JP2017191373A - Multi-agent system, task assignment device, task assignment method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a multi-agent system that can appropriately execute task assignment even when communication between respective agents becomes unstable.SOLUTION: A multi-agent system of the present embodiment comprises a task assignment unit and a task prediction unit. The task assignment unit selects a first agent to which a task can be assigned from a self agent and other agent on the basis of a recognition result of the task, and performs assignment of the task to the first agent when communication with the first agent is possible. The task prediction unit predicts a second agent that has a possibility of performing the assignment of the task to the first agent from the other agent, when the communication with the first agent is impossible.SELECTED DRAWING: Figure 1

Description

  Embodiments described herein relate generally to a multi-agent system, a task assignment device applied to the system, a task assignment method, and a program.

  In recent years, the use of a multi-agent system that communicates with each other between a plurality of agents via a network and performs various tasks (processes) in cooperation has attracted attention. For example, in a control system for controlling a plurality of drones (drone: unmanned aerial vehicles), a multi-agent system can be applied to realize a cooperation function for controlling each drone in cooperation.

  In general, an agent senses a situation (environment), recognizes a task to be executed based on the sensing information, and executes the task using a resource (computer resource) by autonomous determination.

  Here, the multi-agent system includes a task allocation (distribution) function that allocates (distributes), for example, a task for controlling each drone to each agent by communication between the agents. As a method for realizing the task assignment function, there is a method (contract net protocol) in which each agent assigns a task to an appropriate agent by exchanging information using a communication function via a network.

JP 2007-52683 A JP 2002-228396A International Publication WO2006 / 118193 JP 2009-55401 A

  In the multi-agent system, a task assignment function is realized by communication between agents. When communication between agents becomes unstable, there may be a state where communication cannot be temporarily performed, or a time required for communication increases and communication efficiency decreases.

  In such a case, task assignment (task distribution) cannot be executed or a certain time (valid time) such as a situation where a plurality of agents are assigned to one task in duplicate or a situation where an agent responsible for the task cannot be identified. ) May not be possible.

  Thus, there is a problem of realizing a multi-agent system capable of appropriately executing task assignment even when communication between agents becomes unstable.

  In the present embodiment, in a multi-agent system in which a plurality of agents are connected to a network for communication, each agent includes a task recognition unit, a task allocation unit, and a prediction unit. The task recognition means recognizes a task to be executed. The task assigning means selects a first agent that can assign the task from its own and other agents based on the recognition result of the task recognizing means, and can communicate with the first agent. , Assigning the task to the first agent. When the communication with the first agent is impossible, the predicting unit predicts a second agent having a high possibility of assigning the task to the first agent from another agent. .

The block diagram for demonstrating the structure of the multi agent system regarding embodiment. The flowchart for demonstrating operation | movement of the multi-agent system regarding embodiment. The figure which shows an example of the network map information regarding embodiment. The figure which shows an example of the other agent task recognition performance information regarding embodiment. The figure which shows an example of the other agent task allocation rule regarding embodiment. The schematic diagram for demonstrating an example of the task allocation process regarding embodiment. The schematic diagram for demonstrating an example of the task allocation process regarding embodiment.

Embodiments will be described below with reference to the drawings.
[System configuration]
FIG. 1 is a diagram illustrating a configuration of a multi-agent system according to the present embodiment. In this system, each of the agents 10-1 to 10-N communicates with each other in a plurality of computers 1-1 to 1-N connected to the network 2 and includes a task allocation (task distribution) function to be described later. is there.

  The multi-agent system according to the present embodiment is applied to, for example, a control system that controls a plurality of drones, and each agent 10-1 to 10-N controls each drone in cooperation with each other. The agents 10-1 to 10-N may be represented as agents A, B, C, D, etc. for convenience. That is, when the agent 10-1 corresponds to the agent A (parent agent), a plurality of other agents (other agents) 10-N correspond to the agents B, C, D, and the like.

  In the present embodiment, as shown in FIG. 1, the agents 10-1 to 10-N basically have the same internal configuration. Hereinafter, the internal configuration will be described with the agent 10-1 as a representative, and description of the internal configurations of other agents will be omitted.

  The agent 10-1 is configured to have software modules and databases (information) 15, 16, 18-20 for realizing the functions 11-14, 17 respectively. The agent 10-1 includes a task allocation unit 11, a task recognition unit 12, a sensing unit 13, a task execution unit 14, and another agent task allocation prediction unit (hereinafter referred to as a task allocation prediction unit) 17.

  The task allocation unit 11 refers to the task allocation rule database (hereinafter referred to as DB) 15 and the other agent task allocation DB 16 as the task to be executed (referred to as task X for convenience) and the other agent 10 − A task assignment process for assigning (distributing) task X to N is executed.

  The task assignment rule DB 15 is information indicating a task assignment rule (one of the task assignment rules described in FIG. 5) when the agent 10-1 (agent A) executes the task assignment process. The other agent task assignment DB 16 is history information when performing task assignment processing based on a so-called contract net protocol. This history information is information related to task assignment processing when information is exchanged with other agents by communication.

  The task recognition unit 12 recognizes the task X to be executed based on the sensing information of the sensing unit 13 and notifies the task allocation unit 11 of it. The sensing unit 13 senses (detects) a situation (environment) when a multi-agent system is applied, for example, when a plurality of drones are controlled in cooperation. As a specific example, the sensing unit 13 communicates with, for example, the radar 30-1 and executes predetermined detection processing (for example, flight status such as positions of a plurality of drones) according to the detection signal from the radar 30-1. To do. The task recognition unit 12 recognizes, for example, a task X for controlling a designated drone based on the sensing information.

  The task execution unit 14 executes a task assigned by the task assignment unit 11. Here, the agent 10-1 (agent A) uses the task assignment unit 11 to execute a task assigned to the agent A or a task assigned from the other agent 10-N. As a specific example, in the drone control system, the task execution unit 14 communicates with a controller 31-1 that controls each drone, and causes the controller 31-1 to execute an assigned task (for example, a control process). The controller 31-1 may be an actuator of a device to be controlled.

  The task assignment prediction unit 17 cooperates with the task assignment unit 11 to predict an agent that may assign the task X to the corresponding agent (here, B) from the other agent 10-N. The agent B is an agent to which the task X is assigned by the normal task assignment processing of the task assignment unit 11 and cannot communicate with the agent A thereafter.

As will be described later, the task allocation prediction unit 17 refers to the network map DB 18, the other agent task recognition performance DB 19, and the other agent task allocation rule DB 20, and executes a prediction process. Each agent 10-1 to 10-N does not need to share the other agent task assignment rule DB 20 if the task assignment rules in the task assignment rule DB 15 have the same contents.
[Task assignment process]
FIG. 2 is a flowchart for explaining the task assignment processing of this embodiment. With reference to the flowchart of FIG. 2, the task assignment processing of this embodiment will be described.

  In the present embodiment, for the sake of convenience, the agent 10-1 is assumed to be agent A, and the other agent 10-N is assumed to be agents B, C, D, and the like. Agent A is a parent agent that assigns tasks to other agents B, C, D, and the like.

  As shown in FIG. 2, in the agent A, the task recognition unit 12 recognizes the necessity of executing the task X (in other words, the task X to be executed) based on the sensing information of the sensing unit 13 (step S1). . Specifically, the sensing information is information indicating the flight status such as the positions of a plurality of drones detected according to the detection signal from the radar 30-1.

  When the task allocation unit 11 refers to the other agent task allocation DB 16 and recognizes that the task X allocation processing has already started in the other agents B, C, D, etc., the task allocation processing ends. (YES in step S2). Here, the agent A holds the history information when the other agents B, C, D, etc. execute the task assignment processing by exchanging information in the communication between agents as the other agent task assignment DB 16. That is, the history information is information used when performing task assignment processing based on a so-called contract net protocol.

  If the task X allocation process has not been started by another agent (NO in step S2), the task allocation unit 11 may be responsible for task X with respect to the other agents B, C, D, etc. by communication between the agents. (In other words, the possibility of executing task X) is inquired (step S3). When there is no response from any agent for this inquiry (inquiry) (NO in step S4), the task assignment unit 11 ends the task assignment processing for the other agents B, C, D, etc. The task X is executed by its own task execution unit 14. Note that the task assignment unit 11 may execute other processing than that executed by its own task execution unit 14.

  On the other hand, when there is another agent that responds to the inquiry within a certain time (valid time) (YES in step S4), the task allocation unit 11 refers to the task allocation rule DB 15 and sets a predetermined task allocation rule. Based on this, the agent B determined to be optimal as the allocation target is selected (step S5). A specific example is a task assignment rule as shown in FIG.

  If the task allocation unit 11 can communicate with the selected agent B within a predetermined time, that is, if communication is maintained (YES in step S6), the task allocation unit 11 notifies the agent B of task X allocation. (Step S7). As a result, the agent B causes the task execution unit 14 to execute the task X assigned by the agent A. As a specific example, for example, the agent B communicates with the controller 31-N that controls the drone, and causes the assigned task (for example, control process) X to be executed.

  Here, the task allocation unit 11 cooperates with the task allocation prediction unit 17 when communication with the selected agent B is impossible within a predetermined time, that is, when communication is unstable (NO in step S6). The task assignment prediction unit 17 refers to the network map DB 18, the other agent task recognition performance DB 19, and the other agent task assignment rule DB 20, and performs a prediction process for predicting whether or not the other agent assigns the task X to the agent B. Execute (step S8).

  Specifically, the task allocation prediction unit 17 compares the multiplication result of the probability of each condition with the reference value based on the network map information, the other agent task recognition performance information, and the other agent task allocation rule, and determines the reference value. It is predicted that there is a high possibility that the exceeding agent C assigns task X to agent B. In short, the task assignment prediction unit 17 predicts that the possibility that the agent C assigns the task X to the agent B is high based on the probability information indicating the possibility of assigning the task.

  Here, the network map DB 18 includes network map information as shown in FIG. 3, and is specifically information on the communication state of each arc (communication connection between agents). In particular, the connection probability varies greatly when communication between agents is wireless communication. This network map information is acquired periodically by information exchange between agents. The other agent task recognition performance DB 19 includes other agent task recognition performance information as shown in FIG. This information is information that depends on the sensing performance, and is based on, for example, the installation position or detection distance of the radar 30-N. The other agent task assignment rule DB 20 includes task assignment rules as shown in FIG. If the task assignment rules in the task assignment rule DB 15 of each agent have the same contents, it is not necessary to share the other agent task assignment rule DB 20.

  Returning to FIG. 2, the task assignment prediction unit 17 predicts that the agent C communicates with the agent B within a predetermined time and assigns the task X to the agent B (YES in step S9). The task allocation unit 11 is notified. Thereby, the task assignment unit 11 ends the task assignment process.

  If the task allocation prediction unit 17 cannot predict an agent that has a high possibility of assigning the task X to the agent B (NO in step S9), the task allocation unit 11 can again communicate within a certain time other than the agent B. Select other agents. That is, the task allocation unit 11 refers to the task allocation rule DB 15 and selects the agent D determined to be optimal as an allocation target based on a preset task allocation rule (step S10).

  If the task allocation unit 11 can communicate with the selected agent D within a certain period of time, that is, if communication is maintained, the task allocation unit 11 notifies the agent D of task X allocation (step S1). S11). Thereby, the agent D causes the task execution unit 14 to execute the task X assigned by the agent A.

  With regard to the task assignment processing of the present embodiment as described above, the operational effects will be described with reference to the schematic diagrams of FIGS.

  First, after selecting agent B as an assignment target of task X, if agent A can communicate between agents A and B, agent A executes task X assignment processing on agent B and ends. Here, as shown in FIG. 6, when a state (60) in which communication between the agents A and B cannot be performed occurs, the task allocation prediction unit 17 causes the agent C to communicate with the agent B within a certain time (valid time). (61) When it is predicted that the task X is likely to be assigned, the task assignment unit 11 ends without executing anything. Here, the agent A cannot communicate with the predicted agent C.

  That is, when the communication between the agents A and B is unstable, if the agent A predicts the agent C which is highly likely to assign the task X to the agent B, the task assignment unit 11 terminates without executing anything. It becomes.

  Therefore, it is possible to avoid a situation where a plurality of agents are in charge of the same task X and the computer resources are wasted. In addition, if a plurality of agents are responsible for the same task X, each agent cannot allocate resources to other task executions, but such a situation can be avoided. As a result, task allocation (task distribution) can be appropriately executed even when communication between agents becomes unstable.

  Furthermore, as shown in FIG. 7, in a state where communication between agents B and C is not possible (62), when it is not possible to predict an agent C that has a high possibility of assigning task X to agent B within a certain time, the task assignment unit 11 Executes the process of selecting the agent D as the task X allocation target again. If communication between agents A and D is possible within a certain time (valid time), agent A executes task X assignment processing for agent D and ends. Therefore, even when the communication between the agents is unstable, it is possible to avoid the situation where the agent that can handle the task X cannot be specified.

  In the present embodiment, the task allocation prediction unit 17 and each DB 18-20 have been described with respect to the configuration provided in each agent. However, the configuration provided only in the agent A may be used.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

1-1 to 1-N ... computer, 2 ... network, 10-1 to 10-N ... agent,
11 ... task allocation unit, 12 ... task recognition unit, 13 ... sensing unit,
14 ... task execution unit, 15 ... task allocation rule database (DB),
16 ... Other agent task assignment DB, 17 ... Other agent task assignment prediction unit,
18 ... Network map DB, 19 ... Other agent task recognition performance DB,
20 ... Other agent task assignment rule DB, 30-1 to 30-N ... Radar,
31-1 to 31-N: Controller.

Claims (34)

In a multi-agent system in which multiple agents connect to a network and communicate,
Each of the agents
Task recognition means for recognizing tasks to be executed;
When the first agent capable of assigning the task is selected from its own and other agents based on the recognition result of the task recognizing means, and the communication with the first agent is possible, the first agent Task assignment means for executing the task assignment processing for
Predicting means for predicting a second agent having a high possibility of assigning the task to the first agent from another agent when communication with the first agent is impossible Multi-agent system to do. The task assignment means includes:
2. The multi-agent according to claim 1, wherein, when the second agent cannot be predicted by the prediction means, a third agent capable of communicating and allocating the task is selected from the other agent. system. The task assignment means includes:
2. The multi-agent system according to claim 1, wherein when the second agent is predicted by the prediction unit, the task assignment processing is terminated. The task assignment means includes:
Selecting the first agent from the agents that responded to the inquiry as to whether the task can be assigned by communicating with a plurality of other agents;
The multi-agent system according to any one of claims 1 to 3, wherein when there is no response to the inquiry, the task assignment process is terminated.   5. The prediction unit according to claim 1, wherein the prediction unit predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a communication connection probability between agents. The multi-agent system according to any one of the above.   5. The prediction unit according to claim 1, wherein the prediction unit predicts a second agent having a high possibility of assigning the task to the first agent within a predetermined time based on a communication speed between the agents. The multi-agent system according to claim 1.   5. The prediction unit according to claim 1, wherein the predicting unit predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a task recognition probability of another agent. The multi-agent system according to any one of the above.   5. The prediction unit according to claim 1, wherein the predicting unit predicts a second agent having a high possibility of assigning the task to the first agent based on a task assignment rule of another agent. Multi-agent system described in 1.   The prediction means predicts a second agent having a high possibility of assigning the task to the first agent based on a task assignment rule common to all agents. The multi-agent system described in the section. A task assignment device applied to a multi-agent system in which a plurality of agents connect to a network and communicate,
The task allocation device is provided in all of one agent or a plurality of agents,
Task recognition means for recognizing tasks to be executed;
When the first agent capable of assigning the task is selected from its own and other agents based on the recognition result of the task recognizing means, and the communication with the first agent is possible, the first agent Task assignment means for executing the task assignment processing for
Predicting means for predicting a second agent having a high possibility of assigning the task to the first agent from another agent when communication with the first agent is impossible Task assignment device to perform. The task assignment means includes:
11. The task assignment according to claim 10, wherein when the second agent cannot be predicted by the prediction means, a third agent capable of communicating and assigning the task is selected from the other agents. apparatus. The task assignment means includes:
11. The task assignment device according to claim 10, wherein when the second agent is predicted by the prediction means, the task assignment processing is terminated.   13. The prediction unit according to claim 10, wherein the prediction unit predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a communication connection probability between agents. The task assignment device according to any one of claims.   The prediction means predicts a second agent having a high possibility of assigning the task to the first agent within a predetermined time based on a communication speed between the agents. The task assignment device according to claim 1.   The prediction means predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a task recognition probability of another agent. The task assignment device according to any one of claims.   The said prediction means predicts the 2nd agent with high possibility of assigning the said task with respect to the said 1st agent based on the task allocation rule of another agent. The task assignment device described in 1.   The prediction means predicts a second agent having a high possibility of assigning the task to the first agent based on a task assignment rule common to all agents. The task assignment device according to item.   13. The prediction unit according to claim 10, wherein the prediction unit predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a communication connection probability between agents. The task assignment device according to any one of claims. A task assignment method applied to a multi-agent system in which a plurality of agents connect to a network and communicate,
Recognizing the task to be executed,
A process of selecting a first agent capable of assigning the task from the other agents based on the recognition result of the task;
A process of executing the task assignment process for the first agent when communication with the first agent is possible;
A process of predicting a second agent that has a high possibility of assigning the task to the first agent from another agent when communication with the first agent is impossible. Task assignment method.   The process according to claim 19, further comprising a process of selecting a third agent capable of communicating and allocating the task from the other agent when the second agent cannot be predicted by the predicting process. The task assignment method described.   The task assignment method according to claim 19, wherein when the second agent is predicted by the prediction process, the task assignment process is terminated.   The predicting process predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a communication connection probability between agents. The task assignment method according to any one of the above.   The predicting process predicts a second agent having a high possibility of assigning the task to the first agent within a certain period of time based on a communication speed between the agents. The task assignment method according to any one of the above items.   The predicting process predicts a second agent having a high possibility of assigning the task to the first agent based on probability information including a task recognition probability of another agent. The task assignment method according to any one of the above.   The one of claims 19 to 21, wherein the predicting process predicts a second agent having a high possibility of assigning the task to the first agent based on a task assignment rule of another agent. The task assignment method described in the section.   22. The second agent that is likely to assign the task to the first agent is predicted based on a task assignment rule common to all agents by the predicting process. The task allocation method according to item 1. In a computer incorporating a plurality of agents, applied to a multi-agent system in which a plurality of agents connect to a network and communicate,
Recognizing the task to be executed,
A process of selecting a first agent capable of assigning the task from the other agents based on the recognition result of the task;
A process of executing the task assignment process for the first agent when communication with the first agent is possible;
A process of predicting a second agent that is highly likely to assign the task to the first agent from another agent when communication with the first agent is impossible A program for causing the computer to execute.   28. The method according to claim 27, further comprising a process of selecting a third agent capable of communicating and assigning the task from the other agent when the second agent cannot be predicted by the predicting process. The listed program.   28. The program according to claim 27, wherein when the second agent is predicted by the prediction process, the task assignment process is terminated.   30. The second agent having a high possibility of assigning the task to the first agent is predicted based on probability information including a communication connection probability between the agents by the predicting process. The program according to any one of the above.   30. The second agent having a high possibility of allocating the task to the first agent within a predetermined time based on a communication speed between agents by the predicting process. The program according to any one of the above items.   30. The second agent having a high possibility of assigning the task to the first agent is predicted based on the probability information including the task recognition probability of another agent by the predicting process. The program according to any one of the above.   30. Any one of claims 27 to 29, wherein the predicting process predicts a second agent having a high possibility of assigning the task to the first agent based on a task assignment rule of another agent. The program described in the section.   30. The second agent having a high possibility of assigning the task to the first agent based on a task assignment rule common to all agents by the predicting process. The program according to item 1.