KR102153012B1 - Store management apparatus and store management method for recommending suitable employee for task through deep learning model based on big data - Google Patents

Abstract

The present invention relates to a store management apparatus for recommending an employee suitable for a task through a deep learning model based on big data and a store management method thereof. According to the present invention, the store management method comprises the following steps of: receiving store information; extracting a task list including information for a plurality of tasks to be managed in a store corresponding to the store information based on the inputted store information; assigning an employee to perform task information included in the extracted task list; receiving evaluation information for the task executed by the assigned employee; using the task information, employee information for the assigned employee, and the evaluation information as learning data for a predetermined learning machine algorithm to train a first neural network model for extracting task ability information when first task information is executed by an employee corresponding to first employee information; and linking with a manpower database storing employee information of each of a plurality of employees to extract an employee suitable for the predetermined task information based on task ability information outputted by inputting the employee information of the employee to the first neural network model with respect to predetermined task information.

Description

Store management device and store management method that recommends employees suitable for work through a big data-based deep learning model {STORE MANAGEMENT APPARATUS AND STORE MANAGEMENT METHOD FOR RECOMMENDING SUITABLE EMPLOYEE FOR TASK THROUGH DEEP LEARNING MODEL BASED ON BIG DATA}

The present invention relates to a store management apparatus and a store management method. Specifically, the present invention relates to a technology for recommending an employee suitable for performing each task to be performed in a store based on the business type information and employee information of the store through a big data-based deep learning model.

With the spread of various cultures in society due to the development of the Internet and economic development, stores offering a variety of products and services are emerging. In many cases, managers of these stores hire employees who can help with various tasks that occur in the store to conduct business.

As above, as the products and services provided by each store gradually diversify, it is becoming difficult to hire an employee who understands the work that each store must perform from the beginning, and the employee is a part-time student who takes on the job in the short term. It has a characteristic that replacement is frequent due to many cases.

Accordingly, it is very important in terms of store operation that store managers efficiently deliver store-related tasks to employees and manage whether employees are working properly.

In addition, the type of work that employees need to perform varies according to the business characteristics of the store, and the work efficiency of each employee varies depending on the gender and age of the employee at work. It is very important to allocate appropriate staff to perform the task.

For example, a manager who is starting store management for the first time may not know what tasks to instruct employees, and in this case, the store management may be properly managed because the tasks necessary for store management are not performed at the right time. Various problems can occur, such as confusion in performance or inconvenience to guests.

As another example, if the manager assigns tasks without considering the individual characteristics of each employee, the efficiency of each employee's work decreases, resulting in employee dissatisfaction, or the efficiency of store management, such as two employees doing their own work. Falling problems can occur.

Therefore, there is a need for a manager to recommend an appropriate employee to perform each task and the task details to be performed by the staff according to the business characteristics of the store.

An object of the present invention is to provide a store management device that recommends appropriate employees to perform each task based on the business type information and employee information of the store. In particular, it is an object of the present invention to provide a store management device for extracting appropriate employees to perform each task based on employee information of a store operated or operated by a manager and evaluation information on tasks performed by the employee. .

An object of the present invention is to provide a store management device that recommends an employee to perform each job based on evaluation information on the reassigned job when the employee's job is reassigned.

An object of the present invention is to provide a store management device that stores and manages employee capability information for each characteristic and job.

An object of the present invention is to provide a store management device that recommends an employee suitable for a store/task based on previously stored employee-specific business data when a manager wants to hire a new employee.

In the store management method of the store management apparatus according to an embodiment of the present invention, the step of receiving store information, extracting a business list including a plurality of business information to be managed in a store corresponding to the store information based on the input store information Steps to perform, assigning employees to perform work information included in the extracted work list, receiving evaluation information on tasks performed by assigned employees, work information, employee information and evaluation information on assigned employees Is used as learning data for a predetermined machine learning algorithm, and when the first job information and the first employee information are input, the first job information is performed by an employee corresponding to the first employee information. Learning a first neural network model for extracting, and interlocking with a manpower database storing employee information for each of a plurality of employees, inputting employee information on the employees for predetermined work information into the first neural network model and outputting Based on the obtained job capability information, it may include the step of extracting an employee suitable for the predetermined job information.

The step of allocating staff includes reassigning tasks performed by each of the employees whose evaluation of at least one item of the input evaluation information is less than or equal to a preset score, and the step of receiving evaluation information includes reassignment Including the step of receiving evaluation information on the tasks performed by the assigned employee, and the step of learning the first neural network model, the reassigned work information, the reassigned employee information, and the reassigned employee's evaluation information are used as training data. , Retraining the first neural network model.

The step of receiving evaluation information may include receiving at least one of task completion, integrity, responsibility, affinity, kindness, and overall score.

The first neural network model is composed of a neural network including an input layer, a hidden layer, and an output layer, is set to input job information and employee information to the input layer, and is set to input evaluation information to the output layer. And as the employee information is input in the learning process of the first neural network model, the correlation may be learned to derive the job capability information of the employee corresponding to the input job information.

In the step of extracting the work list, when store information including business type information and employee information is input, a second neural network learned through a predetermined machine learning algorithm to extract a plurality of business information to be managed at the store corresponding to the store information. It may include the step of extracting the task list through the model.

The second neural network model is composed of a neural network including an input layer, a hidden layer, and an output layer, and the actual store information including the actual business type information and the actual employee information from a plurality of manager terminals, and the actual store information is actually performed in the store. Based on the store-specific business DB that receives and stores data mapped with actual business information, the input layer is set to input actual store information including actual business type information and actual employee information, and the output layer is set to input actual store information. The actual business information generated by the manager managing the corresponding store is set to be input, and as a number of actual store information is input in the learning process of the second neural network model, the correlation is made to derive business information corresponding to the input store information. Can be learned.

The store-specific business DB further includes data on the number of times actual business items are reported per hour, and the second neural network model is set to input store information including actual business type information and actual employee information, and the actual store in the output layer It is set to input actual business information and data on the number of times generated by the manager who is managing the store according to the information, and as a number of real store information is input in the learning process of the neural network model, business information corresponding to the input store information and Correlation can be learned to derive the number of times a task should be reported per hour.

The step of assigning staff includes assigning staff to perform task information included in the extracted task list, and designating to report the result of the task in accordance with the number of times derived, and receiving evaluation information May include receiving evaluation information on the result of the reported task according to the number of times derived.

In the step of extracting the appropriate employee for the job information, the job data for each employee is mapped to each employee information by mapping the job competency score for each job output when a plurality of job information for each of the plurality of employees is input in the first neural network model. It may include the step of storing.

The step of extracting an employee suitable for the job information may include extracting, based on the job data for each employee, a job evaluated above a preset score among job competency scores as job information suitable for the personnel in the manpower database. .

In the step of extracting employees suitable for job information, when a hiring request for predetermined job information is input from the manager terminal, job capability data for each employee for predetermined job information among employees in the personnel database is mapped to a predetermined score or higher. It may include the step of extracting employees.

The employee information may include one or more of gender, age, language spoken, education, major, height, and weight.

The step of receiving the evaluation information may include receiving an image photographed by an employee assigned to the task information of a result of processing the task, and determining an evaluation of the task information based on the image.

In the store management apparatus according to an embodiment of the present invention, upon receiving store information, a business list extraction unit that extracts a business list including a plurality of business information to be managed in a store corresponding to the store information based on the input store information , Task assignment department that assigns employees to perform task information included in the extracted task list, task management department that receives evaluation information on tasks performed by assigned employees, task information, employee information and evaluation for assigned employees When information is used as learning data for a predetermined machine learning algorithm, when the first job information and the first employee information are input, the first job information is performed by an employee corresponding to the first employee information. In conjunction with a work capability information extraction unit including a first neural network model for extracting information, and a manpower database storing employee information for each of a plurality of employees, a first neural network for employee information on a person for predetermined work information It may include an employee recommendation unit for extracting an employee suitable for predetermined job information based on job capability information input and output in the model.

According to an embodiment of the present invention, since an employee suitable for each job is recommended based on the evaluation information on the job performed by the employees, the work efficiency for each job is increased, and the store management device for improving the job satisfaction of the employees, and There is an advantage that can provide a store management method.

In addition, since the staff is recommended based on the work evaluation information by the manager, there is an advantage that it is possible to provide a store management device and a store management method in which an employee suitable for manager preference is recommended.

In particular, since an employee is recommended based on job information reassigned by the manager, there is an advantage that it is possible to provide a store management device and a store management method in which the employee is recommended by reflecting the manager's employee preference for each job.

In addition, when the manager wants to hire a new employee, the manager is recommended based on the information the manager has evaluated on the work performed by the employee. There is an advantage that can provide a management method.

1 is a conceptual diagram of a store management system including a store management apparatus 100 and an employee terminal 200 according to an embodiment of the present invention.
2 is a functional block diagram of a store management apparatus 100 according to an embodiment of the present invention.
3 is a detailed configuration diagram of the task list extracting unit 112 shown in FIG. 2.
4 is an example of a machine learning-based neural network model that extracts actual work items from actual store information.
5 is a flowchart illustrating a process of a store management method through a store management device according to an embodiment of the present invention.
6 is a diagram illustrating a first example of information input or extracted into a store management apparatus according to the flowchart shown in FIG. 5.
FIG. 7 is a diagram illustrating a second example of information input or extracted from the store management apparatus according to the flowchart illustrated in FIG. 5.
8 is an exemplary diagram of a method for displaying an image capture guideline by the employee terminal 200 according to an embodiment of the present invention.
9 is a flowchart illustrating a method of recommending an employee suitable for predetermined work information by the store management apparatus according to an embodiment of the present invention.
10 is a detailed configuration diagram of the work capability information extracting unit 112 shown in FIG. 2.
11 is an example of a machine learning-based neural network model for extracting work capability information from work information, employee information, and evaluation information.
12 is an exemplary diagram of a method for displaying an employee extracted by a store management apparatus for predetermined job information according to an embodiment of the present invention.

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 together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various forms, and only these embodiments make the disclosure of the present invention complete, and those skilled in the art to which the present invention pertains. It is provided to fully inform the person of the scope of the invention, and the scope of the invention is only defined by the claims.

In describing the embodiments of the present invention, detailed descriptions of known functions or configurations will be omitted except when actually necessary in describing the embodiments of the present invention. In addition, terms to be described later are terms defined in consideration of functions in an embodiment of the present invention, which may vary according to the intention or custom of users or operators. Therefore, the definition should be made based on the contents throughout this specification.

The functional blocks shown in the drawings and described below are only examples of possible implementations. In other implementations, other functional blocks may be used without departing from the spirit and scope of the detailed description. Also, although one or more functional blocks of the present invention are represented as individual blocks, one or more of the functional blocks of the present invention may be a combination of various hardware and software configurations that perform the same function.

In addition, the expression of including certain constituent elements is an open expression and simply refers to the existence of the constituent elements, and should not be understood as excluding additional constituent elements.

Further, when a component is connected to or is referred to as being connected to another component, it should be understood that it may be directly connected or connected to the other component, but other components may exist in the middle.

In addition, expressions such as'first, second', etc. are used only for distinguishing a plurality of elements, and do not limit the order or other features between the elements.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

1 is a conceptual diagram of a store management system including a store management apparatus 100 and an employee terminal 200 according to an embodiment of the present invention.

Referring to FIG. 1, the store management apparatus 100 and the employee terminal 200 are connected to be operable through a network to transmit and receive information, thereby maximizing the efficiency of the store. For example, the network may include a wired or wireless communication network network such as a local area network (LAN), a wide area network (WAN), a virtual network, and remote communication.

The store management device 100 is a device used by a store manager. The store management device 100 may allocate information on a task to be performed in a store to a specific employee and transmit information on the assigned task to the employee terminal 200. In addition, the store management device 100 may receive the results of the work from the employee terminal 200, so that the manager can efficiently manage the store and manage the employees. The store management device 100 may receive evaluation information on the result of the work received from the employee terminal 200.

The employee terminal 200 is a device used by store employees. The employee terminal 200 may receive information on a task to be performed by the employee from the store management device 100. The employee terminal 200 outputs information on a task to be performed to an employee, and the employee terminal 200 may transmit a result (eg, an image file) of a task performed by the employee to the store management device 100.

The store management device 100 and the user terminal 200 may be implemented as various types of devices capable of transmitting and receiving information through a network. For example, it may include a portable communication device (eg, a smart phone), a computer device, a portable multimedia device, a notebook computer, a tablet PC, and the like.

2 is a functional block diagram of a store management apparatus 100 according to an embodiment of the present invention.

Referring to FIG. 2, the store management device 100 includes a store-specific business DB 111, a business list extracting unit 112, a business decision unit 113, a business assignment unit 114, a business management unit 115, and It may include at least some or all of the capability information extracting unit 116, the employee-specific work DB 117, the employee/task extracting unit 118, the camera 119, the display unit 120, and the processor 121. .

The processor 121 may control the overall operation of the store management apparatus 100. The processor 121 includes a store-specific business DB 111, a work list extraction unit 112, a work decision unit 113, a work assignment unit 114, a work management unit 115, a work capability information extraction unit 116, Each of the employee-specific task DB 117, the employee/task extraction unit 118, the camera 119, and the display unit 120 may be controlled.

The store-specific business DB 111 may receive and store data that maps the actual store information including actual business type information and actual employee information from a plurality of manager terminals, and the actual business items actually being performed in the store of the actual store information. have.

In this case, the plurality of manager terminals may mean a plurality of store management apparatuses 100 used by a plurality of managers.

The actual business type information may mean information on the type of product or service provided by the store. For example, the actual business type information may be a hot pot, a pizzeria, a PC cafe, a bowling alley, a sushi restaurant, laundry, academy, beauty, etc. .

The actual employee information may refer to information on employees working in a real store, and for example, the actual employee information may include the number, gender, and age of employees.

Meanwhile, the actual store information may further include information on a region where the store is located in addition to the above-described actual business type information and actual employee information.

The actual business items may include business items that are actually being performed in a store corresponding to the actual store information. For example, actual work items may include exhaust management, gas management, meat wrapping, food ingredients management, store environment management, security/security, sales management, order management, and the like.

Actual work items may be items input through the store management device 100. Actual business items may differ according to actual store information.

The task list extraction unit 112 may extract a task list including task items to be performed in a store corresponding to the actual store information.

Upon receiving the store information, the job list extracting unit 112 may extract a job list including a plurality of job information to be managed in a store corresponding to the store information based on the input store information.

The task list extractor 112 may automatically extract a task list through a neural network model. To this end, when the store information including business type information and employee information is input, the job list extraction unit 112 extracts a job list including a plurality of business items to be managed in a store corresponding to the store information. It may include a neural network model learned through a learning algorithm.

3 is a detailed configuration diagram of the task list extracting unit 112 shown in FIG. 2.

Referring to FIG. 3, the work list extraction unit 112 may include at least some or all of a store information collection unit 1121, an actual work item collection unit 1122, and a model generation unit 1123.

The store information collection unit 1121 collects the actual store information input through the store management device 100, and the actual work item collection unit 1122 collects the actual business items input through the store management device 100. I can.

When the model generation unit 1123 acquires the actual store information collected by the store information collection unit 1121 and the actual business item collected by the real business item collection unit 1122, the actual business item by reflecting the actual store information Actual business items may be extracted by inputting actual store information into a neural network model learned through a predetermined machine learning algorithm that extracts. The model generation unit 1123 may generate such a neural network model through the following method.

4 is an example of a machine learning-based neural network model that extracts actual work items from actual store information.

Machine Learning (Machine Learning) refers to the field of defining various problems dealt with in the field of artificial intelligence and studying methodologies to solve them. Machine learning is also defined as an algorithm that improves the performance of a task through continuous experience.

An artificial neural network (ANN) is a model used in machine learning, and may refer to an overall model with problem-solving capabilities, which is composed of artificial neurons (nodes) that form a network by combining synapses. The artificial neural network may be defined by a connection pattern between neurons of different layers, a learning process for updating model parameters, and an activation function for generating an output value.

The artificial neural network may include an input layer, an output layer, and optionally one or more hidden layers. Each layer includes one or more neurons, and the artificial neural network may include neurons and synapses connecting neurons. In an artificial neural network, each neuron can output a function value of an activation function for input signals, weights, and biases input through synapses.

Model parameters refer to parameters determined through learning, and include weights of synaptic connections and biases of neurons. In addition, hyperparameters refer to parameters that must be set before learning in a machine learning algorithm, and include a learning rate, iteration count, mini-batch size, and initialization function.

The purpose of learning artificial neural networks can be seen as determining model parameters that minimize the loss function. The loss function can be used as an index to determine an optimal model parameter in the learning process of the artificial neural network.

Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning according to the learning method.

Supervised learning refers to a method of training an artificial neural network when a label for training data is given, and a label indicates the correct answer (or result value) that the artificial neural network should infer when training data is input to the artificial neural network. It can mean. Unsupervised learning may refer to a method of training an artificial neural network in a state where a label for training data is not given. Reinforcement learning may mean a learning method in which an agent defined in a certain environment learns to select an action or action sequence that maximizes the cumulative reward in each state.

Among artificial neural networks, machine learning implemented as a deep neural network (DNN) including a plurality of hidden layers is sometimes called deep learning (deep learning), and deep learning is a part of machine learning. Hereinafter, machine learning may be used to include deep learning.

Referring to FIG. 4, the model generator 1123 may train a neural network model using a deep learning algorithm among machine learning algorithms according to an embodiment, and the neural network model according to an embodiment includes an input layer, a hidden layer, and It can be composed of a neural network including an output layer. The model generation unit 1123 sets actual store information including actual business type information and actual employee information as parameters to be input to the input layer, and outputs the actual work items generated by the manager managing the store according to the actual store information. It can be set as a parameter to be input to.

Thereafter, the model generation unit 1123 uses a plurality of data collected through actual store information and actual work items as learning data, or by using external data including variables for parameters to be illustrated below to the setting. We can train the neural network model accordingly. As a plurality of actual store information is input in the learning process of the neural network model, the correlation may be learned to derive a work item corresponding to the input store information. For example, learning data labeled with actual store information (ex. store information shown in FIG. 6) are input to the input layer, and work items corresponding to the actual store information (ex. exhaust management, gas management) are input to the output layer. , Meat wrapping, food ingredient management, etc.) can be set to input the labeled learning data.

In this way, by inputting a plurality of training data labeled with the above example information into the constructed neural network model according to the setting, the weights between neurons and the bias of neurons constituting the neural network model are updated in the direction of minimizing the loss function. , In the neural network model, a correlation for deriving actual business items reflecting the actual store information from the actual store information may be learned in connection information (ex. weights, biases) between hidden layers.

Accordingly, when actual store information for extracting actual business items, for example, information corresponding to the parameters of the input layer, is input into the neural network model, a list of business items including actual business items to be performed at the store corresponding to the actual store information Etc. can be output.

Meanwhile, the number of parameters and the number of hidden layers used for learning in FIG. 4 are examples only, and are not limited thereto. In addition to the deep learning algorithm, decision tree (DT), k-nearest neighbor (KNN), logistic regression (LR), and multilayer An appropriate algorithm can be selected and used according to the type or number of parameters to be used, such as perceptron (MLP), random forest (RF), support vector machine (SVM), and lasso.

Meanwhile, the store-specific business DB 111 may further include data on the number of times that actual business items are reported per hour.

In this case, the neural network model is set to input store information including actual business type information and actual employee information, and data on the actual work items and the number of times generated by the manager managing the store according to the actual store information are input to the output layer. It is set so that the correlation may be learned so as to derive the number of times that the business items corresponding to the input store information and the business items should be reported per hour as a plurality of actual store information is input in the learning process of the neural network model.

Again, Fig. 2 will be described.

The task determination unit 113 may receive an input for determining at least one of the work items included in the work list extracted through the work list extraction unit 112 as an actual work item from the manager. The task determination unit 113 may set only the task items determined through an administrator input among the task items included in the task list extracted through the task list extractor 112 as an actual task item.

In this way, the store management device 100 further includes an input module (not shown) for receiving various inputs from the manager, such as determining the actual work item for the work item included in the work list or manually inputting the work item. You may.

Meanwhile, according to an embodiment, the task determination unit 113 may manually input a task item that is not included in the extracted task list from the manager and add it to the actual task item.

In this case, the neural network model is retrained to additionally input data on work items manually input to the input layer and output layer, so that the correlation between the manager's store information and the work items manually inputted by the manager is updated. It can be re-learned whenever possible.

In this case, the task list extracted through the task list extracting unit 112 can be supplemented by the administrator, and since such data is re-learned by the neural network model, the need for the administrator to separately supplement it in the future is reduced. There is an advantage that can provide convenience.

The task assignment unit 114 may allocate an employee to perform task information included in the extracted task list.

The task assignment unit 114 may allocate an employee to perform each task of the actual work item determined by the task determination unit 113 based on the employee information.

According to an embodiment, the task assignment unit 114 may sequentially allocate employees included in the employee information to each of the actual work items determined by the work determination unit 113.

According to another embodiment, the task assignment unit 114 may allocate an employee to perform a task to each actual task item determined based on the task DB 111 for each store. Specifically, when the data on which the first employee information and the first work item are mapped to the business DB 111 for each store is more than the data on which the second employee information and the first work item are mapped, the An employee with the largest amount of information corresponding to the first employee information may be assigned as an employee who will perform the first work item determined as the work item. To this end, the task assignment unit 114 is composed of a neural network including an input layer, a hidden layer, and an output layer, and is set to input actual work items to the input layer by using the data of the store-specific work DB 111, The output layer contains the neural network model in which the correlation between the work item and the employee information is learned by inputting the actual employee information performing the actual work item, and the actual work item determined by inputting the actual work item determined in the neural network model Based on the employee information output for each, it is possible to assign an employee to perform the job for each of the determined actual job items.

The task management unit 115 may transmit and receive signals with the employee terminal 200. The task management unit 115 may include a short-range communication module, a mobile communication module, and the like. The task management unit 115 may transmit each of the determined actual task items to the assigned employee's terminal 200.

The task management unit 115 may receive an image file including a task execution status from the employee's terminal 200 and determine whether the task is completed based on the image file.

As an example, the task management unit 115 determines the time when the image file is created based on the metadata included in the image file, and if the time at which the image file is created is included in the time to perform the task, It can be judged as completed.

Meanwhile, when transmitting each of the determined actual work items to the assigned employee's terminal, the work management unit 115 may further transmit an image shooting guideline according to the image composition taken by the manager.

The task management unit 115 may receive evaluation information on tasks performed by assigned employees.

When specific task information and specific employee information are input, the work capability information extracting unit 116 may extract work capability information when the inputted work is performed by the inputted employee. To this end, the work capability information extraction unit 116 uses work information, employee information on employees assigned to each job, and evaluation information on tasks performed by each employee as learning data for a predetermined machine learning algorithm, When the first job information and the first employee information are input, a neural network model for extracting job capability information when an employee corresponding to the first employee information performs the first job information may be included.

The employee-specific job DB 117 may store job data for each employee to which job competency scores are mapped for each of a plurality of employee information.

The employee/task extraction unit 118 may extract an employee suitable for each job or extract a job suitable for each employee.

The employee/task extracting unit 118 interlocks with a manpower database that stores employee information for each of a plurality of employees, and inputs employee information on a plurality of employees included in the manpower database for predetermined work information into a neural network model. Based on the outputted job capability information, it is possible to extract employees suitable for predetermined job information.

The camera 119 may take an image for a clear task instruction when transmitting a task instruction to the employee terminal 200 of an employee who will perform each actual task item.

The display unit 120 may display a screen for receiving store information input, a list of tasks extracted according to the store information, and employee information assigned to each task item. That is, the display unit 120 may display various types of information necessary for store management.

Next, FIG. 5 is a flowchart showing a process of a store management method through the store management device according to an embodiment of the present invention, and FIG. 6 is information input or extracted into the store management device according to the flowchart shown in FIG. Is a diagram illustrating a first example of, and FIG. 7 is a diagram illustrating a second example of information input or extracted into a store management apparatus according to the flowchart shown in FIG. 5.

The store management device 100 may receive store information (S1).

Referring to the examples of FIGS. 6 and 7, the store management apparatus 100 may receive a hot pot as business type information, and may receive employee information for eight employees as employee information.

The store management device 100 may extract a work list (S3).

According to an embodiment, as illustrated in FIG. 6, the task list extraction unit 112 includes exhaust management, gas management, meat wrapping, food material management, store environment management, security/security, according to the store information input in step S1. You can extract a list of tasks including sales management and order management.

According to another embodiment, when extracting the task list, the task list extractor 112 may extract the number of times each task item should be reported to a plurality of task items together. As illustrated in FIG. 7, the task list extraction unit 112 includes exhaust management, gas management, meat wrapping, food material management, store environment management, expense/security, sales management, order management, etc. according to the store information input in step S1. Exhaust management reports 5 times a day, gas management reports 3 times a day, meat wrapping reports 3 times a day, food ingredients management reports 3 times a day, store environment management reports once a day, security/security You can extract the number of times that business items such as reporting twice a day, sales management reporting once a day, and order management reporting once a day must be reported.

The store management device 100 may determine a task list (S3).

6 and 7, the task determination unit 113 performs at least one task in the task list, for example, exhaust management, gas management, meat wrapping, food material management, and store environment management as a task item to be actually performed. You can be decided.

The store management device 100 may allocate an employee to each work item included in the determined work list (S4).

6 and 7, the task assignment unit 114 allocates employee 1 as an employee to perform exhaust management, assigns employee 2 as an employee to perform gas management, and an employee to perform meat wrapping As a result, employees 3 and 4 can be assigned, employees 5 and 6 can be assigned as employees to manage food ingredients, and employees 7 and 8 can be assigned as employees to manage the store environment.

On the other hand, the task assignment unit 114 reflects such employee information when exhaust management is mainly performed by men in their 20s and gas management is mainly performed by women in their 20s according to the data stored in the store-specific business DB 111. You can also assign employees to each task item.

The store management device 100 may perform a work report and work check (S5).

6 and 7, the work management unit 115 may transmit the contents of work items assigned to each employee to each employee terminal 200. In this case, the task management unit 115 may further transmit an image file captured by the camera 119. In particular, the task management unit 115 may further transmit an image shooting guideline according to the composition of the image captured by the manager. The employee terminal 200 may also display an image capture guideline. In this case, the manager takes an image to suggest a composition, such as what composition should be taken when the employee reports a job, and the employee refers to the image taken by the manager or the guidelines for shooting images, and in what composition. You know if you need to shoot.

8 is an exemplary diagram of a method for displaying an image capture guideline by the employee terminal 200 according to an embodiment of the present invention.

The store management apparatus 100 may generate an image capture guideline based on an image captured by a manager, and transmit an image and an image capture guideline to the employee terminal 200. As illustrated in FIG. 8, the employee terminal 200 may display a guideline in which an image to be transmitted by an employee should be included and an image shooting guideline including information on text describing the guideline. For example, when exemplifying that the work item is cleaning the front door, since the image to be transmitted by the employee is the front door, it may include information on a guideline (green dotted line) about the range in which the front door should be photographed as shown in FIG. , As a description of the guideline, text information such as "Please take a picture of the front door along the green line" may be included. Accordingly, job information including a guideline and text may be displayed on the employee terminal 200, and the employee may generate an image file by referring to the guideline and text.

Accordingly, in step S5, the store management device 100 determines whether or not the target image included in the image file has been photographed according to the guideline, or the work included in the image file inside the guideline through a predetermined machine learning algorithm. By analyzing the target image, you can judge that the task is complete.

In addition, the store management device 100 may receive evaluation information in step S5. The evaluation information may mean an evaluation result input by a manager for each task performed by each employee. The store management apparatus 100 may receive evaluation information by receiving at least one of job completion, sincerity, responsibility, affinity, kindness, and overall score.

The store management device 100 may input an evaluation on a task performed by each employee. As an example, the manager may view an image captured in the manner as shown in FIG. 8 through the store management device 100 and input evaluation information on tasks performed by each employee. As another example, the store management apparatus 100 may receive an image in which an employee assigned to the job information captures a result of processing a job, and may determine an evaluation of the job information based on the image. However, the above-described examples are merely illustrative for convenience of description. That is, the store management apparatus 100 may input evaluation information on tasks performed by each employee in various ways.

On the other hand, when assigning employees, the processor 121 allocates employees to perform the work information included in the extracted work list, derives the number of times the work must be reported per hour, and processes the work according to the derived number You can specify to report results. In this case, the processor 121 may receive evaluation information on the result of processing the reported task according to the number of times derived. For example, the processor 121 may recognize the average value of evaluation information on the result of processing the tasks reported as many times as the derived number as final evaluation information on the tasks performed by the employee.

According to an embodiment of the present invention, the store management apparatus 100 inputs employee information for a plurality of personnel for predetermined job information into a neural network model, and is suitable for predetermined job information based on the output job capability information. Can recommend staff.

9 is a flowchart illustrating a method of recommending an employee suitable for predetermined work information by the store management apparatus according to an embodiment of the present invention.

The processor 121 learns a neural network model for extracting job capability information when an employee corresponding to the predetermined employee information performs predetermined job information by using job information, employee information, and employee evaluation information as learning data. Can do it (S11).

The task information may mean information on each task item extracted by the task list extraction unit 112 or information on each task item determined by the task determination unit 113. According to the example of FIG. 6, the business information may refer to pear management, gas management, meat wrapping, food material management, and store environment management.

Employee information may include one or more of gender, age, language spoken, education, major, height, and weight.

The evaluation information may mean a result of each employee acquired in step S5 of FIG. 5 performed on the assigned task.

The processor 121 may learn a neural network model for extracting job competency information when an employee corresponding to the predetermined employee information performs predetermined job information through the job competency information extracting unit 112, and for this It will be described in detail with reference to FIG. 10.

10 is a detailed configuration diagram of the work capability information extracting unit 112 shown in FIG. 2.

Referring to FIG. 10, the work capability information extraction unit 112 includes at least some or all of the work information collection unit 1161, the employee information collection unit 1162, the evaluation information collection unit 1163, and the model generation unit 1164 It may include.

The business information collection unit 1161 collects business information in the store input through the store management device 100, and the employee information collection unit 1162 performs each job input through the store management device 100. Information of employees is collected, and the evaluation information collection unit 1163 may collect evaluation information input for each employee who has performed each task. The evaluation information collection unit 1163 may collect evaluation information input for a task performed by an employee assigned to each task in step S5 of FIG. 5.

The model generation unit 1164 includes the work information collected by the work information collection unit 1161, the employee information collected by the employee information collection unit 1162, and the evaluation information collection unit 1164 When the evaluation information of the employees who performed each task is acquired, the work information and employee information are applied to the neural network model learned through a predetermined machine learning algorithm that extracts the work capability information of each employee by reflecting the work information and employee information. By inputting, it is possible to extract the work capability information of the employee corresponding to the input work information. The model generator 1164 may generate such a neural network model through the following method.

11 is an example of a machine learning-based neural network model for extracting work capability information from work information, employee information, and evaluation information.

Machine Learning (Machine Learning) refers to the field of defining various problems dealt with in the field of artificial intelligence and studying methodologies to solve them. Machine learning is also defined as an algorithm that improves the performance of a task through continuous experience.

An artificial neural network (ANN) is a model used in machine learning, and may refer to an overall model with problem-solving capabilities, which is composed of artificial neurons (nodes) that form a network by combining synapses. The artificial neural network may be defined by a connection pattern between neurons of different layers, a learning process for updating model parameters, and an activation function for generating an output value.

The artificial neural network may include an input layer, an output layer, and optionally one or more hidden layers. Each layer includes one or more neurons, and the artificial neural network may include neurons and synapses connecting neurons. In an artificial neural network, each neuron can output a function value of an activation function for input signals, weights, and biases input through synapses.

Model parameters refer to parameters determined through learning, and include weights of synaptic connections and biases of neurons. In addition, hyperparameters refer to parameters that must be set before learning in a machine learning algorithm, and include a learning rate, iteration count, mini-batch size, and initialization function.

The purpose of learning artificial neural networks can be seen as determining model parameters that minimize the loss function. The loss function can be used as an index to determine an optimal model parameter in the learning process of the artificial neural network.

Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning according to the learning method.

Supervised learning refers to a method of training an artificial neural network when a label for training data is given, and a label indicates the correct answer (or result value) that the artificial neural network should infer when training data is input to the artificial neural network. It can mean. Unsupervised learning may refer to a method of training an artificial neural network in a state where a label for training data is not given. Reinforcement learning may mean a learning method in which an agent defined in a certain environment learns to select an action or action sequence that maximizes the cumulative reward in each state.

Among artificial neural networks, machine learning implemented as a deep neural network (DNN) including a plurality of hidden layers is sometimes called deep learning (deep learning), and deep learning is a part of machine learning. Hereinafter, machine learning may be used to include deep learning.

Referring to FIG. 11, the model generator 1164 may train a neural network model using a deep learning algorithm among machine learning algorithms according to an embodiment, and the neural network model according to an embodiment includes an input layer, a hidden layer, and It can be composed of a neural network including an output layer. The model generator 1164 may set work information and employee information as parameters to be input to the input layer, and set employee evaluation information corresponding to the input work information as parameters to be input to the output layer.

Thereafter, the model generation unit 1164 uses job information, employee information on employees assigned to each job, and evaluation information on jobs performed by each employee as learning data, or variables for parameters to be illustrated below. A neural network model may be trained according to the above setting by using the included external data. As job information and employee information on employees assigned to each job are input in the learning process of the neural network model, the correlation can be learned so that job capability information, which is the data that predicts the evaluation information of the employee on the input job, is predicted. . For example, in the input layer, learning data labeled with job information and employee information are input, and in the output layer, employee evaluation information for each job (ex. job completion, integrity, kindness, and responsibility) Learning data labeled with) can be set to be input.

In this way, by inputting a plurality of training data labeled with the above example information into the constructed neural network model according to the setting, the weights between neurons and the bias of neurons constituting the neural network model are updated in the direction of minimizing the loss function. , In the neural network model, a correlation for deriving actual business items reflecting the actual store information from the actual store information may be learned in connection information (ex. weights, biases) between hidden layers.

Accordingly, by inputting information corresponding to the parameters of the input layer in which the work information and employee information for extracting work capability information is input into the neural network model, the work capability information of the employees for each task, for example, work on the input work information When an employee corresponding to the inputted employee information performs it, a score for each expected job completion, integrity, kindness, and responsibility can be output.

Meanwhile, the number of parameters or hidden layers used for learning in FIG. 11 are examples only, and are not limited thereto, and in addition to the deep learning algorithm, decision tree (DT), k-nearest neighbor (KNN), logistic regression (LR), and multilayer An appropriate algorithm can be selected and used according to the type or number of parameters to be used, such as perceptron (MLP), random forest (RF), support vector machine (SVM), and lasso.

The processor 121 may extract the work capability information of each employee through the neural network model learned in step S11 (S13).

The processor 121 may store work data for each employee (S15).

For each employee's job data, each employee's employee information and various types of job information included in the manpower database including employee information for a plurality of employees are inputted into the neural network model of FIG. It may be data mapped to employee information.

For example, the processor 121 uses five kinds of work information (first to fifth tasks) and five employees, and extracts each of the work information and employee information through the neural network model of FIG. It is possible to store job data for each employee representing job competency information (ex. average of scores for each of job completion, integrity, kindness, and responsibility output in the example of FIG. 11), which may be as shown in Table 1 below. , This is only exemplary.

Employee's ability to work in the first task Employee's job ability for the second job Employee's work ability for the third task Employee's job ability for the fourth job Employee's work ability for the fifth task Employee information 1 1/10 5/10 4/10 9/10 7/10 Employee information 2 8/10 8/10 10/10 3/10 8/10 Employee information 3 4/10 10/10 6/10 1/10 1/10 Employee information 4 9/10 9/10 2/10 3/10 2/10 Employee information 5 5/10 5/10 7/10 5/10 9/10

The processor 121 may determine whether the work of the employees has been reassigned (S17).

The processor 121 may reassign work of employees based on the evaluation information input by the manager. Specifically, the processor 121 may reassign tasks performed by each of the employees whose evaluation of at least one item of evaluation information input by the manager is less than or equal to a preset score.

When the work of the employees is reassigned, the processor 121 may retrain the neural network model by using the reassigned work information, the reassigned employee information, and the reassigned employee's evaluation information as training data (S19).

When the work of the employees is reassigned, the processor 121 may receive evaluation information on the work performed by the reassigned employee.

The processor 121 may retrain the neural network model by using the reassigned work information, the reassigned employee information, and the reassigned employee's evaluation information as training data.

When the work of the employees is not reassigned, the processor 121 may extract work information suitable for each employee based on work data for each employee (S21).

As an example, the processor 121 provides job information suitable for the personnel of the employee database based on the job data for each employee, and the job evaluated as a preset score or higher (ex. 7 points or more out of 10 points) among the job competency points for each job. It is possible to control the employee/task extraction unit 118 to extract into. Referring to the example in Table 1 above, the employee task extraction unit 118 allocates an employee corresponding to employee information 2 or 4 to the first task, and responds to the employee information 2, 3 or 4 to the second task. Assign employees, assign employees corresponding to employee information 2 or 5 to the third job, assign employees corresponding to employee information 1 to the fourth job, and respond to employee information 1, 2 or 5 to the fifth job You can allocate the staff.

As another example, the processor 121 may control the employee/task extraction unit 118 to extract the employee with the highest job ability score for each job as job information suitable for the job based on job data for each employee. have. Referring to the example in Table 1 above, the employee/task extraction unit 118 extracts the employee corresponding to employee information 4 as an employee suitable for job 1, and an employee responding to employee information 3 as an employee suitable for job 2 Extracts the employee corresponding to employee information 2 as an employee suitable for job 3, extracts the employee corresponding to employee information 1 as an employee suitable for job 5, and responds to employee information 5 as an employee suitable for job 6 Employees can be extracted.

The above examples are only examples for convenience of explanation, and the employee/task extraction unit 118 may extract work information suitable for each employee based on work data for each employee in various ways.

In step S21, the manpower database may mean a database of employees included in work data for each employee.

The processor 121 may determine whether a hiring request for predetermined job information has been input (S23).

In the case of hiring a new employee, the manager may input job information to be performed by the employee to be employed and a job request for the job through the store management device 100. The store management apparatus 100 may receive a hiring request for predetermined job information from a manager.

When a hiring request for predetermined job information is input, the processor 121 may extract an employee suitable for predetermined job information based on job data for each employee (S25).

When a hiring request for predetermined job information is input, the processor 121 may extract an employee whose job competency score for predetermined job information is mapped to a predetermined score or higher among employees of the external database.

In step S25, the manpower database may mean a database having information on job seekers registered on a predetermined recruitment site.

That is, when a hiring request for predetermined job information is input, the processor 121 extracts employee information mapped to a predetermined score or higher for predetermined job information from job data for each employee, and responds to the extracted employee information. The employees who respond to hiring requests can be extracted by extracting the employees who are working from an external database.

12 is an exemplary diagram of a method for displaying an employee extracted by a store management apparatus for predetermined job information according to an embodiment of the present invention.

When the hiring request for the first task, the second task, and the third task is input, the store management device 100 is an employee A corresponding to the employee information mapped to the task ability score higher than a preset score for the first task. Is extracted, and employee B corresponding to employee information mapped with a job competency score higher than a preset score for the second job, and employee information mapped with a job competency score higher than a preset score for the third job The corresponding employee C may be extracted, and the extracted employee may be displayed on the display unit 120. At this time, the processor 121 may control the display unit 120 to further display the degree of suitability for the employee extracted for each task, and the degree of suitability is the employee information of the employee extracted according to the job competency score for each employee or employee information It may be determined according to the matching rate for.

The embodiments of the present invention described above can be implemented through various means. For example, embodiments of the present invention may be implemented by hardware, firmware, software, or a combination thereof.

In the case of implementation by hardware, the method according to embodiments of the present invention includes one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), and Programmable Logic Devices (PLDs). , Field Programmable Gate Arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, and the like.

In the case of implementation by firmware or software, the method according to the embodiments of the present invention may be implemented in the form of a module, procedure, or function that performs the functions or operations described above. A computer program in which software codes and the like are recorded may be stored in a computer-readable recording medium or a memory unit and driven by a processor. The memory unit may be located inside or outside the processor, and may exchange data with the processor through various known means.

In addition, combinations of each block of the block diagram attached to the present invention and each step of the flowchart may be performed by computer program instructions. Since these computer program instructions can be mounted on the encoding processor of a general purpose computer, special purpose computer or other programmable data processing equipment, the instructions executed by the encoding processor of the computer or other programmable data processing equipment are each block of the block diagram or Each step of the flow chart will create a means to perform the functions described. These computer program instructions can also be stored in computer-usable or computer-readable memory that can be directed to a computer or other programmable data processing equipment to implement a function in a particular way. It is also possible to produce an article of manufacture in which the instructions stored in the block diagram contain instruction means that perform the functions described in each block of the block diagram or each step of the flow chart. Computer program instructions can also be mounted on a computer or other programmable data processing equipment, so that a series of operating steps are performed on a computer or other programmable data processing equipment to create a computer-executable process to create a computer or other programmable data processing equipment. It is also possible for the instructions to perform the processing equipment to provide steps for performing the functions described in each block of the block diagram and each step of the flowchart.

In addition, each block or each step may represent a module, segment, or part of code including one or more executable instructions for executing a specified logical function. It should also be noted that in some alternative embodiments, the functions mentioned in blocks or steps may occur out of order. For example, two blocks or steps shown in succession may in fact be performed substantially simultaneously, or the blocks or steps may sometimes be performed in the reverse order depending on the corresponding function.

As such, those skilled in the art to which the present invention pertains will be able to understand that the present invention can be implemented in other specific forms without changing the technical spirit or essential features. Therefore, the embodiments described above are illustrative in all respects and should be understood as non-limiting. The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. .

100: store management device
111: Business DB for each store
112: task list extraction unit
113: business decision
114: task assignment
115: business management department
116: business capability information extraction unit
117: DB for each employee's work
118: employee/task extraction unit
119: camera
120: display unit
121: processor
200: employee terminal

Claims (15)

In the store management method performed by the store management device,
Receiving store information;
Extracting a business list including a plurality of business information to be managed in a store corresponding to the store information based on the input store information;
Allocating a first employee to perform first job information included in the job list;
The first task information performed by the first employee, reported according to the number of times the first task information should be reported, based on a store-specific task DB that stores data on the number of times a predetermined task information should be reported Receiving the first evaluation information, which is an average value of evaluation information for, to be input;
Using the first job information, first employee information for the first employee, and the first evaluation information as training data for generating a first neural network model based on a predetermined machine learning algorithm, the first job information and the First training the first neural network model to learn a correlation of the first evaluation information from the first employee information;
Reassigning the first employee to perform the second job information included in the job list when the evaluation of at least one item of the first evaluation information is less than or equal to a preset score;
Based on the store-specific business DB, input second evaluation information, which is the average value of evaluation information on the second business information performed by the first employee, reported according to the number of times the second business information should be reported Receiving step;
Adding the first employee information, the second job information, and the second evaluation information as the learning data, so that the correlation of the second evaluation information is further learned from the second job information and the first employee information Secondary training the first neural network model; And
Inputting third employee information included in a manpower database storing employee information for each of a plurality of employees and a plurality of job information that can be assigned to employees of the third employee information into the second learned first neural network, Comprising the step of extracting a job competency score for each job that the employee of the third employee information is in charge of the plurality of job information,
Store management method of store management device.
delete The method of claim 1,
The first evaluation information
Including at least one of the items for job completeness, integrity, responsibility, affinity, kindness, and overall score.
Store management method of store management device.
The method of claim 1,
The first neural network model is
It is composed of a neural network including an input layer, a hidden layer, and an output layer, is set to input work information and employee information to the input layer, and is set to input evaluation information to the output layer, and is set to input a plurality of work information and employee information. As is input in the learning process of the first neural network model, the correlation is learned to derive the work capability information of the employee corresponding to the input work information.
Store management method of store management device.
The method of claim 1,
The step of extracting the task list
When store information including business type information and employee information is input, the work list is performed through a second neural network model learned through a predetermined machine learning algorithm to extract a plurality of work information to be managed in a store corresponding to the store information. Comprising the step of extracting
Store management method of store management device.
The method of claim 5,
The second neural network model is
Consisting of a neural network including an input layer, a hidden layer, and an output layer, actual store information including actual business type information and actual employee information from a plurality of manager terminals, and actual business information actually being performed in the store of the actual store information Based on the store-specific business DB receiving and storing the mapped data, the actual store information including the actual business type information and the actual employee information is set to be input to the input layer, and the actual store information is input to the output layer. The actual business information generated by the manager managing the store according to the store information is set to be input, and as a plurality of real store information is input in the learning process of the second neural network model, business information corresponding to the input store information is provided. The correlation is learned to derive,
Store management method of store management device.
The method of claim 6,
The second neural network model is
Store information including the actual business type information and the actual employee information is set to be input, and the actual business information generated by the manager managing the store according to the actual store information and data on the number of times are input to the output layer It is set so that, as a plurality of real store information is input in the learning process of the second neural network model, the correlation is learned to derive the number of reports that the business information and the business information corresponding to the input store information should be reported per hour.
Store management method of store management device.
The method of claim 7,
The step of allocating the first employee,
Including the step of designating to report the processing result of the job information corresponding to the input store information according to the number of reports derived by the second neural network model while allocating an employee to perform the job information included in the job list. doing,
Store management method of store management device.
delete The method of claim 1,
The step of extracting the task competency score for each task
Comprising the step of extracting a task that is evaluated above a preset score among the task competency scores as task information suitable for personnel in the personnel database
Store management method of store management device.
The method of claim 1,
The step of extracting the task competency score for each task
When a hiring request for predetermined job information is input from the manager terminal, extracting a number of personnel in which the job competency score for each job for the predetermined job information is mapped to a predetermined score or higher among personnel in the personnel database doing
Store management method of store management device.
The method of claim 1,
The first employee information,
Including one or more of gender, age, language spoken, education, major, height, and weight,
Store management method of store management device.
The method of claim 1,
The step of receiving the first evaluation information,
Receiving an image photographing a result of processing the first job information of the first employee; And
Including the step of determining the evaluation of the first task information based on the image,
Store management method of store management device.
A computer program stored in a computer-readable recording medium for causing a processor to perform the method of any one of claims 1, 3 to 8, and 10 to 13.
In the store management device,
Upon receiving the store information, a business list extracting unit for extracting a business list including a plurality of business information to be managed in a store corresponding to the store information;
A task assignment unit for allocating a first employee to perform first task information included in the task list;
The first task information performed by the first employee, reported according to the number of times the first task information should be reported, based on a store-specific task DB that stores data on the number of times a predetermined task information should be reported A task management unit receiving the first evaluation information, which is an average value of evaluation information for; And
Using the first job information, first employee information for the first employee, and the first evaluation information as training data for generating a first neural network model based on a predetermined machine learning algorithm, the first job information and the And a work capability information extracting unit for primary training the first neural network model so that the correlation of the first evaluation information is learned from the first employee information,
The task assignment unit, when the evaluation of at least one item of the first evaluation information is less than or equal to a preset score, reassigns the first employee to perform the second task information included in the task list,
The business management unit, based on the store-specific business DB, the second business information, which is an average value of evaluation information on the second business information performed by the first employee, reported according to the number of times the second business information should be reported, a second Receive evaluation information,
The work capability information extracting unit adds the first employee information, the second work information, and the second evaluation information as the learning data, and the second evaluation information from the second work information and the first employee information The first neural network model is secondarily trained to further learn the correlation of,
The store management device,
Inputting third employee information included in a manpower database storing employee information for each of a plurality of employees and a plurality of job information that can be assigned to employees of the third employee information into the second learned first neural network, Further comprising an employee recommendation unit for extracting a job competency score for each job in which an employee of the third employee information is in charge of the plurality of job information,
Store management device.

Images