KR102261793B1 - System for rpa robot agent clutstering - Google Patents

Abstract

The present invention relates to an RPA robot agent clustering system. According to the present invention, the RPA robot agent clustering system includes: an input unit for receiving work schedule information; a memory which stores a program performing clustering for an RPA robot agent using job schedule information; and a processor which executes the program. The processor allocates a plurality of different tasks to a cluster composed of the plurality of robot agents when a random waiting time is required because it is impossible to predict RPA task duration, or when the execution times of the plurality of RPA tasks overlap. Therefore, it is possible to improve the productivity of script code writing, and an RPA-related script writing work is possible on a web without installing a studio.

Description

RPA Robot Agent Clustering System {SYSTEM FOR RPA ROBOT AGENT CLUTSTERING}

The present invention relates to an RPA robot agent clustering system.

RPA (Robotic Process Automation) refers to applying software to repetitive and simple business processes during business processes and automating them.

According to the prior art, a technology for standardizing and logically automatically performing data generated in the business process has been proposed. However, in the case of a work script written only in text, certain parts on the screen are intuitively clicked, double clicked, and typed. There is a problem that it is difficult to determine whether it is the target of (Type).

According to the prior art, when a large amount of robot agent operation is required, the required time of the RPA task cannot be predicted, so a random waiting time is required for the next task assignment, or the time when a plurality of RPA tasks should be executed overlaps. There is a problem.

The present invention has been proposed to solve the above problems, and an object of the present invention is to provide an RPA robot agent clustering system capable of zero-wait processing for multiple tasks.

The RPA robot agent clustering system according to the present invention includes an input unit for receiving task schedule information, a memory in which a program for clustering the RPA robot agent using the task schedule information is stored, and a processor for executing the program, the processor comprising: When a random waiting time is required because it is impossible to predict the RPA task duration, or when the execution times of a plurality of RPA tasks overlap, a plurality of different tasks are allocated to a cluster composed of a plurality of robot agents.

The processor allocates tasks in cluster units, and assigns tasks to available robot agents with no work in progress among a plurality of robot agents belonging to the cluster.

The processor considers the capability information of each robot agent, grasps the progress of the current task, and selects available robot agents to assign the next task.

The robot agent includes an Image Capture Module (ICM), a Script Runtime Engine (SRE), an Image Detect Module (IDM), and an Image Processing Module (IPM).

According to the present invention, by providing a system and method for providing a web-based integrated development environment for RPA script development, images targeted for click, double click, and type are displayed in an integrated development environment (IDE). ) It is possible to improve the productivity of script code writing by supporting visual confirmation along with the code implementing the logic on the editor, and it is possible to write RPA-related scripts on the web without installing a studio.

In addition, by tying multiple agents into a single cluster so that multiple tasks are assigned, problems such as delay and error occurrence of subsequent tasks due to the performance of the PC running the agent (processing speed and system resources) or errors in previous tasks It is possible to solve the problem, and there is an effect that it is possible to manage Saas-based RPA robot tasks without installing a separate studio.

Effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 shows a system for providing an integrated development environment according to an embodiment of the present invention.
Figure 2 shows before/after conversion for image display on a code mirror according to an embodiment of the present invention.
3 illustrates an image capture process of a web-based integrated development environment providing system for RPA script development according to an embodiment of the present invention.
4 shows an image capture process of a web-based integrated development environment providing system for RPA script development for performing capture image version management according to another embodiment of the present invention.
5 illustrates a method for providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention.
6 shows the problem of generating a waiting time for a job according to the prior art.
7 shows a zero-waits implementation using an RPA robot agent clustering system according to another embodiment of the present invention.
8 shows an operation structure of an RPA according to another embodiment of the present invention.
9 shows an RPA robot agent clustering system according to another embodiment of the present invention.
10 and 11 show the cluster operation structure of the RPA robot agent clustering system according to another embodiment of the present invention.
12 shows a message queue operation structure according to another embodiment of the present invention.
13 illustrates a Publisher/Subscribe pattern according to another embodiment of the present invention.
14 shows a Task Queue pattern according to another embodiment of the present invention.

The above and other objects, advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings.

However, the present invention is not limited to the embodiments disclosed below, but can be implemented in various different forms, and only the following examples are provided to those of ordinary skill in the art to which the present invention pertains to the purpose of the invention, It is only provided to easily inform the configuration and effect, and the scope of the present invention is defined by the description of the claims.

On the other hand, the terms used herein are for the purpose of describing the embodiments and are not intended to limit the present invention. As used herein, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, "comprises" and/or "comprising" means that a referenced element, step, operation and/or element is the presence of one or more other elements, steps, operations and/or elements. or added.

1 shows a system for providing an integrated development environment according to an embodiment of the present invention.

Studio 110 is a working script development environment integrated with the administrator web dashboard.

The robot agent 120 drives the robot agent application program.

The robot agent 120 provides a communication channel for receiving information (including source code) related to the work created in the studio 110, and a web socket channel in which the robot agent 120 uses the server and the studio 110 as clients. make up

The robot agent 120 receives the commands related to the work written in the studio 110, and provides a communication channel for returning the result, and the message between the channels has the form of a JSON format {event, payload}.

The studio 110 creates/edits a working source code script.

The studio 110 calls the work script stored in the local repository (local repository, 140) by requesting the robot agent 120 running locally (open command) (open a project).

The studio 110 transmits the edited work source code by requesting the robot agent 120 running locally (save command) (send the project to save).

The robot agent 120 stores the work source code delivered from the studio 110 in an arbitrary local repository 140 (save the project to the storage).

The robot agent 120 executes the work script delivered from the studio 110 (exec the project), and tests it.

The robot agent 120 transmits the execution log and result (Log) of the work script in the studio 110 to the studio 110 .

The robot agent 120 builds a deployment module using the job-related information transmitted from the studio 110 (build the project to create a deploy module).

The robot agent 120 delivers the generated deployment module and job information to the backend server 150 (deploy the module).

The backend server 150 stores the delivered distribution module and job information in a database, and registers a new job.

According to an embodiment of the present invention, when writing a work script in the integrated development environment, images are inserted by using CodeMirror, and images that are the target of click, double click, and type are displayed. By supporting the visual confirmation along with the code implementing the logic on the integrated development environment (IDE) editor, the productivity of script code writing is improved.

According to an embodiment of the present invention, it is possible to utilize an open source project capable of editing program source code on a web page.

According to an embodiment of the present invention, by utilizing the code mirror custom language definition function (eg, CodeMirror.defineMode('python') // definition of grammar interpretation for python language), the final result of the source code displayed on the screen is displayed in HTML expressed in documents.

According to an embodiment of the present invention, when a capture function is requested from the robot agent at the point in time when capture is required in the integrated development environment, the robot agent, a native application program that can access all screens of the desktop PC, uploads the captured image to the server. After that, an accessible URL is created for the uploaded image, and the generated URL is converted into an <img> tag in the code on the integrated development environment.

By adding an onClick event handler as the <img> tag attribute, it can be used to set the image click location, and the <img> tag added with the code is displayed as an image on the code mirror.

Fig. 2 shows before (Fig. 2(a)) and after (Fig. 2(b)) conversion for image display on a code mirror according to an embodiment of the present invention.

3 shows an image capture flow of a web-based integrated development environment providing system for RPA script development according to an embodiment of the present invention.

The processes of (1) and (8), which will be described later, are performed through the WebSocket channel, and the process of (6) is performed through the REST API channel.

(1) In the integrated development environment, the studio 110 selects a screen capture menu to request a screen capture function.

(2) The robot agent 120 calls the function of the screen capture module 130 .

(3) The screen capture module 130 captures by dragging a certain part of the desktop PC.

(4) The screen capture module 130 stores the captured image in the local repository 140 .

(5) The screen capture module 130 returns the path stored in the local repository 140 to the robot agent 120 .

(6) The robot agent 120 uploads the image file to the backend server 150 .

(7) The backend server 150 stores the received image in the server repository 160 and returns the URL to the robot agent 120 .

(8) The robot agent 120 transmits the returned image URL to the studio 110 .

(9) The studio 110 of the integrated development environment uses the code mirror custom language redefinition function to search the script code for a pattern to be converted into an image, and then <img> tag with the URL returned from the robot agent 120 convert to

4 shows an image capture flow of a web-based integrated development environment providing system for RPA script development for performing captured image version management according to another embodiment of the present invention.

The processes of (1) and (11), which will be described later, are performed through the websocket channel, and the process of (9) is performed through the REST API channel.

(1) In the integrated development environment, the studio 110 selects a screen capture menu to request a screen capture function.

(2) The robot agent 120 calls the function of the screen capture module 130 .

(3) The screen capture module 130 captures by dragging a certain part of the desktop PC.

(4) The screen capture module 130 stores the captured image in the local repository 140 .

(5) The screen capture module 130 sends a COMMIT to the local Git repository, and (6), (7) sends a PUSH to the remote Git repository.

(8) The screen capture module 130 returns the path stored in the local repository 140 to the robot agent 120 .

(9) The robot agent 120 uploads an image file to the backend server 150 .

(10) The backend server 150 stores the received image in the server repository 160 and returns the URL to the robot agent 120 .

(11) The robot agent 120 transmits the returned image URL to the studio 110 .

(12) The studio 110 of the integrated development environment uses the code mirror custom language redefinition function to search the script code for a pattern to be converted into an image, and then <img> tag with the URL returned from the robot agent 120 convert to

5 illustrates a method for providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention.

The method of providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention includes the steps of calling a screen capture module function to capture a PC screen (S510), and storing the captured image in a local repository, , returning the path stored in the local repository (S520), uploading the image file to the backend server, storing it in the server repository, and passing the returned image URL (S530) and converting it into an image in the script code It includes a step (S540) of searching for a pattern to be used and converting it into a tag together with the returned image URL.

In step S510, the robot agent that has received the screen capture function request from the studio of the integrated development environment calls the function of the screen capture module and captures it by dragging a certain part of the PC.

In step S520, the screen capture module stores the captured image in the local repository, and transmits the PUSH to the remote repository and stores it.

In step S540, the studio uses the code mirror custom language redefinition function to search the script code for a pattern to be converted into an image, and then converts it into an <img> tag along with the URL returned from the robot agent.

Hereinafter, an RPA robot agent clustering system and a method according to another embodiment of the present invention will be described.

6 shows the problem of generating a waiting time for a job according to the prior art.

When the maximum concurrent processing task of a single agent is 1, there is a problem that the waiting time between the jobs cannot be 0 in a situation where the first to third jobs are scheduled.

7 shows a zero-waits implementation using an RPA robot agent clustering system according to another embodiment of the present invention.

Referring to FIG. 7 , the maximum concurrent processing task of the agent cluster 700 is 3, and the first agent 710 , the second agent 720 , and the third agent 730 assign an assignment to concurrently process the job execution. receive

8 shows an operation structure of an RPA according to another embodiment of the present invention.

The web management UI of the RPA client and the RPA robot agent communicate through a websocket.

The web management UI includes statistics monitor, task management, project management, integrated development environment, shared information management, agent cluster management, log monitor, and system management.

The web management UI is a web application that runs in a web browser environment, and includes the following main functions.

Statistical Monitor (Dashboard) provides statistical graphs and charts for job execution results, and provides a status monitoring view of RPA robot agents and agent clusters.

Job management provides functions of registration, inquiry, modification, and deletion of RPA jobs.

Project management provides functions of registration, inquiry, modification, and deletion of RPA development projects.

The integrated development environment provides the functions of editing and running test scripts for task automation belonging to the RPA project.

Shared information management provides registration, inquiry, modification, and deletion functions for sharing information that multiple RPA scripts should share and sensitive information that should not be displayed directly in the script (eg, other system accounts, passwords, etc.).

Agent cluster management improves the execution efficiency of RPA automation tasks by tying multiple RPA robot agents into one cluster, and provides agent list inquiry, selection, cluster registration, inquiry, modification, and deletion functions.

When explaining the cluster execution model, when there is a lot of work processing data, it is not allocated to one robot agent that processes it, but is allocated to a plurality of agents that process the same task by dividing the data section.

The target work condition is a task that requires processing a large amount of data, for example, thousands or tens of thousands of Excel files, or Excel or DB data in which thousands or tens of thousands of data are input.

The log monitor provides a function to search the log transmitted from the robot agents as a result of the RPA operation, and provides a function to determine the status of RPA operation such as normal termination and failure.

System management provides user registration and inquiry functions.

The RPA robot agent performs native functions that cannot be performed with web functions according to the request of the web management UI.

The RPA robot agent executes the RPA task and sends the resulting logs to the server.

The RPA robot agent sends status information to the server.

The RPA robot agent is registered as a member of the cluster and handles tasks assigned to the cluster it belongs to.

The RPA robot agent stores and manages project sources in a local repository.

RPA robot agents include Image Capture Module (ICM), Script Runtime Engine (SRE), Image Detect Module (IDM), and Image Processing Module (IPM).

ICM is an image capture function for the entire desktop screen, a function to capture a partial area of the desktop by dragging the mouse, a function to save the captured image as an image file in PNG format, a function to create a file name without duplicates for the captured image, and C++ Provides the function to capture images at high speed using the written native module.

IDM uses a PNG format image to detect the same area as the input image on the full screen of the desktop, a function to compare images at high speed using a native module written in C++, and a similarity between 1 to 99% ( The higher the value, the higher the similarity), the function to detect, the function to provide a UI dialog for setting the similarity to the detection target image, mouse click position, etc., and the function to designate the mouse click position during RPA execution based on the detected image. to provide.

SRE provides functions to interpret and execute RPA scripts, mouse and keyboard input functions required for RPA operations, and various functions (eg, wait, click, paste, etc.) using detected images using IDM.

IPM is a high-speed native module written in C++ and provides image capture function, image comparison function, image file creation function, and multi-monitor support function.

The RPA backend API microservices in the RPA server include Realtime Communication API (RTC API), File System Management Interface (FSMI), Scheduler Interface (SCI), Scheduler Engine (SCE), DB Interface (DBI), and Message Queue Interface (MQI). Included.

RPA servers include File System Management (FSM) and Message Queue Service (MQS) microservices.

9 shows an RPA robot agent clustering system according to another embodiment of the present invention.

The RPA robot agent clustering system according to another embodiment of the present invention includes an input unit 910 for receiving work schedule information, a memory 920 storing a program for clustering the RPA robot agent using the work schedule information, and It includes a processor 930 for executing a program, and the processor 930 executes a plurality of different tasks when a random waiting time is required because it is impossible to predict the RPA task duration, or when the execution times of a plurality of RPA tasks overlap. It is assigned to a cluster consisting of multiple robot agents.

The processor 930 allocates tasks in cluster units, and assigns tasks to available robot agents having no ongoing task among a plurality of robot agents belonging to the cluster.

The processor 930 considers the capability information of each of the robot agents, grasps the progress of the work currently in progress, and selects available robot agents to assign tasks.

The robot agent includes an Image Capture Module (ICM), a Script Runtime Engine (SRE), an Image Detect Module (IDM), and an Image Processing Module (IPM).

10 and 11 show the cluster operation structure of the RPA robot agent clustering system according to another embodiment of the present invention.

According to another embodiment of the present invention, by allocating a plurality of different tasks that cannot predict the time taken to complete the task to a cluster consisting of a plurality of robot agents, the execution waiting time of the plurality of tasks is minimized, and ideally sets the waiting time to zero.

The target work conditions are those that require the operation of a large number of robot agents.

Hereinafter, a problem in the case where a random waiting time is required for the next task assignment because the required time of the RPA task cannot be predicted will be described.

For example, if task A executed at 17:00 takes 1 hour, but since the execution time of task B cannot be set to 18:01 afterwards, 30 minutes of free time is added and set to 18:30 , there is a problem in that the RPA robot agent's work gap of up to 30 minutes occurs, and the efficiency of the robot agent decreases.

In addition, if the time that a plurality of RPA tasks should be executed overlaps, for example, if the estimated duration of task A to be executed at 18:00 is 1 hour and task B is to start at 18:30, before task A ends, B Unable to start work.

In this case, it is possible to set the robot agents per task one-to-one, but the number of robot agents increases, and it becomes difficult to grasp the task assignment status of each robot agent, which causes problems and errors that complicate the task assignment process of the manager. There is a problem.

According to another embodiment of the present invention, a plurality of robot agents are assigned to one cluster, and tasks are distributed in cluster units so that available robot agents with no ongoing tasks among the robot agents belonging to the cluster are automatically assigned tasks. It is possible to process

12 shows a message queue operation structure according to another embodiment of the present invention.

Job Queue delivers job execution commands to individual agents through a message queue channel allocated to each cluster according to a schedule.

The log collection (Log Queue) delivers the output log of the job scripts executed by the agent to the backend server.

Hereinafter, [Table 1] is the type of work load.

Event name state Contents Schedule: callback Receive notification of job start from the scheduler according to the registered schedule projectexec start Notify that the agent has started working projectexec done Notify that the agent has successfully completed its task projectexec error Notify the agent that an error occurred while executing the job

A shared information change notification (Config Queue) delivers changes to RPA job scripts or other global information to each agent.

Hereinafter, a RabbitMQ processing method for each message type according to another embodiment of the present invention will be described.

Hereinafter, P stands for Publisher, X stands for Exchange, and C stands for Consumer.

13 illustrates a Publisher/Subscribe pattern according to another embodiment of the present invention.

When explaining the application message of the Publisher/Subscribe pattern, Config broadcasts and transmits server environment change information to all agents. Log sends the execution result log to the server, and can also be received by a separate monitoring program. In the case of parallel execution of a job, a message is sent to the exchange designated for each cluster.

When explaining the characteristics of the Publisher/Subscribe pattern, when the consumer process is in the terminated state, the transmitted message is lost.

14 shows a Task Queue pattern according to another embodiment of the present invention.

When explaining the application message of the task queue pattern, the job transmits the job execution message to the job queue set for each cluster according to the job schedule registered in the server.

If we explain the characteristics of the task queue pattern, even if the consumer process is in the terminated state, unreceived messages can be received upon re-execution.

Meanwhile, the method for providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention may be implemented in a computer system or recorded in a recording medium. The computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and storage. Each of the above-described components performs data communication through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in a memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the method for providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention may be implemented as a computer-executable method. When the method for providing a web-based integrated development environment for RPA script development according to an embodiment of the present invention is performed in a computer device, computer-readable commands provide a web-based integrated development environment for RPA script development according to the present invention method can be performed.

On the other hand, the method for providing a web-based integrated development environment for RPA script development according to the present invention described above may be implemented as a computer-readable code on a computer-readable recording medium. The computer-readable recording medium includes any type of recording medium in which data that can be read by a computer system is stored. For example, there may be a read only memory (ROM), a random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be stored and executed as code readable in a distributed manner by being distributed in a computer system connected through a computer communication network.

Claims (8)

an input unit for receiving work schedule information;
a memory storing a program for clustering the RPA robot agent using the job schedule information; and
a processor for executing the program;
The processor allocates a plurality of different tasks to a cluster consisting of a plurality of robot agents when a random waiting time is required because it is impossible to predict the RPA task duration or when the execution times of a plurality of RPA tasks overlap,
The RPA robot agent has an image capture function for the entire desktop screen, a desktop partial area capture function through mouse drag, a function to save the captured image as a PNG format image file, a function to create a file name without duplicates for the captured image , which includes an Image Capture Module (ICM) that provides the ability to capture images at high speed using native modules written in C++.
An RPA Robot Agent Clustering System.
According to claim 1,
The processor distributes tasks in units of the cluster, and assigns tasks to available robot agents that do not have a work in progress among the plurality of robot agents belonging to the cluster
An RPA Robot Agent Clustering System.
According to claim 1,
The processor considers the capability information of each of the robot agents, grasps the progress of the currently ongoing task, and selects available robot agents to assign the next task
An RPA Robot Agent Clustering System.
delete delete an input unit for receiving work schedule information;
a memory storing a program for clustering the RPA robot agent using the job schedule information; and
a processor for executing the program;
The processor allocates a plurality of different tasks to a cluster consisting of a plurality of robot agents when a random waiting time is required because it is impossible to predict the RPA task duration or when the execution times of a plurality of RPA tasks overlap,
The RPA robot agent has a function to detect the same area as the input image on the entire desktop screen using a PNG format image, a function to compare images at high speed using a native module written in C++, and a function to detect the similarity of the comparison image. Function, similarity to detection target image, UI dialog window for setting mouse click position, etc., including IDM (Image Detect Module) that provides a function to specify the mouse click position during RPA execution based on the detected image. that
An RPA Robot Agent Clustering System.
an input unit for receiving work schedule information;
a memory storing a program for clustering the RPA robot agent using the job schedule information; and
a processor for executing the program;
The processor allocates a plurality of different tasks to a cluster consisting of a plurality of robot agents when a random waiting time is required because it is impossible to predict the RPA task duration or when the execution times of a plurality of RPA tasks overlap,
The RPA robot agent interprets and executes RPA scripts, mouse and keyboard input functions required for RPA work, and SRE (SRE) that provides wait, click, and paste functions using images detected using IDM (Image Detect Module). Script Runtime Engine)
An RPA Robot Agent Clustering System.
an input unit for receiving work schedule information;
a memory storing a program for clustering the RPA robot agent using the job schedule information; and
a processor for executing the program;
The processor allocates a plurality of different tasks to a cluster consisting of a plurality of robot agents when a random waiting time is required because it is impossible to predict the RPA task duration or when the execution times of a plurality of RPA tasks overlap,
The RPA robot agent is a high-speed native module written in C++, and includes an Image Processing Module (IPM) that provides an image capture function, an image comparison function, an image file creation function, and a multi-monitor support function.
An RPA Robot Agent Clustering System.

Images