Intelligent auditing decision-making method based on industrial big data

Abstract

The invention discloses an intelligent auditing decision-making method based on industrial big data, and is applied to the technical field of production decision-making. Comprising the following steps: acquiring demand data of different products in a plurality of historical periods, and converting the demand data into time sequence data; constructing a product demand prediction network, and training based on the data set; acquiring demand data of different products in a preset time period, converting the demand data into time sequence data, inputting the time sequence data into the corresponding demand prediction network, and obtaining demand prediction data of the different products; taking the demand prediction data as a production decision value, and taking the minimum total production cost and the minimum total production time as production scheduling targets to obtain an intelligent decision scheme for production scheduling; auditing the intelligent decision scheme; and arranging factories to produce different products based on the intelligent decision-making scheme. According to the invention, the decision scheme can be generated based on industrial production big data for reference of workers, so that the production scheme better meets the market demand.

Description

Technical Field

[0001] The present invention relates to the technical field of production decision-making, and more specifically to an intelligent audit decision-making method based on industrial big data. Background Art

[0002] Industrial big data is a general term for data on the entire life cycle of products and services in the industrial field, including data generated and used by industrial enterprises in R&D design, production and manufacturing, business management, operation and maintenance services, and data in the industrial Internet platform. Production system data originates from the front line of industrial production, mainly generated by equipment, machines and products on the production line. The data generated by production equipment through the Internet of Things provides instant information support for the scheduling, quality control and performance management of smart factories. The large amount of time series data collected by sensors covers detailed information such as equipment status parameters, workload and operating environment. These data are of great significance for improving equipment operation efficiency, enhancing the reliability of production processes and predictive maintenance. However, existing industrial big data applications are still mainly based on equipment monitoring and maintenance, production quality control and automated production. Production decisions still use manual judgment and lack the assistance of big data. Therefore, how to provide an intelligent audit decision-making method based on industrial big data is an urgent problem that technicians in this field need to solve. Summary of the invention

[0003] In view of this, the present invention provides an intelligent audit decision-making method based on industrial big data, which makes intelligent decisions and audits based on historical product demand data and production data to obtain a production plan.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] An intelligent audit decision-making method based on industrial big data includes the following steps:

[0006] S1. Obtain demand data of different products in multiple historical periods, convert the demand data into time series data, and use the time series data as a data set for training the product demand prediction network;

[0007] S2. Build a product demand prediction network, train it based on the data set, and obtain demand prediction networks for different products;

[0008] S3, collecting demand data of different products within a predetermined time period, converting the demand data into time series data and inputting it into the corresponding demand forecasting network to obtain demand forecast data of different products;

[0009] S4. Using demand forecast data as production decision values, taking the lowest total production cost and the shortest total production time as production scheduling goals, and combining production data to obtain an intelligent decision-making solution for production scheduling;

[0010] S5. Audit the intelligent decision-making scheme;

[0011] S6. Arrange factories to produce different products based on intelligent decision-making solutions.

[0012] Optionally, S1 is specifically as follows: obtain historical demand data of different products, divide the historical data into multiple historical periods according to the set time interval, calculate the average different products in each historical period, and convert the demand data into time series data x t =(y t ,δ t ,w t ), where t∈[1,n], n is the number of historical periods, y t represents the demand in the tth historical period, δ t represents the total number of historical periods with zero demand before the tth historical period, w t Represents the number of production days in the t-th historical period.

[0013] Optionally, S2 specifically includes: constructing an LSTM network, determining the training time step, the number of hidden layers, and the number of hidden neurons of the LSTM network; normalizing the data set, and mapping the data to the range of [0,1]; using the time series data of different products in the data set as training samples, using the Adam optimization algorithm for training, and using the trained LSTM network as the demand prediction network for the corresponding product.

[0014] Optional, S3 is:

[0015] S31. Collect demand data of different products within a predetermined time period and convert the demand data into time series data x t =(y t ,δ t ,w t );

[0016] S32, the hidden state h of the previous historical period t-1 With the current input x t Input the demand forecasting network and get the retention degree f t ;

[0017] S33, demand forecasting network based on the hidden state h of the previous historical period t-1 With the current input x t Get candidate values Update value i t and the output value o t ;

[0018] S34, based on the state value c of the previous historical period t-1, retention degree t , candidate value and update the value i t Calculate the state value c of the current historical period t ;

[0019] S35. Status value c based on the current historical period t and the output value o t Calculate the hidden state h of the current historical period t .

[0020] Optionally, S4 specifically includes: using the demand forecast data of different products as the production decision value to determine the production quantity of the corresponding products, combining the production data to calculate the production time and production cost of the products, taking the minimum total production cost and the minimum total production time as the objective function of production scheduling, setting constraints, constructing a mathematical model for the production scheduling optimization problem, solving the mathematical model of the production scheduling optimization problem based on the genetic algorithm, and obtaining an intelligent decision-making plan for production scheduling.

[0021] Optionally, the total production cost objective function of the product is:

[0022]

[0023] In the formula, n is the number of products, n i represents the number of processes for product i, m represents the number of production machines, and a i is the material cost of product i, b ijk represents the cost of processing product i in process j on machine k, c ijk represents the labor cost of processing product i in process j on machine k, d ijk Indicates whether process j of product i is processed on machine k.

[0024] Optionally, the total production time objective function of the product is:

[0025]

[0026] Where n i represents the number of processes for product i, s ijk represents the time when process j of product i starts processing on machine k, t ijk represents the time it takes for process j of product i to be processed on machine k, d ijk Indicates whether process j of product i is processed on machine k.

[0027] Optionally, S5 specifically includes: obtaining the cost data, production process data, market information data, demand forecast data, decision process log data and industry standard data required for the production scheduling plan to execute the intelligent decision-making solution, storing them in the database, standardizing and auditing the data based on the audit project, and generating an audit report.

[0028] It can be seen from the above technical solution that compared with the prior art, the present invention provides an intelligent audit decision-making method based on industrial big data, which has the following beneficial effects: the present invention can generate decision plans based on industrial production big data for reference by staff, which can make production plans more in line with market demand, and at the same time consider demand and production equipment to make production scheduling decision plans, improve production efficiency, adopt intelligent processing methods, and improve decision-making efficiency; audit intelligent decision plans to ensure the rationality of production scheduling decisions, so as to facilitate the grasp of production processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0030] Figure 1 It is a flow chart of the intelligent audit decision-making method of the present invention;

[0031] Figure 2 This is a demand forecasting network structure diagram of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] The embodiment of the present invention discloses an intelligent audit decision-making method based on industrial big data, such as Figure 1 As shown, the following steps are included:

[0034] S1. Obtain demand data of different products in multiple historical periods, convert the demand data into time series data, and use the time series data as a data set for training the product demand prediction network;

[0035] S2. Build a product demand prediction network, train it based on the data set, and obtain demand prediction networks for different products;

[0036] S3, collecting demand data of different products within a predetermined time period, converting the demand data into time series data and inputting it into the corresponding demand forecasting network to obtain demand forecast data of different products;

[0037] S4. Using demand forecast data as production decision values, taking the lowest total production cost and the shortest total production time as production scheduling goals, and combining production data to obtain an intelligent decision-making solution for production scheduling;

[0038] S5. Audit the intelligent decision-making scheme;

[0039] S6. Arrange factories to produce different products based on intelligent decision-making solutions.

[0040] Optionally, S1 is specifically as follows: obtain historical demand data of different products, divide the historical data into multiple historical periods according to the set time interval, calculate the average different products in each historical period, and convert the demand data into time series data x t =(y t ,δ t ,w t ), where t∈[1,n], n is the number of historical periods, y t represents the demand in the tth historical period, δ t represents the total number of historical periods with zero demand before the tth historical period, w t Represents the number of production days in the t-th historical period.

[0041] Optionally, S2 is specifically: constructing an LSTM network, the structure of the LSTM network is as follows Figure 2 As shown in the figure, the training time step, number of hidden layers and number of hidden neurons of the LSTM network are determined; the data set is normalized and mapped to the range of [0,1]; the time series data of different products in the data set are used as training samples, and the Adam optimization algorithm is used for training, and the trained LSTM network is used as the demand prediction network for the corresponding products.

[0042] In the embodiment of the present invention, the normalization adopts the following formula:

[0043]

[0044] In the formula, represents the jth data in the i-th feature of the normalized input data x, x ij represents the jth data in the i-th feature of the input data x before normalization, max(x i) represents the maximum value of the i-th feature, min(x i ) represents the minimum value of the i-th feature;

[0045] In the embodiment of the present invention, the number of training time steps, that is, historical periods, is 10, the number of hidden layers is 1, and the number of hidden neurons is 128.

[0046] Optional, S3 is:

[0047] S31. Collect demand data of different products within a predetermined time period and convert the demand data into time series data x t =(y t ,δ t ,w t );

[0048] S32, the hidden state h of the previous historical period t-1 With the current input x t Input the demand forecasting network and get the retention degree f t :

[0049] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0050] The LSTM network consists of three gate structures, among which the forget gate outputs a value between 0 and 1 through the Sigmoid activation function to determine the degree of retention of each state, that is, f t ; The closer to 0, the higher the degree of forgetting, and the closer to 1, the higher the degree of retention. In the formula, W f is the weight matrix of the forget gate, b f is the bias vector of the forget gate, σ(·) represents the Sigmoid activation function; h t-1 is the hidden state at the previous moment, x t is the current input;

[0051] S33, demand forecasting network based on the hidden state h of the previous historical period t-1 With the current input x t Get candidate values Update value i t and the output value o t :

[0052] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0053]

[0054] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0055] The input gate of the LSTM network mainly controls the input at the current moment and updates the unit state. First, the hidden state h of the previous period is t-1 and the current input x t Passed to the Sigmoid activation function, the updated value i is obtained t Decide which information to update, and secondly, update the hidden state h of the previous period t-1 and the current input x t To the tanh function, generate new candidate values The output gate generates the output value o through the Sigmoid function t ;

[0056] Where W i , W C are the weight matrices of the input gate and candidate values, respectively, and b i 、b C are the bias vectors of the input gate and candidate value, respectively, W o is the weight matrix of the output gate, b o is the bias vector for the output gate.

[0057] S34, based on the state value c of the previous historical period t-1 , retention degree t , candidate value and update the value i t Calculate the state value c of the current historical period t :

[0058]

[0059] S35. Status value c based on the current historical period t and the output value o t Calculate the hidden state h of the current historical period t :

[0060] h t =o t *tanh(C t ).

[0061] Optionally, S4 specifically includes: using the demand forecast data of different products as the production decision value to determine the production quantity of the corresponding products, combining the production data to calculate the production time and production cost of the products, taking the minimum total production cost and the minimum total production time as the objective function of production scheduling, setting constraints, constructing a mathematical model for the production scheduling optimization problem, solving the mathematical model of the production scheduling optimization problem based on the genetic algorithm, and obtaining an intelligent decision-making plan for production scheduling.

[0062] Optionally, the total production cost objective function of the product is:

[0063]

[0064] In the formula, n is the number of products, n i represents the number of processes for product i, m represents the number of production machines, and a i is the material cost of product i, b ijk represents the cost of processing product i in process j on machine k, c ijk represents the labor cost of processing product i in process j on machine k, d ijk Indicates whether process j of product i is processed on machine k.

[0065] Optionally, the total production time objective function of the product is:

[0066]

[0067] Where n i represents the number of processes for product i, s ijk represents the time when process j of product i starts processing on machine k, t ijk represents the time it takes for process j of product i to be processed on machine k, d ijk Indicates whether process j of product i is processed on machine k.

[0068] In the embodiment of the present invention, the constraints are:

[0069] 1. All product processes must be carried out in order:

[0070] C ij -C i(j-1) ≥t ijk d ijk

[0071] In the formula, C ij , C i(j-1) are the time when process j of product i is completed and the time when process j-1 of product i is completed respectively;

[0072] 2. The same machine can only process one process at a time;

[0073] 3. Processing is completed continuously after it starts

[0074]

[0075] Genetic Algorithm (GA) is a computational model of biological evolution that simulates the natural selection and genetic mechanism of Darwin's theory of biological evolution. It is a method for searching for the optimal solution by simulating the natural evolution process. It is suitable for dealing with complex and nonlinear problems that are difficult to solve with traditional search methods. It has the remarkable characteristics of being simple and universal, strong in robustness, suitable for parallel processing, and having a wide range of applications.

[0076] Optionally, S5 specifically includes: obtaining the cost data, production process data, market information data, demand forecast data, decision process log data and industry standard data required for the production scheduling plan to execute the intelligent decision-making solution, storing them in the database, standardizing and auditing the data based on the audit project, and generating an audit report.

[0077] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0078] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent audit decision-making method based on industrial big data, characterized in that: The following steps are involved: S1. Obtain demand data of different products in multiple historical periods, convert the demand data into time series data, and use the time series data as a data set for training the product demand prediction network; S2. Build a product demand prediction network, train it based on the data set, and obtain demand prediction networks for different products; S3, collecting demand data of different products within a predetermined time period, converting the demand data into time series data and inputting it into the corresponding demand forecasting network to obtain demand forecast data of different products; S4. Using demand forecast data as production decision values, taking the lowest total production cost and the shortest total production time as production scheduling goals, and combining production data to obtain an intelligent decision-making solution for production scheduling; S5. Audit the intelligent decision-making scheme; S6. Arrange factories to produce different products based on intelligent decision-making solutions.

2. According to the intelligent audit decision-making method based on industrial big data according to claim 1, it is characterized in that: S1 is specifically as follows: obtain the historical demand data of different products, divide the historical data into multiple historical periods according to the set time interval, calculate the average different products in each historical period, and convert the demand data into time series data x t =(y t ,δ t ,w t ), where t∈[1,n], n is the number of historical periods, y t represents the demand in the tth historical period, δ t represents the total number of historical periods with zero demand before the tth historical period, w t Represents the number of production days in the t-th historical period.

3. According to the intelligent audit decision-making method based on industrial big data according to claim 1, it is characterized in that: S2 is specifically as follows: construct an LSTM network, determine the training time step, number of hidden layers and number of hidden neurons of the LSTM network; normalize the data set and map the data to the range of [0,1]; use the time series data of different products in the data set as training samples, use the Adam optimization algorithm for training, and use the trained LSTM network as the demand prediction network for the corresponding products.

4. According to the intelligent audit decision-making method based on industrial big data according to claim 1, it is characterized in that: S3 is specifically: S31. Collect demand data of different products within a predetermined time period and convert the demand data into time series data x t =(y t , δ t , w t ); S32, the hidden state h of the previous historical period t-1 With the current input x t Input the demand forecasting network and get the retention degree f t ; S33, demand forecasting network based on the hidden state h of the previous historical period t-1 With the current input x t Get candidate value Update value i t and the output value o t ; S34, based on the state value c of the previous historical period t-1 , retention degree t , candidate value and update the value i t Calculate the state value c of the current historical period t ; S35. Status value c based on the current historical period t and the output value o t Calculate the hidden state h of the current historical period t .

5. According to the intelligent audit decision-making method based on industrial big data according to claim 1, it is characterized in that: S4 is specifically as follows: using the demand forecast data of different products as the production decision value to determine the production quantity of the corresponding products, combining the production data to calculate the production time and production cost of the products, taking the minimum total production cost and the minimum total production time as the objective function of production scheduling, setting constraints, constructing a mathematical model for the production scheduling optimization problem, solving the mathematical model of the production scheduling optimization problem based on the genetic algorithm, and obtaining an intelligent decision-making solution for production scheduling.

6. According to claim 5, an intelligent audit decision-making method based on industrial big data is characterized in that: The total production cost objective function of the product is: In the formula, n is the number of products, n i represents the number of processes for product i, m represents the number of production machines, and a i is the material cost of product i, b ijk represents the cost of processing product i in process j on machine k, c ijk represents the labor cost of processing product i in process j on machine k, d ijk Indicates whether process j of product i is processed on machine k.

7. According to claim 5, an intelligent audit decision-making method based on industrial big data is characterized in that: The total production time objective function of the product is: Where n i represents the number of processes for product i, s ijk represents the time when process j of product i starts processing on machine k, t ijk represents the time it takes for process j of product i to be processed on machine k, d ijk Indicates whether process j of product i is processed on machine k.

8. According to claim 1, an intelligent audit decision-making method based on industrial big data is characterized in that: S5 is specifically: obtaining the cost data, production process data, market information data, demand forecast data, decision-making process log data and industry standard data required for the production scheduling plan to execute the intelligent decision-making solution, storing them in the database, standardizing the data and performing audit analysis based on the audit project, and generating an audit report.

Images