelligence agents is still in an earlier stage. Most institutions, universities, enterprises, etc. are in the exploration stage. Some of the more popular mainstream frameworks have relatively weak control capabilities over the states of multiple artificial intelligence agents, and none of them can achieve the monitoring, scheduling and switching of multiple artificial intelligence agents in terms of their life cycles and operating states.

[0004] Therefore, how to effectively manage the states of multiple artificial intelligence agents during collaboration is an urgent problem to be solved. Summary of the Invention

[0005] This application aims to at least solve the technical problems existing in the prior art. To this end, a first aspect of this application proposes a method for managing multiple artificial intelligence agents, and the method includes:

[0006] Obtain multiple currently registered artificial intelligence agents;

[0007] For each currently registered artificial intelligence agent, perform a life cycle transformation on the currently registered artificial intelligence agent based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the currently registered artificial intelligence agent;

[0008] Perform a state transformation on the currently registered artificial intelligence agent based on real-time state information and a preset state transformation rule to obtain a target state corresponding to the currently registered artificial intelligence agent;

[0009] Synchronize the target life cycle and the target state to other currently registered artificial intelligence agents related to the currently registered artificial intelligence agent.

[0010] In a possible implementation manner, performing a life cycle transformation on the currently registered artificial intelligence agent based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the currently registered artificial intelligence agent includes:

[0011] Obtain a set of multiple stages of a life cycle formulated in advance and preset life cycle transformation rules corresponding to each stage;

[0012] Determine the current stage in the life cycle corresponding to the currently registered artificial intelligence agent;

[0013] Perform a lifecycle transformation on the current artificial intelligence agent based on a preset lifecycle transformation rule corresponding to the current stage to obtain the target lifecycle corresponding to the current artificial intelligence agent.

[0014] In a possible implementation, performing a state transformation on the current artificial intelligence agent based on real-time state information and a preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent includes:

[0015] Obtain preset message queues corresponding to each preset state of the current artificial intelligence agent; wherein, the preset message queues include the real-time state information of the current artificial intelligence agent;

[0016] Obtain the real-time state information in the preset message queue;

[0017] Perform a state transformation on the current artificial intelligence agent based on the preset queue operation rule, real-time state information, and preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent.

[0018] In a possible implementation, the method further includes:

[0019] Obtain a state transformation failure rate and a state transformation benefit rate; wherein, the state transformation failure rate is used to represent the probability of a state switch failure, and the state transformation benefit rate is used to represent the efficiency of a state switch;

[0020] Calculate a comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate;

[0021] Update the preset queue operation rule based on the comprehensive state change benefit coefficient to obtain a new queue operation rule.

[0022] In a possible implementation, calculating a comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate includes:

[0023] Obtain a first preset weight corresponding to the state transformation failure rate and a second preset weight corresponding to the state transformation benefit rate;

[0024] Calculate a comprehensive state change benefit coefficient based on the state transformation failure rate, the state transformation benefit rate, the first preset weight, and the second preset weight.

[0025] In a possible implementation, updating the preset queue operation rule based on the comprehensive state change benefit coefficient to obtain a new queue operation rule includes:

[0026] Match the comprehensive state change benefit coefficient with a preset benefit threshold to obtain a matching result;

[0027] Update the preset queue running rules based on the matching results to obtain new queue running rules.

[0028] In a possible implementation, synchronize the target life cycle and target status to other current artificial intelligence agents related to the current artificial intelligence agent, including:

[0029] For each current artificial intelligence agent, identify the role information corresponding to the current artificial intelligence agent;

[0030] Based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information, obtain the execution sequence logic information and dependency relationship information of the current artificial intelligence agent;

[0031] Based on the execution sequence logic information and dependency relationship information, determine other current artificial intelligence agents related to the current artificial intelligence agent;

[0032] Synchronize the target life cycle and target status to other current artificial intelligence agents.

[0033] A multi-artificial intelligence agent management device is proposed in the second aspect of this application. The device includes:

[0034] An acquisition module, configured to acquire multiple registered current artificial intelligence agents;

[0035] A first transformation module, configured to perform life cycle transformation on each current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent;

[0036] A second transformation module, configured to perform status transformation on the current artificial intelligence agent based on real-time status information and a preset status transformation rule to obtain the target status corresponding to the current artificial intelligence agent;

[0037] A synchronization module, configured to synchronize the target life cycle and target status to other current artificial intelligence agents related to the current artificial intelligence agent.

[0038] In a possible implementation, the above first transformation module is specifically configured to:

[0039] Perform life cycle transformation on the current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent, including:

[0040] Obtain a set of multiple stages of the life cycle formulated in advance and the preset life cycle transformation rules corresponding to each stage;

[0041] Determine the current stage in the life cycle corresponding to the current artificial intelligence agent;

[0042] Perform a lifecycle transformation on the current artificial intelligence agent based on a preset lifecycle transformation rule corresponding to the current stage to obtain the target lifecycle corresponding to the current artificial intelligence agent.

[0043] In a possible implementation manner, the above-mentioned second transformation module is specifically configured to:

[0044] Obtain preset message queues corresponding to each preset state of the current artificial intelligence agent; wherein, the preset message queues include real-time state information of the current artificial intelligence agent;

[0045] Obtain the real-time state information in the preset message queue;

[0046] Perform a state transformation on the current artificial intelligence agent based on a preset queue operation rule, real-time state information, and a preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent.

[0047] In a possible implementation manner, the above-mentioned multi-artificial intelligence agent management device is further configured to:

[0048] Obtain a state transformation failure rate and a state transformation benefit rate; wherein, the state transformation failure rate is used to represent the probability of state switching failure, and the state transformation benefit rate is used to represent the efficiency of state switching;

[0049] Calculate a comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate;

[0050] Update the preset queue operation rule based on the comprehensive state change benefit coefficient to obtain a new queue operation rule.

[0051] In a possible implementation manner, the above-mentioned multi-artificial intelligence agent management device is further configured to:

[0052] Obtain a first preset weight corresponding to the state transformation failure rate and a second preset weight corresponding to the state transformation benefit rate;

[0053] Calculate a comprehensive state change benefit coefficient based on the state transformation failure rate, the state transformation benefit rate, the first preset weight, and the second preset weight.

[0054] In a possible implementation manner, the above-mentioned multi-artificial intelligence agent management device is further configured to:

[0055] Match the comprehensive state change benefit coefficient with a preset benefit threshold to obtain a matching result;

[0056] Update the preset queue operation rule based on the matching result to obtain a new queue operation rule.

[0057] In a possible implementation manner, the above-mentioned synchronization module is specifically configured to:

[0058] For each current artificial intelligence agent, identify the role information corresponding to the current artificial intelligence agent;

[0059] Based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information, obtain the sequence logic information and dependency information for the current artificial intelligence agent to execute;

[0060] Based on the sequence logic information and dependency information, determine other current artificial intelligence agents related to the current artificial intelligence agent;

[0061] Synchronize the target life cycle and target state to other current artificial intelligence agents.

[0062] A third aspect of this application proposes an electronic device, which includes a processor and a memory. At least one instruction, at least one program, a code set, or an instruction set is stored in the memory. The at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the multi-artificial intelligence agent management method as described in the first aspect.

[0063] A fourth aspect of this application proposes a computer-readable storage medium. At least one instruction, at least one program, a code set, or an instruction set is stored in the storage medium. The at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the multi-artificial intelligence agent management method as described in the first aspect.

[0064] The embodiments of this application have the following beneficial effects:

[0065] The multi-artificial intelligence agent management method provided by the embodiments of this application includes: obtaining multiple registered current artificial intelligence agents, for each current artificial intelligence agent, performing a life cycle transformation on the current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent, performing a state transformation on the current artificial intelligence agent based on real-time state information and a preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent, and synchronizing the target life cycle and target state to other current artificial intelligence agents related to the current artificial intelligence agent. This solution makes the survival and change management of each current artificial intelligence agent more refined by managing the life cycle and state of multiple current artificial intelligence agents, thereby making subsequent state monitoring and scheduling more efficient; in addition, by synchronizing the target life cycle and target state to other current artificial intelligence agents related to the current artificial intelligence agent, the communication ability between each current artificial intelligence agent is enhanced, and multiple current artificial intelligence agents can cooperate more effectively, improving the coordination and efficiency of multi-current artificial intelligence agent cooperation. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A block diagram of a computer device provided by an embodiment of the present application;

[0067] Figure 2 A flowchart of steps of a multi-artificial intelligence agent management method provided by an embodiment of the present application;

[0068] Figure 3 A flowchart of steps of obtaining a target life cycle provided by an embodiment of the present application;

[0069] Figure 4 A flowchart of steps of obtaining a target state provided by an embodiment of the present application;

[0070] Figure 5 A flowchart of steps of obtaining a new queue operation rule provided by an embodiment of the present application;

[0071] Figure 6 A flowchart of steps of calculating a beneficial coefficient of comprehensive state change provided by an embodiment of the present application;

[0072] Figure 7 A flowchart of steps of obtaining a new queue operation rule provided by an embodiment of the present application;

[0073] Figure 8 A flowchart of steps of synchronizing a target life cycle and a target state provided by an embodiment of the present application;

[0074] Figure 9 A block diagram of the structure of a multi-artificial intelligence agent management device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0075] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0076] Hereinafter, the terms "first" and "second" are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present disclosure, unless otherwise specified, the meaning of "a plurality" is two or more. Additionally, the use of "based on" or "in accordance with" implies openness and inclusiveness, because a process, step, calculation, or other action "based on" or "in accordance with" one or more of the stated conditions or values may, in practice, be based on additional conditions or values beyond those stated.

[0077] The multi-artificial intelligence agent management method provided in this application can be applied to a computer device (electronic device). The computer device can be a server or a terminal. Among them, the server can be a single server or a server cluster composed of multiple servers. The embodiments of this application do not make specific limitations in this regard. The terminal can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, and portable wearable devices.

[0078] Taking the computer device being a server as an example, Figure 1 shows a block diagram of a server, as Figure 1 shown. The server can include a processor and a memory connected through a system bus. Among them, the processor of the server is used to provide computing and control capabilities. The memory of the server includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the computer program is executed by the processor, it realizes a multi-artificial intelligence agent management method.

[0079] Those skilled in the art can understand that Figure 1 the structure shown in

[0080] is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation on the server to which the solution of this application is applied. Optionally, the server can include more or fewer components than shown in the figure, or combine certain components, or have a different component layout.

[0081] Figure 2 is a flowchart of the steps of a multi-artificial intelligence agent management method provided in an embodiment of this application. As Figure 2 shown, the method includes the following steps:

[0082] Step 202: Obtain multiple currently registered artificial intelligence agents.

[0083] Among them, after registering the relevant information of multiple artificial intelligence agents, multiple currently registered artificial intelligence agents can be obtained. Optionally, standardization processing can be performed on each artificial intelligence agent to be registered, including but not limited to standardizing information such as task itself information, environmental information, and unique information of the artificial intelligence agent.

[0084] Among them, the task itself information may include task objectives, task risk points, possible data volume, the number of artificial intelligence agents involved, etc.; the environmental information may include hardware and software information, network information, upstream and downstream other resource information, etc.; the unique information of the artificial intelligence agent may include detailed information, functions, descriptions, permissions, frameworks used, business scenario information, list of executable tools, etc. of the artificial intelligence agent.

[0085] Next, according to different business scenarios, identify the role information corresponding to each artificial intelligence agent in the multi-artificial intelligence agent. Exemplarily, in the artificial intelligence agent of the sales team, it may include roles such as sales manager, sales representative, sales consultant, sales operations manager, sales trainer, sales analyst, sales specialist, regional manager, sales consultant, etc. The above identification of sales roles can be obtained by analyzing the information of all standardized artificial intelligence agents with the help of a preset large prediction model and then summarizing.

[0086] Optionally, the execution sequence information and dependency relationship information of the current artificial intelligence agent can also be obtained based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information.

[0087] Step 204: For each current artificial intelligence agent, perform a life cycle transformation on the current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent.

[0088] Among them, after obtaining the current artificial intelligence agent, a life cycle transformation can be performed on the current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent. In some optional embodiments, as Figure 3 shown, Figure 3 is a flowchart of steps for obtaining a target life cycle provided by an embodiment of the present application, including:

[0089] Step 302: Obtain a set of multiple stages of the life cycle formulated in advance and the preset life cycle transformation rules corresponding to each stage.

[0090] Step 304: Determine the current stage in the life cycle corresponding to the current artificial intelligence agent.

[0091] Step 306: Perform a life cycle transformation on the current artificial intelligence agent based on the preset life cycle transformation rule corresponding to the current stage to obtain the target life cycle corresponding to the current artificial intelligence agent.

[0092] Among them, for different tasks, strategies can be formulated for the life cycles of each artificial intelligence agent, that is, a set of multiple stages of the life cycle can be pre-formulated. Optionally, a set of life cycles can include a neonatal period, an operating period, a decline period, a death period, etc.

[0093] Next, for each stage, a preset life cycle transformation rule corresponding to each stage can be pre-formulated. Exemplarily, a possible preset life cycle transformation rule can be: The initial stage of the life cycle of the artificial intelligence agent is defaulted to the neonatal period. Once the running state of its task changes, the stage where the overall life cycle of the artificial intelligence agent is located is the operating period. If there are multiple state switches in the state transformation rule that are rolled back, it enters the decline period, and after returning to normal, it enters the operating period again. After normal operation ends or ends abnormally, it enters the death period.

[0094] Therefore, after determining the current stage in the life cycle corresponding to the current artificial intelligence agent, the life cycle of the current artificial intelligence agent can be transformed based on the preset life cycle transformation rule corresponding to the current stage to obtain the target life cycle corresponding to the current artificial intelligence agent.

[0095] In some optional embodiments, for the above life cycle, real-time monitoring can be performed. In the case where multiple artificial intelligence agents enter the decline period simultaneously in a complex task, manual intervention can be performed.

[0096] Step 206: Perform a state transformation on the current artificial intelligence agent based on the real-time state information and the preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent.

[0097] Among them, the states of each current artificial intelligence agent can also be transformed. In some optional embodiments, as Figure 4 shown Figure 4 is a flowchart of the steps for obtaining the target state provided by the embodiment of the present application, including:

[0098] Step 402: Obtain the preset message queue corresponding to each preset state of the current artificial intelligence agent.

[0099] Step 404: Obtain the real-time state information in the preset message queue.

[0100] Step 406: Transform the state of the current artificial intelligence agent based on the preset queue operation rules, real-time status information, and preset state transformation rules to obtain the target state corresponding to the current artificial intelligence agent.

[0101] Among them, different state sets can be set for different business scenarios. Exemplarily, a typical state set may include ready state, waiting state, running state, paused state, revoked state, ended state, etc.

[0102] Next, corresponding preset state transformation rules can be formulated for each state to ensure the orderly scheduling and switching of each state. Exemplarily, a feasible preset state transformation rule can be: the ready state can be switched to the waiting state at any time, and the waiting state can be switched to the running state according to a certain strategy. For example, in the queue where the waiting state is located, according to the first-come-first-served policy, the task that enters the queue first is obtained for switching. Then, the running state can be switched to the paused state or the revoked state according to other information of the task execution. If the running state runs until the task ends, it can enter the ended state, and the task in the ended state will not undergo any transformation.

[0103] Therefore, when transforming the states of each current artificial intelligence agent, the preset message queues corresponding to the preset states of the current artificial intelligence agent can be obtained first. Among them, the preset message queues may include the real-time status information of the current artificial intelligence agent. After obtaining the real-time status information in the preset message queues, the state of the current artificial intelligence agent can be transformed based on the preset queue operation rules, real-time status information, and preset state transformation rules, so as to obtain the target state corresponding to the current artificial intelligence agent.

[0104] Among them, when formulating the preset queue operation rules, the preset message queues corresponding to each state can be initialized first, and each initialized preset message queue is empty. Then, the preset queue operation rules can be formulated to ensure the orderly scheduling and switching of each queue.

[0105] Exemplarily, a feasible preset queue operation rule can be: set a real-time or timing mechanism to detect the real-time status information in each queue in real-time or at regular intervals. For example, in the queue of the waiting state, all stored tasks are in the waiting state, and this task also carries other information, such as the entry time into the queue, task itself information, etc. If a certain real-time status information in a certain queue reaches the above preset state transformation rules, the current state is popped out from this preset message queue and enters the next state to be accepted. It should be noted that which state to enter specifically is still determined by the preset state transformation rules.

[0106] Finally, based on the above-mentioned preset queue operation rules, real-time status information, and preset status transformation rules, the status of the current artificial intelligence agent is transformed, and the target status corresponding to the current artificial intelligence agent can be determined.

[0107] In some alternative embodiments, for certain scenarios, the preset queue operation rules formulated are not the optimal rules. Therefore, an intelligent operation mechanism can be designed to correct the queue operation strategy. For example, Figure 5 as shown, Figure 5 is a flowchart of steps for obtaining a new queue operation rule provided by an embodiment of the present application, including:

[0108] Step 502, obtain the state transformation failure rate and the state transformation benefit rate.

[0109] Step 504, calculate the comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate.

[0110] Step 506, update the preset queue operation rule based on the comprehensive state change benefit coefficient to obtain a new queue operation rule.

[0111] Among them, the state transformation failure rate is used to represent the probability of state switching failure, and the state transformation benefit rate is used to represent the efficiency of state switching.

[0112] When calculating the state transformation failure rate, the state transformation failure rate is also the probability that, during operation, according to the preset queue operation rule and the preset state transformation rule, the state should have been able to be transformed but the transformation fails or it is found that subsequent tasks cannot be carried out after the transformation. The state switching failure rate can be denoted as switch_err_rate, where switch_err_count is the number of change failures, and switch_count is the total number of transformations.

[0113] When calculating the state transformation benefit rate, when a state needs to be transformed, first perform the state transformation of all artificial intelligence agents corresponding to the tasks, run them in sequence and record the total elapsed time. The total elapsed time can be denoted as switch_time(x,j), where x represents the xth task and j represents the transformable state.

[0114] Next, the current elapsed time for state transformation according to the preset queue operation rule and the preset state transformation rule during operation can be calculated. It should be noted that generally, at a certain moment, there is only one state that meets the preset switching rule in a certain preset message queue, and the current elapsed time is denoted as switch_time(x,current), where current is a pre-customized identification constant.

[0115] The state transformation benefit rate can be denoted as switch_help_rate. switch_time(x, current) is the current elapsed time calculated according to the set queue operation rules and preset state transformation rules. It is the sum of the total elapsed times ∑ that can be switched in the current state. j＝1 It is the average elapsed time obtained by dividing the sum of switch_time(x, j) by the total number of states count(j) that can be transformed. It should be noted that the smaller the value of this state transformation benefit rate, the better, indicating that the state transformation saves more time.

[0116] In some alternative embodiments, when calculating the comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate, as Figure 6 shown, Figure 6 This is a flowchart of the steps for calculating the comprehensive state change benefit coefficient provided by the embodiment of the present application, including:

[0117] Step 602: Obtain the first preset weight corresponding to the state transformation failure rate and the second preset weight corresponding to the state transformation benefit rate.

[0118] Step 604: Calculate the comprehensive state change benefit coefficient based on the state transformation failure rate, the state transformation benefit rate, the first preset weight, and the second preset weight.

[0119] Among them, the first preset weight corresponding to the state transformation failure rate can be denoted as switch_err_w, and the second preset weight corresponding to the state transformation benefit rate can be denoted as switch_help_w. Thus, the comprehensive state change benefit coefficient can be calculated based on the state transformation failure rate, the state transformation benefit rate, the first preset weight, and the second preset weight. The above first preset weight and second preset weight can be customarily set in advance.

[0120] Among them, the comprehensive state change benefit coefficient can be denoted as task_score.

[0121]

[0122] In some alternative embodiments, after obtaining the comprehensive state change benefit coefficient, the preset queue operation rules can be updated based on the comprehensive state change benefit coefficient, as Figure 7 shown, Figure 7 This is a flowchart of the steps for obtaining a new queue operation rule provided by the embodiment of the present application, including:

[0123] Step 702: Match the comprehensive state change benefit coefficient with a preset benefit threshold to obtain a matching result.

[0124] Step 704: Update the preset queue operation rules based on the matching result to obtain new queue operation rules.

[0125] Among them, for each task, a preset beneficial threshold can be set in advance, denoted as task_score_0, and the preset beneficial threshold for task x is task_score_0(x). All the set preset beneficial thresholds task_score_0 should be a sequence, which includes the preset beneficial thresholds of all tasks.

[0126] Thus, the comprehensive state change beneficial coefficient can be matched with the preset beneficial threshold to obtain a matching result, and then the preset queue operation rules can be updated based on the matching result to obtain new queue operation rules.

[0127] Optionally, for a certain task x, the matching degree between the comprehensive state change beneficial coefficient task_score(x) of task x and its preset beneficial threshold task_score_0(x) can be calculated first. Optionally, a calculation mode of the matching degree can be: compare task_score(x) with its preset beneficial threshold task_score_0(x), and use a floating range. If the floating range is within 30%, it is considered an acceptable floating.

[0128] If the calculated comprehensive state change beneficial coefficient is unacceptable, that is, the floating range of the matching result of the matching process with the preset beneficial threshold exceeds 30%, then update the preset queue operation rules to obtain new queue operation rules.

[0129] Thus, according to the above new queue operation rules, multiple currently registered artificial intelligence agents are incorporated into each preset message queue, and the initialized preset message queue is entrusted to a preset state scheduler for scheduling. The preset state scheduler uses preset state transformation rules for state scheduling transformation. Optionally, the preset state transformation rules can adopt the first-come-first-served rule, the elevator scheduling rule, the shortest path rule, etc.

[0130] Optionally, according to the real-time progress of each task, the state scheduling in the corresponding preset message queue can be evoked according to the preset state transformation rule corresponding to the current task x. Exemplarily, for a certain task x, the initial state is the ready state. After the system loads the necessary hardware and software resources, it is about to change it to the waiting state. At this time, it needs to wait for its business task, such as the return information of other upstream tasks, and then perform the switching of the running state through the preset state transformation rule. By using the life cycle of each current artificial intelligence agent in combination with the preset message queue, when the number of current artificial intelligence agents or tasks is large, the target state can be retrieved without a full scan of the preset message queue, improving the performance of state scheduling.

[0131] Step 208: Synchronize the target life cycle and target state to other current AI agents related to the current AI agent.

[0132] Among them, the target life cycle and target state can also be synchronized to other relevant current AI agents and other monitoring points, ensuring that all scheduling transformations of tasks or states dependent on the current AI agent can obtain relevant information at any time. In some optional embodiments, as Figure 8 shown Figure 8 FIG. is a flowchart of steps for synchronizing the target life cycle and target state provided by an embodiment of the present application, including:

[0133] Step 802: For each current AI agent, identify the role information corresponding to the current AI agent.

[0134] Step 804: Based on the role information, the preset function description information corresponding to the current AI agent, and the preset total task information, obtain the execution sequence logic information and dependency relationship information of the current AI agent.

[0135] Step 806: Based on the execution sequence logic information and dependency relationship information, determine other current AI agents related to the current AI agent.

[0136] Step 808: Synchronize the target life cycle and target state to other current AI agents.

[0137] Among them, the specific determination process of the role information, the execution sequence logic information and the dependency relationship information of the current AI agent can refer to the above embodiments and will not be elaborated here. Thus, based on the execution sequence logic information and the dependency relationship information, other current AI agents related to the current AI agent can be determined, and finally the target life cycle and target state can be synchronized to other current AI agents.

[0138] The present application provides a method for managing multiple artificial intelligence agents. The method includes: obtaining multiple currently registered artificial intelligence agents; for each currently registered artificial intelligence agent, performing a life cycle transformation on the currently registered artificial intelligence agent based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the currently registered artificial intelligence agent; performing a state transformation on the currently registered artificial intelligence agent based on real-time state information and a preset state transformation rule to obtain a target state corresponding to the currently registered artificial intelligence agent; and synchronizing the target life cycle and the target state to other currently registered artificial intelligence agents related to the currently registered artificial intelligence agent. By managing the life cycles and states of multiple currently registered artificial intelligence agents, the present solution makes the survival and change management of each currently registered artificial intelligence agent more refined, thereby making subsequent state monitoring and scheduling more efficient. In addition, by synchronizing the target life cycle and the target state to other currently registered artificial intelligence agents related to the currently registered artificial intelligence agent, the communication ability between each currently registered artificial intelligence agent is enhanced, and the multiple currently registered artificial intelligence agents can cooperate more effectively, improving the coordination and efficiency of the cooperation of multiple currently registered artificial intelligence agents.

[0139] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0140] Figure 9 It is a structural block diagram of a device for managing multiple artificial intelligence agents provided by an embodiment of the present application.

[0141] As Figure 9 shown, the device 900 for managing multiple artificial intelligence agents includes:

[0142] An obtaining module 902, configured to obtain multiple currently registered artificial intelligence agents.

[0143] A first transformation module 904, configured to perform a life cycle transformation on each currently registered artificial intelligence agent based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the currently registered artificial intelligence agent.

[0144] A second transformation module 906, configured to perform a state transformation on the current artificial intelligence agent based on real-time state information and a preset state transformation rule, so as to obtain a target state corresponding to the current artificial intelligence agent.

[0145] A synchronization module 908, configured to synchronize the target lifecycle and the target state to other current artificial intelligence agents related to the current artificial intelligence agent.

[0146] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here. Each module in the above multi-artificial intelligence agent management device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or stored in a memory in the computer device in the form of software, so that the processor can call and execute the operations of the above modules.

[0147] In an embodiment of the present application, a computer device is provided. The computer device includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0148] Obtain multiple registered current artificial intelligence agents;

[0149] For each current artificial intelligence agent, perform a lifecycle transformation on the current artificial intelligence agent based on a preset lifecycle transformation rule, so as to obtain a target lifecycle corresponding to the current artificial intelligence agent;

[0150] Perform a state transformation on the current artificial intelligence agent based on real-time state information and a preset state transformation rule, so as to obtain a target state corresponding to the current artificial intelligence agent;

[0151] Synchronize the target lifecycle and the target state to other current artificial intelligence agents related to the current artificial intelligence agent.

[0152] In an embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0153] Obtain multiple stages of a preset lifecycle and preset lifecycle transformation rules corresponding to each stage;

[0154] Determine the current stage in the lifecycle corresponding to the current artificial intelligence agent;

[0155] Perform a lifecycle transformation on the current artificial intelligence agent based on the preset lifecycle transformation rule corresponding to the current stage, so as to obtain a target lifecycle corresponding to the current artificial intelligence agent.

[0156] In one embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0157] Obtain the preset message queues corresponding to the preset states of the current artificial intelligence agent; wherein, the preset message queues include the real-time status information of the current artificial intelligence agent;

[0158] Obtain the real-time status information in the preset message queue;

[0159] Perform state transformation on the current artificial intelligence agent based on the preset queue operation rules, the real-time status information, and the preset state transformation rules to obtain the target state corresponding to the current artificial intelligence agent.

[0160] In one embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0161] Obtain the state transformation failure rate and the state transformation benefit rate; wherein, the state transformation failure rate is used to represent the probability of state switching failure, and the state transformation benefit rate is used to represent the efficiency of state switching;

[0162] Calculate the comprehensive state change benefit coefficient based on the state transformation failure rate and the state transformation benefit rate;

[0163] Update the preset queue operation rules based on the comprehensive state change benefit coefficient to obtain the new queue operation rules.

[0164] In one embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0165] Obtain the first preset weight corresponding to the state transformation failure rate and the second preset weight corresponding to the state transformation benefit rate;

[0166] Calculate the comprehensive state change benefit coefficient based on the state transformation failure rate, the state transformation benefit rate, the first preset weight, and the second preset weight.

[0167] In one embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0168] Match the comprehensive state change benefit coefficient with the preset benefit threshold to obtain a matching result;

[0169] Update the preset queue operation rules based on the matching result to obtain the new queue operation rules.

[0170] In one embodiment of the present application, when the processor executes the computer program, the following steps are further implemented:

[0171] For each current artificial intelligence agent, identify the role information corresponding to the current artificial intelligence agent;

[0172] Based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information, obtain the sequence logic information and dependency information for the execution of the current artificial intelligence agent;

[0173] Based on the sequence logic information and dependency information, determine other current artificial intelligence agents related to the current artificial intelligence agent;

[0174] Synchronize the target life cycle and target state to other current artificial intelligence agents.

[0175] The computer device provided by the embodiments of the present application has a similar implementation principle and technical effect to the above method embodiments, and will not be elaborated here.

[0176] In an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0177] Obtain multiple currently registered artificial intelligence agents;

[0178] For each current artificial intelligence agent, perform a life cycle transformation on the current artificial intelligence agent based on a preset life cycle transformation rule to obtain the target life cycle corresponding to the current artificial intelligence agent;

[0179] Perform a state transformation on the current artificial intelligence agent based on real-time state information and a preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent;

[0180] Synchronize the target life cycle and target state to other current artificial intelligence agents related to the current artificial intelligence agent.

[0181] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0182] Obtain a plurality of stages of a pre-defined set of life cycles and preset life cycle transformation rules corresponding to each stage;

[0183] Determine the current stage in the life cycle corresponding to the current artificial intelligence agent;

[0184] Perform a life cycle transformation on the current artificial intelligence agent based on the preset life cycle transformation rule corresponding to the current stage to obtain the target life cycle corresponding to the current artificial intelligence agent.

[0185] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0186] Obtain preset message queues corresponding to each preset state of the current artificial intelligence agent; wherein, the preset message queues include real-time state information of the current artificial intelligence agent;

[0187] Obtain the real-time state information in the preset message queue;

[0188] Perform state transformation on the current artificial intelligence agent based on the preset queue operation rules, real-time state information, and preset state transformation rules to obtain the target state corresponding to the current artificial intelligence agent.

[0189] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0190] Obtain the state transformation failure rate and the state transformation beneficial rate; wherein, the state transformation failure rate is used to characterize the probability of state switching failure, and the state transformation beneficial rate is used to characterize the efficiency of state switching;

[0191] Calculate the comprehensive state change beneficial coefficient based on the state transformation failure rate and the state transformation beneficial rate;

[0192] Update the preset queue operation rules based on the comprehensive state change beneficial coefficient to obtain new queue operation rules.

[0193] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0194] Obtain the first preset weight corresponding to the state transformation failure rate and the second preset weight corresponding to the state transformation beneficial rate;

[0195] Calculate the comprehensive state change beneficial coefficient based on the state transformation failure rate, the state transformation beneficial rate, the first preset weight, and the second preset weight.

[0196] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0197] Match the comprehensive state change beneficial coefficient with a preset beneficial threshold to obtain a matching result;

[0198] Update the preset queue operation rules based on the matching result to obtain new queue operation rules.

[0199] In an embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0200] For each current artificial intelligence agent, identify the role information corresponding to the current artificial intelligence agent;

[0201] Based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information, obtain the sequence logic information and dependency information for the current artificial intelligence agent to execute.

[0202] Based on the sequence logic information and dependency information, determine other current artificial intelligence agents related to the current artificial intelligence agent.

[0203] Synchronize the target life cycle and target state to other current artificial intelligence agents.

[0204] The computer-readable storage medium provided in this embodiment has the same implementation principle and technical effects as the above method embodiment, and will not be elaborated here.

[0205] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. This computer program can be stored in a non-volatile computer-readable storage medium. When this computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in this application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0206] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0207] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A method for managing multiple artificial intelligence agents, characterized in that: The method comprises: Get multiple current AI agents that have been registered; For each of the current artificial intelligence agents, the life cycle of the current artificial intelligence agent is transformed based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the current artificial intelligence agent; Based on the real-time state information and the preset state transformation rules, the state of the current artificial intelligence agent is transformed to obtain the target state corresponding to the current artificial intelligence agent; Synchronize the target lifecycle and the target state to other current artificial intelligence agents related to the current artificial intelligence agent.

2. The method according to claim 1, characterized in that The performing lifecycle transformation on the current artificial intelligence agent based on the preset lifecycle transformation rule to obtain the target lifecycle corresponding to the current artificial intelligence agent includes: Obtaining a pre-defined set of multiple life cycle stages and preset life cycle transformation rules corresponding to each of the stages; Determining a current stage in a life cycle corresponding to the current artificial intelligence agent; The life cycle of the current artificial intelligence agent is transformed based on the preset life cycle transformation rules corresponding to the current stage to obtain the target life cycle corresponding to the current artificial intelligence agent.

3. The method according to claim 1 or 2, characterized in that: The state transformation of the current artificial intelligence agent based on the real-time state information and the preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent includes: Obtaining a preset message queue corresponding to each preset state of the current artificial intelligence agent; wherein the preset message queue includes real-time state information of the current artificial intelligence agent; Obtaining real-time status information in the preset message queue; Based on the preset queue operation rules, the real-time status information and the preset status change rules, the status of the current artificial intelligence agent is changed to obtain the target state corresponding to the current artificial intelligence agent.

4. The method according to claim 3, characterized in that: The method further comprises: Acquire a state change failure rate and a state change benefit rate; wherein the state change failure rate is used to characterize the probability of state switching failure, and the state change benefit rate is used to characterize the efficiency of state switching; Calculating a comprehensive state change benefit coefficient based on the state change failure rate and the state change benefit rate; The preset queue operation rule is updated based on the comprehensive state change benefit coefficient to obtain a new queue operation rule.

5. The method according to claim 4, characterized in that The calculating of the comprehensive state change benefit coefficient based on the state change failure rate and the state change benefit rate comprises: Obtaining a first preset weight corresponding to the state change failure rate and a second preset weight corresponding to the state change benefit rate; The comprehensive state change benefit coefficient is calculated based on the state change failure rate, the state change benefit rate, the first preset weight and the second preset weight.

6. The method according to claim 4 or 5, characterized in that: The updating of the preset queue operation rule based on the comprehensive state change beneficial coefficient to obtain a new queue operation rule includes: Matching the comprehensive state change beneficial coefficient with a preset beneficial threshold to obtain a matching result; The preset queue operation rule is updated based on the matching result to obtain the new queue operation rule.

7. The method according to claim 1 or 2, characterized in that The step of synchronizing the target lifecycle and the target state to other current artificial intelligence agents related to the current artificial intelligence agent includes: For each of the current artificial intelligence agents, identifying role information corresponding to the current artificial intelligence agent; Based on the role information, the preset function description information corresponding to the current artificial intelligence agent, and the preset total task information, the sequential logic information and dependency relationship information executed by the current artificial intelligence agent are obtained; Based on the sequential logic information and the dependency information, determine and obtain other current artificial intelligence agents related to the current artificial intelligence agent; Synchronize the target lifecycle and the target state to the other current artificial intelligence agents.

8. A multi-artificial intelligence agent management device, characterized in that: The device comprises: The acquisition module is used to obtain multiple currently registered artificial intelligence agents; A first transformation module is used to transform the life cycle of each current artificial intelligence agent based on a preset life cycle transformation rule to obtain a target life cycle corresponding to the current artificial intelligence agent; A second transformation module is used to transform the state of the current artificial intelligence agent based on the real-time state information and the preset state transformation rule to obtain the target state corresponding to the current artificial intelligence agent; A synchronization module is used to synchronize the target life cycle and the target state to other current artificial intelligence agents related to the current artificial intelligence agent.

9. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the multi-artificial intelligence agent management method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the multi-artificial intelligence agent management method as described in any one of claims 1-7.

Images