CN113780655A - Steel multi-variety demand prediction method based on intelligent supply chain - Google Patents

Abstract

The invention relates to a demand forecasting method for multiple varieties of steel based on a smart supply chain. Demand data time series to construct SARIMA time series model; 3) Input the training sample set to be tested into SARIMA time series model to obtain demand forecast results for multiple varieties of steel; 4) Based on the demand data time series of multiple varieties of steel, construct SARIMAX time series Sequence model; 5) Input the test sample set to be tested into the SARIMAX time series model, and obtain the demand forecast results of multiple varieties of steel revised by the algorithm; 6) Obtain the final forecast according to the results output by the SARIMA model and the correction results output by the SARIMAX model result. Compared with the prior art, the present invention has the advantages of making the time series more stable and having higher prediction accuracy.

Description

A method of forecasting demand for steel products based on smart supply chain

technical field

The invention relates to the technical field of supply chain production management, in particular to a demand forecasting method for multiple varieties of steel based on a smart supply chain.

Background technique

With the development of the Industrial Internet, the downstream supply chain of the steel industry has also ushered in transformation and upgrading, and the production organization method of its main raw material steel also needs to be changed accordingly. Transformed into a flexible production and marketing balance of production organization. It is necessary to use the information exchange with the downstream supply chain enterprises in the process of manufacturing industrial products, and combine effective technical means to mine the laws of industrial product production data, and predict in advance the demand for multiple varieties of steel in the downstream supply chain of the steel industry for a certain period of time in the future, so as to improve The steel production efficiency and capacity resource utilization rate of steel production enterprises can meet the steel consumption demand of the downstream supply chain more quickly and in a timely manner.

However, at present, the technical means of demand analysis, processing and decision-making for the downstream supply chain are relatively backward. For the industrial product production data of downstream supply chain enterprises collected in the smart supply chain, it mainly relies on manual experience or simple mathematical formulas to determine the corresponding steel varieties. demand forecast. Based on the current demand forecast, it is difficult for steel mills to reasonably and efficiently arrange production plans for different steel varieties, resulting in difficulty in improving production efficiency and capacity resource utilization. This is a technical problem that needs to be improved and solved urgently. In other fields, when forecasting demand through various big data forecasting techniques, more consideration is given to non-technical issues such as corporate profitability and cost. The current technical means and methods of demand forecasting are time-consuming, poorly systematic, low in integrity, low in accuracy, and slow in response and output.

To sum up, the existing multi-variety steel demand forecasting methods cannot improve production efficiency and capacity resource utilization, and cannot meet the intelligent linkage between the production of intelligent supply chain industrial products and the multi-variety steel demand reserve and production scheduling of steel production enterprises.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a method for predicting the demand for multiple varieties of steel based on a smart supply chain in order to overcome the above-mentioned defects in the prior art. The processing method can make timely, accurate and stable forecasts for the demand for multiple varieties of steel in a certain period of time in the future.

The object of the present invention can be realized through the following technical solutions:

A method for forecasting demand for multiple varieties of steel based on a smart supply chain, the method includes the following steps:

S1: Based on the industrial product production data of the smart supply chain system, obtain the demand data time series of various types of steel. The industrial product production data of the smart supply chain system includes production BOM data, main engine production plant production data, and component production plant production data. The specific steps to obtain the time series of demand data for multiple varieties of steel include:

11) Take the acquired production data of industrial products in the smart supply chain system as the original data time series;

12) Based on the technical framework of the SARIMA model, clean the original data time series, adjust the date format of the data, fill in the missing data with the moving average method, set the threshold number of the month, and fill in zeros for the data above the threshold number of consecutive missing months , and filter out the top 90% of the data in a certain period as the available data source;

13) According to the production BOM data, the production data of the main engine production plant and the production data of the parts production plant, establish the functional relationship between the production data and the demand data W _t =g(P ₁ ,P ₂ ,...,P _n ), where W _t is the demand value of different varieties of steel in different months, P ₁ , P ₂ ,...,P _n is the production value of different industrial products, using the functional relationship to solve and convert the production data of industrial products in the smart supply chain into relative Corresponding time series of demand data for multiple varieties of steel;

14) Perform sliding average processing on the demand data time series of multiple varieties of steel obtained in step 13) to obtain the sample data length and initially stable demand data time series used in the rolling forecast of the demand forecasting model.

S2: Build a SARIMA time series model based on the demand data time series of multiple varieties of steel.

Specific steps include:

21) Based on the seasonal law, construct the SARIMA time series model:

In the formula, W _t , w _tn , and w _t-sn are the demand values of different varieties of steel in months t, tn, and t-sn respectively, μ is a constant term, ε _t , ε _tn , and ε _t-sn are respectively the demand values of different varieties of steel Errors of demand values in months t, tn, and t-sn, p is the number of trend autoregressive items, q is the number of trend moving average items, P is the number of seasonal autoregressive items, and Q is the number of seasonal moving average items . α _n is the trend autoregressive coefficient, θ _n is the trend moving average coefficient, φ _n is the seasonal autoregressive coefficient, and η _n is the seasonal moving average coefficient.

22) According to the results of the three characteristic sub-items obtained by the model decomposition, calculate the standard deviation between the seasonal characteristic components and the random characteristic components, and adjust the abnormal values of all demand data time series within the set interval; the seasonal characteristic components and The standard deviation σ between random feature components is calculated as:

In the formula, Cs and Cr are the seasonal feature components and random feature components automatically decomposed from the demand data time series;

Use the trend feature component Ct±standard deviation to obtain the upper and lower limits of the demand data time series values before each forecast, and adjust the abnormal values of all demand data time series within the program setting interval, namely [Ct-σ,Ct+σ] .

23) Use the grid search training and optimization selected by the model to obtain different combinations of SARIMA(p,d,q)(P,D,Q)s parameters, d is the number of trend differences made to make it the final stationary sequence, D is the number of seasonal differences made to make it the final stationary series, s is the number of time steps in a single seasonal period, p is the number of trend autoregressive items, q is the number of trend moving average items, and P is seasonal autoregression The number of items, Q is the number of seasonal moving average items; calculate the AIC values of different combinations, select the parameter combination with the smallest AIC value as the best parameter combination, and complete the final conditions for establishing the SARIMA time series model.

S3: Input the training sample set to be tested into the constructed SARIMA time series model, and obtain the demand forecast results of various types of steel.

S4: Build a SARIMAX time series model based on the demand data time series of multiple varieties of steel in S1.

Specific steps include:

41) Based on the time series of the demand data of various types of steel corresponding to the production data of the main engine manufacturer and the production data of the parts manufacturer obtained from S1, normalize it to obtain two groups of (0, 1) mapped in the range. data;

42) Use correlation analysis to calculate the correlation coefficient of the two groups of mapped data, filter out the data corresponding to the correlation coefficient ≥ 0.7 as the advance data, and use the production data of the parts factory before normalization to correspond to the demand for multiple varieties of steel The data, as the exogenous variable X of the SARIMAX model, have:

为零部件生产厂的生产数据对应的钢材多品种的需求值，β_n为外部变量

的回归系数，r为外部变量的回归项数；In the formula,

is the demand value of multiple varieties of steel corresponding to the production data of the parts manufacturer, and β _n is an external variable

The regression coefficient of , r is the number of regression items of external variables;

43) Based on the demand data time series obtained after S2 processing, use the grid search selected by the model to retrain and optimize to obtain different combinations of SARIMAX(p,d,q)(P,D,Q)s parameters, and calculate the AIC value, the parameter combination with the smallest AIC value is selected as the optimal parameter combination to complete the final condition for establishing the SARIMAX time series model.

S5: Input the training sample set to be tested to the constructed SARIMAX time series model, and obtain the demand forecast results of various steel varieties corrected by the algorithm.

S6: Obtain a final forecast result according to the demand forecast result output by the SARIMA time model obtained in S3 and the revised demand forecast result output by the SARIMAX time series model obtained in S5.

The specific contents are:

与S5得到的SARIMAX时间序列模型输出的修正后的需求量预测结果

获取最终的预测结果

其中

为使用SARIMA时间序列模型输出、除去可用SARIMAX时间序列模型输出结果的需求时间序列。Demand forecast results based on SARIMA time model output from S3

The revised demand forecast results from the SARIMAX time series model output obtained with S5

Get the final forecast

in

For the demand time series output using the SARIMAX time series model, remove the available SARIMAX time series model output results.

Further, in step 12), taking twelve months as a complete cycle, the data of the top 90% of the data volume of a complete cycle are screened out as available data sources.

Further, in step 42), filter out the data with correlation coefficient ≥ 0.7 and W _b is the leading data of W _a , W _a and W _b are the two groups (0, 1) mapped in the range obtained in step 41). data.

Compared with the prior art, the method for predicting the demand for multiple varieties of steel based on the smart supply chain provided by the present invention at least includes the following beneficial effects:

1) The present invention can realize the use of the rules of industrial product production data in the smart supply chain and a targeted data processing method to obtain the length of the sample data used in the rolling forecast of the demand forecasting model and the initially stable demand data time series, and the demand data time series. The sequence is more stable, which helps to improve the accuracy of prediction;

2) The SARIMA and SARIMAX time series models are established, which can predict the time series operation data of the demand data of multiple varieties of steel, that is, the non-stationary time series, which can not only convert the non-stationary time series into a stationary time series, but also consider seasonal factors. In addition, by mining the natural laws existing between the production data of different industrial products, the effective characteristics of the model are formed, and the algorithm efficiency and complexity of the data source and correlation factor analysis dimensions are reduced, so that the time series based on the above can be calculated. Obtain prediction results with high accuracy and engineering application;

3) Through the multi-variety demand forecasting method for steel based on the smart supply chain of the present invention, a timely, accurate and stable forecast is made for the multi-variety demand for steel in a certain period of time in the future, so as to solve the problem that the steel mill cannot reasonably and efficiently arrange different steel varieties. It is a technical problem that it is difficult to improve the production efficiency and the utilization rate of capacity resources.

Description of drawings

FIG. 1 is a schematic flowchart of a method for predicting demand for multiple varieties of steel based on a smart supply chain in an embodiment.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

Example

As shown in FIG. 1 , the present invention relates to a method for forecasting demand for multiple varieties of steel based on a smart supply chain, the method comprising the following steps:

Step 1: Obtain the time series of demand data for multiple varieties of steel.

Obtain the production data of industrial products from the smart supply chain system, including production BOM (Bill of Materials) data, main engine production plant production data, parts production plant production data and other original data time series; based on the SARIMA model technical framework, the original data time series Data preprocessing such as cleaning, screening, conversion, and moving average is performed on the sequence to obtain the length of the multi-variety sample data of steel used in the rolling forecast of the demand forecast model and the required stable demand data time series. specifically:

Step101: Obtain the production data of industrial products in the smart supply chain system, including the production BOM (Bill of Materials) data, the production data of the main engine production plant, the production data of the parts production plant and other original data time series;

Step102: Clean the original data time series based on the technical framework of the SARIMA model, use the program to adjust the data date format, use the sliding average method to fill in the missing data, set the threshold to 6 months, and fill in the data with more than 6 consecutive months of missing data as 0. In order to improve the stability of the time series, and combined with the production data of industrial products, 12 months is a complete cycle, and the top 90% of the data in a complete cycle is selected as the available data source;

Step103: According to the production BOM (Bill of Materials) data, the production data of the main engine manufacturer, and the production data of the parts manufacturer, establish the functional relationship between the production data and the demand data, W _t =g(P ₁ ,P ₂ ,..., P _n ), solve and convert the production data of industrial products in the smart supply chain into the corresponding time series of demand data for multiple varieties of steel, where W _t is the demand value of different varieties of steel in different months, P ₁ , P ₂ ,. .., P _n is the production value of different industrial products;

Step 104: Perform 3-month moving average processing on the demand data time series of multiple varieties of steel obtained in Step 103, and obtain the sample data length and initial stable demand data time series used in the rolling forecast of the demand forecast model.

Step 2: Establish a SARIMA time series model based on the demand data time series obtained in step 1.

The SARIMA algorithm model is continuously adjusted and optimized, and the trend feature component, seasonal feature component and random feature component are automatically decomposed from the demand data time series; The standard deviation between the random feature components, the upper and lower limits of the demand data time series values are obtained by the trend feature components ± standard deviation, and the abnormal values of all demand data time series are adjusted within the program setting range to avoid extreme demand data in individual months Influence on the overall data law; use the grid search training and optimization of model selection to obtain different combinations of SARIMA (p, d, q) × (P, D, Q) parameters, and use the program to calculate the AIC (Akaike Information Criterion) of different combinations. ) value, the parameter combination with the smallest AIC value is selected as the optimal parameter combination, and the final conditions for establishing the SARIMA time series model are completed. Specific steps include:

Step201: Due to the obvious seasonality in the production data of industrial products in the smart supply chain related to the consumption of steel varieties, that is, the production volume in summer and winter is greater than that in spring and autumn, so the demand data for multiple varieties of steel has seasonal patterns. Variation in nature, the SARIMA time series model can be constructed:

In the formula, W _t , w _tn , and w _t-sn are the demand values of different varieties of steel in months t, tn, and t-sn respectively, μ is a constant term, ε _t , ε _tn , and ε _t-sn are respectively the demand values of different varieties of steel Errors of demand values in months t, tn, and t-sn, p is the number of trend autoregressive items, q is the number of trend moving average items, P is the number of seasonal autoregressive items, and Q is the number of seasonal moving average items . α _n is the trend autoregressive coefficient, θ _n is the trend moving average coefficient, φ _n is the seasonal autoregressive coefficient, and η _n is the seasonal moving average coefficient.

Using the composition of the SARIMA algorithm model, automatic adjustment and optimization, the trend characteristic component, seasonal characteristic component and random characteristic component are automatically decomposed from the demand data time series, which are recorded as Ct, Cs, and Cr respectively.

Step202: According to the results of the three feature sub-items obtained by the model decomposition, use the program to calculate the standard deviation between the seasonal feature component and the random feature component. The calculation formula is as follows:

In order to avoid the influence of extreme demand data of individual months on the overall data law, the upper and lower limits of the demand data time series values before each forecast are obtained by using the trend characteristic component ± standard deviation, and the abnormal values of all demand data time series are adjusted in the program setting. within a certain interval, namely [Ct-σ, Ct+σ].

Step203: Use the grid search training and optimization selected by the model to obtain different combinations of SARIMA(p,d,q)(P,D,Q)s parameters, d is the number of trend differences made to make it the final stationary sequence, D is the number of seasonal differencing done to make it the final stationary series, and s is the number of time steps in a single seasonal period. The AIC (Akaike Information Criterion) values of different combinations are calculated by the program, and the parameter combination with the smallest AIC value is selected as the optimal parameter combination to complete the final condition for establishing the SARIMA time series model.

Step 3. Based on the SARIMA time series model obtained in Step 2, and based on the final stationary demand data time series processed in Steps 1 and 2, select the data of the time period of [T-24, T] as the training sample set, and input the data to be tested. The training sample set for T+2 months (the next 30 to 60 days) is rolled out to forecast the demand for multiple varieties of steel, and the forecast results of the demand for multiple varieties of steel are obtained.

Step 4. Since there are obvious natural laws in the production data of the main engine production plant and the production data of the parts production plant, the production data of the parts production plant is always ahead of the changes in the production data of the main engine production plant. Using this law, based on the steps Once the time series of the demand data for various types of steel corresponding to the production data of the main engine manufacturer and the production data of the parts manufacturer is obtained, normalize it to obtain two sets of data mapped in the range of (0, 1); Based on the SARIMA algorithm model in step 2 and step 3, the grid search training and optimization of model selection are used to obtain SARIMAX(p,d ,q)(P,D,Q)s parameters for different combinations, use the program to calculate the AIC (Akaike Information Criterion) values of different combinations, select the parameter combination with the smallest AIC value as the best parameter combination, and complete the establishment of the SARIMAX time series model. final condition. Specific steps include:

Step 401: Based on the time series of the demand data for multiple varieties of steel corresponding to the production data of the main engine manufacturer and the production data of the parts manufacturer obtained in step 1, normalize it to obtain two groups of (0, 1) mapped in the range. data, respectively denoted as W _a , W _b ;

Step 402: Use correlation analysis to calculate the correlation coefficient of the two groups of mapped data, filter out the data with the correlation coefficient ≥ 0.7 and W _b is the leading data of W _a , and normalize the production data of the former parts factory corresponding to the multi-variety of steel The demand data of , as the exogenous variable X of the SARIMAX model, there are:

为零部件生产厂的生产数据对应的钢材多品种的需求值，β_n为外部变量

的回归系数，r为外部变量的回归项数；in,

is the demand value of multiple varieties of steel corresponding to the production data of the parts manufacturer, and β _n is an external variable

The regression coefficient of , r is the number of regression items of external variables;

Step403: Based on the demand data time series obtained after the processing in step 2, in order to avoid the influence of the combination selection of SARIMAX(p,d,q)(P,D,Q)s parameters after adding the exogenous variable X, use the model selected Grid search is retrained and optimized to obtain different combinations of SARIMAX(p,d,q)(P,D,Q)s parameters, and the program is used to calculate the AIC (Akaike Information Criterion) values of different combinations, and the parameter combination with the smallest AIC value is selected As the optimal parameter combination, the final conditions for establishing the SARIMAX time series model are completed.

Step 5. Based on the SARIMAX time series model obtained in step 4, and based on the final stationary demand data time series processed in steps 1 and 2, select the data with the time period length of [T-24, T] as the training sample set, and input the data to be tested. The training sample set of T+2 months (the next 60 days) is used to predict the demand for multiple varieties of steel, and the forecast result of the demand for multiple varieties of steel after the algorithm is revised.

和基于步骤五得到的SARIMAX模型输出的修正后的需求量预测结果

则最终的预测结果

其中

为使用SARIMA模型输出、除去可用SARIMAX模型输出结果的需求时间序列。Step 6. Demand forecast results based on the SARIMA model output obtained in Step 3

and the revised demand forecast result based on the SARIMAX model output obtained in step 5

the final prediction result

in

For the demand time series using the SARIMA model output, excluding the available SARIMAX model output results.

The invention can realize the use of the rules of industrial product production data in the smart supply chain and the targeted data processing method to obtain the length of the sample data used in the rolling forecast of the demand forecasting model and the initially stable demand data time series, and the demand data time series is more accurate. Stable, which helps improve the accuracy of predictions. The SARIMA and SARIMAX time series models are established to predict the time series operation data of the demand data of various types of steel, that is, the non-stationary time series, which can not only convert the non-stationary time series into a stationary time series, but also consider the influence of seasonal factors. , and by mining the natural laws existing between the production data of different industrial products, the effective features of the model are formed, and the algorithm efficiency and complexity of the data source and correlation factor analysis dimensions are reduced, so that the above-mentioned time series can be accurately obtained. Predictive results for high-rate, engineerable applications.

The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited to this. Any person familiar with the technical field can easily think of various equivalents within the technical scope disclosed by the present invention. Modifications or substitutions should be included within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims (10)

1. a kind of steel multi-variety demand forecasting method based on smart supply chain, is characterized in that, comprises the following steps: 1) Based on the industrial product production data of the smart supply chain system, obtain the demand data time series of various types of steel; 2) Based on the demand data time series of multiple varieties of steel, build a SARIMA time series model; 3) Input the training sample set to be tested into the constructed SARIMA time series model, and obtain the demand forecast results of various types of steel; 4) Based on the demand data time series of multiple varieties of steel in step 1), construct a SARIMAX time series model; 5) Input the training sample set to be tested into the constructed SARIMAX time series model, and obtain the demand forecast results of multiple varieties of steel revised by the algorithm; 6) Obtain the final forecast result according to the demand forecast result output by the SARIMA time model obtained in step 3) and the revised demand forecast result output by the SARIMAX time series model obtained in step 5). 2. The multi-variety demand forecasting method for steel based on a smart supply chain according to claim 1, wherein the industrial product production data of the smart supply chain system comprises production BOM data, main engine production plant production data and parts production Factory production data. 3. The method for forecasting demand for multiple varieties of steel based on a smart supply chain according to claim 2, wherein the specific steps of obtaining the demand data time series of multiple varieties of steel include: 11) Take the acquired production data of industrial products in the smart supply chain system as the original data time series; 12) Based on the technical framework of the SARIMA model, clean the original data time series, adjust the date format of the data, fill in the missing data with the moving average method, set the threshold number of the month, and fill in zeros for the data above the threshold number of consecutive missing months , and filter out the top 90% of the data in a certain period as the available data source; 13) According to the production BOM data, the production data of the main engine production plant and the production data of the parts production plant, establish the functional relationship between the production data and the demand data W _t =g(P ₁ ,P ₂ ,...,P _n ), where W _t is the demand value of different varieties of steel in different months, P ₁ , P ₂ ,...,P _n is the production value of different industrial products, using the functional relationship to solve and convert the production data of industrial products in the smart supply chain into relative Corresponding time series of demand data for multiple varieties of steel; 14) Perform sliding average processing on the demand data time series of multiple varieties of steel obtained in step 13) to obtain the sample data length and initially stable demand data time series used in the rolling forecast of the demand forecasting model. 4. the multi-variety demand forecasting method for steel based on smart supply chain according to claim 1, is characterized in that, the concrete steps of step 2) comprise: 21) Based on seasonal law, build SARIMA time series model; 22) According to the results of the three feature sub-items obtained by the model decomposition, calculate the standard deviation between the seasonal feature component and the random feature component, and adjust the abnormal values of all demand data time series within the set interval; 23) Use the grid search training and optimization selected by the model to obtain different combinations of SARIMA(p,d,q)(P,D,Q)s parameters, d is the number of trend differences made to make it the final stationary sequence, D is the number of seasonal differences made to make it the final stationary series, s is the number of time steps in a single seasonal period, p is the number of trend autoregressive items, q is the number of trend moving average items, and P is seasonal autoregression The number of items, Q is the number of seasonal moving average items; calculate the AIC values of different combinations, select the parameter combination with the smallest AIC value as the best parameter combination, and complete the final conditions for establishing the SARIMA time series model. 5. the multi-variety demand forecasting method for steel based on smart supply chain according to claim 4, is characterized in that, in step 21), the expression of the SARIMA time series model of construction is:

In the formula, W _t , w _tn , and w _t-sn are the demand values of different varieties of steel in months t, tn, and t-sn respectively, μ is a constant term, ε _t , ε _tn , and ε _t-sn are respectively the demand values of different varieties of steel Errors of demand values in months t, tn, and t-sn, p is the number of trend autoregressive items, q is the number of trend moving average items, P is the number of seasonal autoregressive items, and Q is the number of seasonal moving average items , α _n is the trend autoregressive coefficient, θ _n is the trend moving average coefficient, φ _n is the seasonal autoregressive coefficient, and η _n is the seasonal moving average coefficient. 6. The multi-variety demand forecasting method for steel based on an intelligent supply chain according to claim 5, wherein the concrete steps of step 4) comprise: 41) Based on the time series of demand data for multiple varieties of steel corresponding to the production data of the main engine manufacturer and the production data of the parts manufacturer obtained in step 1), normalize it to obtain two sets of (0, 1) within the range mapped data; 42) Use correlation analysis to calculate the correlation coefficient of the two groups of mapped data, filter out the data corresponding to the correlation coefficient ≥ 0.7 as the advance data, and use the production data of the parts factory before normalization to correspond to the demand for multiple varieties of steel The data, as the exogenous variable X of the SARIMAX model, have:

In the formula,

is the demand value of multiple varieties of steel corresponding to the production data of the parts manufacturer, and β _n is an external variable

The regression coefficient of , r is the number of regression items of external variables; 43) Based on the demand data time series obtained after processing in step 2), use the grid search selected by the model to retrain and optimize to obtain different combinations of SARIMAX(p,d,q)(P,D,Q)s parameters, and the calculation is different. The AIC value of the combination, the parameter combination with the smallest AIC value is selected as the optimal parameter combination, and the final conditions for establishing the SARIMAX time series model are completed. 7. the multi-variety demand forecasting method for steel based on smart supply chain according to claim 5, is characterized in that, the specific content of step 6) is: Demand forecast results based on the SARIMA time model output obtained in step 3)

The revised demand forecast result output by the SARIMAX time series model obtained in step 5)

Get the final forecast

in

For the demand time series output using the SARIMAX time series model, remove the available SARIMAX time series model output results. 8. The multi-variety demand forecasting method for steel based on an intelligent supply chain according to claim 3, characterized in that, in step 12), taking twelve months as the law of a complete cycle, and screening out the data volume of a complete cycle Sort the top 90% of the data as the available data source. 9. The multi-variety demand forecasting method for steel based on an intelligent supply chain according to claim 4, wherein in step 22), the calculation formula of the standard deviation σ between the seasonal characteristic component and the random characteristic component is:

In the formula, Cs and Cr are the seasonal feature components and random feature components automatically decomposed from the demand data time series; Use the trend feature component Ct±standard deviation to obtain the upper and lower limits of the demand data time series values before each forecast, and adjust the abnormal values of all demand data time series within the program setting interval, namely [Ct-σ,Ct+σ] . 10. The multi-variety demand forecasting method for steel based on a smart supply chain according to claim 6, characterized in that, in step 42), screening out the data with correlation coefficient ≥ 0.7 and W _b is the leading data of W _a , W _a and W _b are the mapped data in the range of two sets of (0, 1) obtained in step 41).

Images