CN113780655A - Steel multi-variety demand prediction method based on intelligent supply chain - Google Patents
Steel multi-variety demand prediction method based on intelligent supply chain Download PDFInfo
- Publication number
- CN113780655A CN113780655A CN202111049815.5A CN202111049815A CN113780655A CN 113780655 A CN113780655 A CN 113780655A CN 202111049815 A CN202111049815 A CN 202111049815A CN 113780655 A CN113780655 A CN 113780655A
- Authority
- CN
- China
- Prior art keywords
- data
- demand
- steel
- model
- production
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 88
- 239000010959 steel Substances 0.000 title claims abstract description 88
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000004519 manufacturing process Methods 0.000 claims abstract description 120
- 241000728173 Sarima Species 0.000 claims abstract description 44
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000012937 correction Methods 0.000 claims abstract description 5
- 230000001932 seasonal effect Effects 0.000 claims description 36
- 238000012545 processing Methods 0.000 claims description 14
- 238000013277 forecasting method Methods 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 9
- 238000005096 rolling process Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 239000000463 material Substances 0.000 description 3
- 230000008520 organization Effects 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000000556 factor analysis Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Databases & Information Systems (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Entrepreneurship & Innovation (AREA)
- Manufacturing & Machinery (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- General Factory Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method for predicting the multi-variety demand of steel based on an intelligent supply chain, which comprises the following steps: 1) acquiring a time sequence of multi-variety demand data of steel based on industrial product production data of an intelligent supply chain system; 2) constructing a SARIMA time series model based on the time series of the demand data of various steel products; 3) inputting a training sample set to be tested to a SARIMA time sequence model, and obtaining the multi-variety demand forecasting result of steel; 4) constructing a SARIMAX time sequence model based on the time sequence of the demand data of multiple varieties of steel; 5) inputting a test sample set to be tested to a SARIMAX time sequence model, and obtaining the demand prediction result of the steel products of multiple varieties after algorithm correction; 6) and obtaining a final prediction result according to the result output by the SARIMA model and the correction result output by the SARIMAX model. Compared with the prior art, the method has the advantages of more stable time sequence, higher prediction accuracy and the like.
Description
Technical Field
The invention relates to the technical field of supply chain production management, in particular to a method for predicting multi-variety requirements of steel based on an intelligent supply chain.
Background
With the development of the industrial internet, the downstream supply chain in the steel industry is also transformed and upgraded, the production organization mode of the main raw material steel needs to be changed, and the steel production enterprises need to change from the rigid production organization mode with fixed production pins to the flexible production organization mode with balanced production pins. Information intercommunication with a downstream supply chain enterprise industrial product manufacturing process is needed, effective technical means are combined, the rule of industrial product production data is mined, the multi-variety demand of steel of a downstream supply chain in the steel industry is predicted in advance for a certain period in the future, the steel production efficiency and the capacity resource utilization rate of the steel production enterprise are improved, and the steel consumption demand of the downstream supply chain is met more quickly and timely.
However, the technical means of analyzing, processing and deciding the demands of the downstream supply chain is laggard at present, and the demands of corresponding steel varieties are predicted mainly by means of manual experience or simple mathematical formulas for industrial product production data of downstream supply chain enterprises collected in the intelligent supply chain. The steel mill is difficult to reasonably and efficiently arrange production plans of different steel varieties based on the current demand prediction, so that the production efficiency and the utilization rate of capacity resources are difficult to improve, and the technical problem to be improved and solved is urgent. In other fields, when demand prediction is carried out through various big data prediction technical means, non-technical problems such as enterprise profit, cost conditions and the like are more considered. The current demand forecasting technique and method is long in time consumption, poor in systematicness, low in integrity, low in accuracy and slow in response and output speed.
In conclusion, the existing method for predicting the multi-variety demand of the steel cannot improve the production efficiency and the utilization rate of capacity resources, and cannot meet the intelligent linkage from the production of industrial products of an intelligent supply chain to the multi-variety demand storage and production scheduling of the steel of an iron and steel production enterprise.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a method for predicting the multi-variety demand of steel based on an intelligent supply chain.
The purpose of the invention can be realized by the following technical scheme:
a multi-variety demand forecasting method for steel based on an intelligent supply chain comprises the following steps:
s1: based on industrial product production data of the intelligent supply chain system, the time sequence of the demand data of multiple varieties of steel is obtained. The industrial product production data of the intelligent supply chain system comprises production BOM data, host production plant production data and part production plant production data. The specific steps of acquiring the time series of the demand data of multiple varieties of steel products comprise:
11) taking the obtained production data of the industrial product in the intelligent supply chain system as an original data time sequence;
12) cleaning an original data time sequence based on a SARIMA model technical framework, adjusting a data date format, filling missing data by using a moving average method, setting a threshold number of months, filling data with the number more than the threshold number of consecutive missing months to be zero, and screening out data with the first 90% of data volume in a certain period as an available data source;
13) establishing a functional relation W between the production data and the demand data according to the production BOM data, the production data of the host production factory and the production data of the part production factoryt＝g(P1,P2,...,Pn) Wherein W istFor the demand values of different steel varieties in different months, P1,P2,...,PnSolving the production data of the intelligent supply chain industrial products for the production values of different industrial products by utilizing a functional relation and converting the production data into corresponding time series of the demand data of multiple steel products;
14) and (3) performing moving average processing on the multi-variety demand data time series of the steel obtained in the step 13) to obtain the sample data length used by the rolling prediction of the demand prediction model and the preliminary stable demand data time series.
S2: and constructing a SARIMA time series model based on the time series of the demand data of various steel products.
The method comprises the following specific steps:
21) based on seasonal regularity, a SARIMA time series model is constructed:
in the formula, Wt、wt-n、wt-snRespectively the required values of different steel varieties in the months of t, t-n and t-sn, mu is a constant term, epsilont、εt-n、εt-snThe method comprises the steps of calculating the required values of different steel varieties in months t, t-n and t-sn respectively, wherein P is the number of trend autoregressive terms, Q is the number of trend moving average terms, P is the number of seasonal autoregressive terms, and Q is the number of seasonal moving average terms. Alpha is alphanIs a trend autoregressive coefficient, θnIs a trend moving average coefficient, phinIs a seasonal autoregressive coefficient, etanIs a seasonal moving average coefficient.
22) Calculating the standard deviation between the seasonal characteristic component and the random characteristic component according to the results of the three characteristic items obtained by model decomposition, and adjusting abnormal values of all the required data time sequences in a set interval; the standard deviation σ between the seasonal and random feature components is calculated as:
in the formula, Cs and Cr are seasonal characteristic components and random characteristic components which are automatically decomposed from a demand data time sequence respectively;
and (3) obtaining upper and lower limits of the demand data time sequence value before each prediction by using the trend characteristic component Ct +/-standard deviation, and adjusting all abnormal values of the demand data time sequence within a program set interval, namely [ Ct-sigma, Ct + sigma ].
23) Different combinations of SARIMA (P, D, Q) (P, D, Q) s parameters are obtained by utilizing the grid search training and optimization selected by a model, D is the trend difference times for leading the SARIMA (P, D, Q) (P, D, Q), D is the seasonal difference times for leading the SARIMA (P, D, Q) (P, D, Q) s is the time step number in a single season period, P is the trend autoregressive term number, Q is the trend moving average term number, P is the seasonal autoregressive term number, and Q is the seasonal moving average term number; calculating the AIC values of different combinations, selecting the parameter combination with the minimum AIC value as the optimal parameter combination, and finishing the final condition of establishing the SARIMA time series model.
S3: inputting a training sample set to be tested to the constructed SARIMA time series model, and obtaining the demand prediction results of multiple varieties of steel.
S4: and constructing a SARIMAX time series model based on the time series of the required data of the steel products of multiple varieties in S1.
The method comprises the following specific steps:
41) based on the production data of the host production plant and the time series of the demand data of the multiple varieties of steel corresponding to the production data of the part production plant obtained in the step S1, normalization processing is carried out on the production data to obtain two groups of data mapped in the range of (0, 1);
42) calculating the data correlation coefficients of the two groups of mappings by utilizing correlation analysis, screening out data corresponding to the correlation coefficient being more than or equal to 0.7 as antecedent data, and taking the demand data of multiple varieties of steel corresponding to the production data of the part production plant before normalization processing as an exogenous variable X of the SARIMAX model, wherein the data comprises the following steps:
in the formula,
the requirement value of multiple varieties of steel corresponding to the production data of the part production plant, betanAs an external variable
R is the number of regression terms of the external variable;
43) based on the required data time sequence obtained after the processing of S2, different combinations of SARIMAX (P, D, Q) (P, D, Q) S parameters are obtained by training again and optimizing through the grid search selected by the model, AIC values of the different combinations are calculated, the parameter combination with the minimum AIC value is selected as the optimal parameter combination, and the final condition of establishing the SARIMAX time sequence model is completed.
S5: inputting a training sample set to be tested to the constructed SARIMAX time sequence model, and obtaining the demand prediction result of the steel products of multiple varieties after algorithm correction.
S6: and acquiring a final prediction result according to the demand prediction result output by the SARIMA time series model obtained in the step S3 and the corrected demand prediction result output by the SARIMAX time series model obtained in the step S5.
The concrete contents are as follows:
demand prediction result output based on SARIMA time model obtained in S3
And the modified demand prediction result output by the SARIMAX time series model obtained in S5
Obtaining the final prediction result
Wherein
The required time series of results are output for use of the SARIMA time series model, excluding the available SARIMAX time series model.
Further, in step 12), with a rule that twelve months are taken as a complete cycle, screening out data with the top 90% of the data quantity ranking of the complete cycle as an available data source.
Further, in step 42), the correlation coefficient W not less than 0.7 is screened outbIn is WaPreceding data of (2), Wa、WbThe two sets of (0, 1) range mapped data obtained in step 41).
Compared with the prior art, the intelligent supply chain-based steel multi-variety demand prediction method provided by the invention at least has the following beneficial effects:
1) according to the method, the sample data length used by the rolling prediction of the demand prediction model and the preliminary stable demand data time sequence can be obtained by utilizing the rule of industrial product production data in an intelligent supply chain and a targeted data processing method, the demand data time sequence is more stable, and the accuracy of prediction is improved;
2) the SARIMA and SARIMAX time sequence models are established, the time sequence running data, namely the non-stationary time sequence, of the multi-variety required data of steel can be predicted, the non-stationary time sequence can be converted into the stationary time sequence, the influence of seasonal factors is considered, the effective characteristics of the model can be formed by excavating natural rules existing among production data of different industrial products, the algorithm efficiency and the complexity of data source searching and correlation factor analysis dimensionality are reduced, and the prediction result with high accuracy and engineering application can be obtained based on the time sequence;
3) by the method for predicting the multi-variety demand of the steel based on the intelligent supply chain, the multi-variety demand of the steel in a certain period of time in the future is predicted timely, accurately and stably, and the technical problem that the production efficiency and the utilization rate of capacity resources are difficult to improve because a steel mill cannot arrange production plans of different steel varieties reasonably and efficiently is solved.
Drawings
FIG. 1 is a schematic flow chart illustrating an embodiment of a method for predicting a demand of a plurality of varieties of steel based on an intelligent supply chain.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
Examples
As shown in FIG. 1, the present invention relates to a method for predicting the demand of multiple varieties of steel based on an intelligent supply chain, which comprises the following steps:
step one, acquiring a time sequence of the demand data of multiple varieties of steel.
Acquiring production data of industrial products from an intelligent supply chain system, wherein the production data comprises time series of original data such as BOM (bill of material) production data, production data of a host production plant, production data of a part production plant and the like; and (3) performing data preprocessing such as cleaning, screening, converting and sliding averaging on the original data time sequence based on a SARIMA model technical framework to obtain the length of the sample data of multiple varieties of steel used for rolling prediction of the demand prediction model and the required stable demand data time sequence. Specifically, the method comprises the following steps:
step 101: acquiring production data of industrial products in an intelligent supply chain system, wherein the production data comprises time series of original data such as BOM (bill of material) production data, host production plant production data, part production plant production data and the like;
step 102: cleaning an original data time sequence based on a SARIMA model technical framework, adjusting a data date format by using a program, filling missing data by using a moving average method, setting a threshold value to be 6 months, continuously missing data for more than 6 months, and filling 0, in order to improve the stability of the time sequence, and screening out 90% of data quantity in the first order of 1 complete cycle as an available data source by combining with the rule that the production data of industrial products takes 12 months as 1 complete cycle;
step 103: establishing a functional relationship between the production data and the demand data, W, based on the BOM (bill of materials) data, the production data of the host manufacturing plant, and the production data of the component manufacturing plantt＝g(P1,P2,...,Pn) Solving and converting the production data of the intelligent supply chain industrial products into corresponding time series of the demand data of multiple steel products, wherein WtIs made of steelDemand values of different varieties in different months, P1,P2,...,PnProduction values for different industrial products;
step 104: and (4) performing 3-month moving average processing on the multi-variety demand data time series of the steel obtained in Step103 to obtain the sample data length used by the rolling prediction of the demand prediction model and the preliminary stable demand data time series.
And step two, establishing a SARIMA time series model based on the demand data time series obtained in the step one.
Continuously adjusting and optimizing by using a SARIMA algorithm model, and automatically decomposing trend characteristic components, seasonal characteristic components and random characteristic components from a required data time sequence; calculating the standard deviation between seasonal characteristic components and random characteristic components by using a program according to the results of the three characteristic items obtained by model decomposition, obtaining the upper limit and the lower limit of a time sequence value of the demand data by using the standard deviation of the trend characteristic components +/-and adjusting abnormal values of all the time sequences of the demand data within a program setting interval, and avoiding the influence of extreme demand data of individual months on the overall data rule; different combinations of SARIMA (P, D, Q) x (P, D, Q) parameters are obtained by utilizing the grid search training and optimization selected by the model, the AIC (Red pool information criterion) values of the different combinations are calculated by utilizing a program, the parameter combination with the minimum AIC value is selected as the optimal parameter combination, and the final condition for establishing the SARIMA time sequence model is completed. The method comprises the following specific steps:
step 201: because the production data of the intelligent supply chain industrial products related to the consumed steel varieties has obvious seasonal regularity, namely the production in summer and winter is greater than that in spring and autumn, the demand data of multiple steel varieties has seasonal changes, and a SARIMA time series model can be constructed:
in the formula, Wt、wt-n、wt-snRespectively the required values of different steel varieties in the months of t, t-n and t-sn, mu is a constant term, epsilont、εt-n、εt-snThe method comprises the steps of calculating the required values of different steel varieties in months t, t-n and t-sn respectively, wherein P is the number of trend autoregressive terms, Q is the number of trend moving average terms, P is the number of seasonal autoregressive terms, and Q is the number of seasonal moving average terms. Alpha is alphanIs a trend autoregressive coefficient, θnIs a trend moving average coefficient, phinIs a seasonal autoregressive coefficient, etanIs a seasonal moving average coefficient.
By utilizing the constitution, automatic adjustment and optimization of the SARIMA algorithm model, a trend characteristic component, a seasonal characteristic component and a random characteristic component are automatically decomposed from a time sequence of the demand data and are respectively marked as Ct, Cs and Cr.
Step 202: and calculating the standard deviation between the seasonal characteristic component and the random characteristic component by using a program according to the results of the three characteristic items obtained by model decomposition, wherein the calculation formula is as follows:
in order to avoid the influence of extreme demand data of individual months on the overall data rule, the trend characteristic component plus or minus standard deviation is used for obtaining the upper limit and the lower limit of the demand data time sequence value before each prediction, and the abnormal values of all the demand data time sequences are adjusted in the program setting interval, namely [ Ct-sigma, Ct + sigma ].
Step 203: different combinations of SARIMA (P, D, Q) (P, D, Q) s parameters are obtained by utilizing the grid search training and optimization selected by the model, D is the trend difference times for enabling the SARIMA (P, D, Q) (P, D, Q) s parameters to become the final stable sequence, D is the seasonal difference times for enabling the SARIMA (P, D, Q) parameters to become the final stable sequence, and s is the time step number in a single seasonal period. And calculating the AIC (Red pool information criterion) values of different combinations by using a program, selecting the parameter combination with the minimum AIC value as the optimal parameter combination, and finishing the final condition for establishing the SARIMA time series model.
And step three, selecting data with the time period length of [ T-24, T ] as a training sample set based on the SARIMA time series model obtained in the step two and the final stable demand data time series obtained by processing in the step one and the step two, inputting the training sample set to be tested, and performing rolling prediction on the demand of the steel products of the T +2 months (30-60 days in the future) to obtain the demand prediction result of the steel products of various varieties.
Step four, because the production data of the host production plant and the production data of the part production plant have obvious natural rules, the production data of the part production plant always leads the change of the production data of the host production plant, and by utilizing the rules, the time series of the demand data of various steel products corresponding to the production data of the host production plant and the production data of the part production plant are obtained based on the step one, and the normalization processing is carried out on the time series of the demand data to obtain two groups of data mapped in the range of (0, 1); and calculating data correlation coefficients of the two groups of mappings by utilizing correlation analysis, screening an exogenous variable X which can be used as a SARIMAX model, searching, training and optimizing by utilizing grids selected by the model to obtain different combinations of SARIMAX (P, D, Q) (P, D, Q) s parameters based on the SARIMA algorithm model in the second step and the third step, calculating AIC (Chichi information criterion) values of the different combinations by utilizing a program, selecting a parameter combination with the minimum AIC value as an optimal parameter combination, and finishing the final condition for establishing the SARIMAX time sequence model. The method comprises the following specific steps:
step 401: based on the time sequence of the multi-variety demand data of the steel corresponding to the production data of the host production plant and the production data of the part production plant obtained in the first step, normalization processing is carried out on the time sequence to obtain two groups of (0, 1) range mapping data which are respectively marked as Wa、Wb；
Step 402: calculating data correlation coefficient of two groups of mappings by using correlation analysis, and screening out the correlation coefficient which is not less than 0.7 and WbIn is WaThe antecedent data of (2) is obtained by using, as exogenous variables X of the SARIMAX model, demand data of multiple steel varieties corresponding to production data of the parts manufacturing plant before the normalization, and includes:
wherein,
the requirement value of multiple varieties of steel corresponding to the production data of the part production plant, betanAs an external variable
R is the number of regression terms of the external variable;
step 403: based on the time sequence of the demand data obtained after the processing in the second step, in order to avoid the influence on the combination selection of SARIMAX (P, D, Q) (P, D, Q) s parameters after an exogenous variable X is added, different combinations of the SARIMAX (P, D, Q) (P, D, Q) s parameters are obtained by training again and optimizing through the grid search selected by the model, the AIC (Red pool information criterion) values of the different combinations are calculated through a program, the parameter combination with the minimum AIC value is selected as the optimal parameter combination, and the final condition for establishing the SARIMAX time sequence model is completed.
And step five, selecting data with the time period length of [ T-24, T ] as a training sample set based on the SARIMAX time sequence model obtained in the step four and the final steady demand data time sequence obtained by processing in the step one and the step two, inputting the training sample set to be tested, predicting the demand of the steel products of T +2 months (60 days in the future), and obtaining the demand prediction result of the steel products of multiple varieties after algorithm correction.
Step six, based on the demand prediction result output by the SARIMA model obtained in the step three
And a corrected demand prediction result based on the SARIMAX model output obtained in step five
The final predicted result
Wherein
For outputting by using SARIMA model,The required time series of results that can be output with the SARIMAX model are removed.
The method can obtain the sample data length used by the rolling prediction of the demand prediction model and the preliminary stable demand data time sequence by utilizing the rule of industrial product production data in the intelligent supply chain and a targeted data processing method, and the demand data time sequence is more stable, thereby being beneficial to improving the accuracy of prediction. The SARIMA and SARIMAX time sequence models are established, the time sequence running data, namely the non-stationary time sequence, of the multi-variety required data of steel can be predicted, the non-stationary time sequence can be converted into the stationary time sequence, the influence of seasonal factors is considered, the effective characteristics of the model can be formed by excavating natural rules existing among production data of different industrial products, the algorithm efficiency and the complexity of data source searching and correlation factor analysis dimensionality are reduced, and the prediction result with high accuracy and engineering application can be obtained based on the time sequence.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and those skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A multi-variety demand forecasting method for steel based on an intelligent supply chain is characterized by comprising the following steps:
1) acquiring a time sequence of multi-variety demand data of steel based on industrial product production data of an intelligent supply chain system;
2) constructing a SARIMA time series model based on the time series of the demand data of various steel products;
3) inputting a training sample set to be tested to a constructed SARIMA time series model to obtain the demand prediction results of multiple varieties of steel;
4) constructing a SARIMAX time series model based on the time series of the demand data of the steel products of multiple varieties in the step 1);
5) inputting a training sample set to be tested to a constructed SARIMAX time sequence model, and obtaining the demand prediction result of the steel products of multiple varieties after algorithm correction;
6) and acquiring a final prediction result according to the demand prediction result output by the SARIMA time model obtained in the step 3) and the corrected demand prediction result output by the SARIMAX time series model obtained in the step 5).
2. The intelligent supply chain based steel product multi-item demand forecasting method as claimed in claim 1, wherein the industrial product production data of the intelligent supply chain system includes BOM production data, mainframe production plant production data and part production plant production data.
3. The intelligent supply chain-based steel product multi-variety demand forecasting method as claimed in claim 2, wherein the specific step of obtaining the time series of the demand data of the steel products multi-variety includes:
11) taking the obtained production data of the industrial product in the intelligent supply chain system as an original data time sequence;
12) cleaning an original data time sequence based on a SARIMA model technical framework, adjusting a data date format, filling missing data by using a moving average method, setting a threshold number of months, filling data with the number more than the threshold number of consecutive missing months to be zero, and screening out data with the first 90% of data volume in a certain period as an available data source;
13) establishing a functional relation W between the production data and the demand data according to the production BOM data, the production data of the host production factory and the production data of the part production factoryt＝g(P1,P2,...,Pn) Wherein W istFor the demand values of different steel varieties in different months, P1,P2,...,PnSolving the production data of the intelligent supply chain industrial products for the production values of different industrial products by utilizing a functional relation and converting the production data into corresponding time series of the demand data of multiple steel products;
14) and (3) performing moving average processing on the multi-variety demand data time series of the steel obtained in the step 13) to obtain the sample data length used by the rolling prediction of the demand prediction model and the preliminary stable demand data time series.
4. The intelligent supply chain-based steel multi-variety demand forecasting method as claimed in claim 1, wherein the specific steps of step 2) include:
21) constructing a SARIMA time series model based on seasonal rules;
22) calculating the standard deviation between the seasonal characteristic component and the random characteristic component according to the results of the three characteristic items obtained by model decomposition, and adjusting abnormal values of all the required data time sequences in a set interval;
23) different combinations of SARIMA (P, D, Q) (P, D, Q) s parameters are obtained by utilizing the grid search training and optimization selected by a model, D is the trend difference times for leading the SARIMA (P, D, Q) (P, D, Q), D is the seasonal difference times for leading the SARIMA (P, D, Q) (P, D, Q) s is the time step number in a single season period, P is the trend autoregressive term number, Q is the trend moving average term number, P is the seasonal autoregressive term number, and Q is the seasonal moving average term number; calculating the AIC values of different combinations, selecting the parameter combination with the minimum AIC value as the optimal parameter combination, and finishing the final condition of establishing the SARIMA time series model.
5. The intelligent supply chain-based steel multi-variety demand forecasting method as claimed in claim 4, wherein in step 21), the expression of the SARIMA time series model is constructed as follows:
in the formula, Wt、wt-n、wt-snRespectively the required values of different steel varieties in the months of t, t-n and t-sn, mu is a constant term, epsilont、εt-n、εt-snThe error of the required values of different steel varieties in the months of t, t-n and t-sn respectively, P is the number of trend autoregressive terms, Q is the number of trend moving average terms, P is the number of seasonal autoregressive terms, Q is the number of seasonal moving average terms, alphanIs a trend autoregressive coefficient, θnIs a trend moving average coefficient, phinIs a seasonal autoregressive coefficient, etanIs a seasonal moving average coefficient.
6. The intelligent supply chain-based steel multi-variety demand forecasting method as claimed in claim 5, wherein the specific steps of step 4) include:
41) normalizing the production data of the host production plant and the required data time series of multiple steel varieties corresponding to the production data of the part production plant obtained in the step 1) to obtain two groups of data mapped within the range of (0, 1);
42) calculating the data correlation coefficients of the two groups of mappings by utilizing correlation analysis, screening out data corresponding to the correlation coefficient being more than or equal to 0.7 as antecedent data, and taking the demand data of multiple varieties of steel corresponding to the production data of the part production plant before normalization processing as an exogenous variable X of the SARIMAX model, wherein the data comprises the following steps:
in the formula,
the requirement value of multiple varieties of steel corresponding to the production data of the part production plant, betanAs an external variable
R is the number of regression terms of the external variable;
43) based on the required data time sequence obtained after the processing of the step 2), different combinations of SARIMAX (P, D, Q) (P, D, Q) s parameters are obtained by training again and optimizing through the grid search selected by the model, AIC values of the different combinations are calculated, the parameter combination with the minimum AIC value is selected as the optimal parameter combination, and the final condition of establishing the SARIMAX time sequence model is completed.
7. The method for predicting the multi-variety demand of steel based on the intelligent supply chain as claimed in claim 5, wherein the specific contents of the step 6) are as follows:
predicting the result of demand output based on the SARIMA time model obtained in step 3)
And the corrected demand prediction result output by the SARIMAX time series model obtained in the step 5)
Obtaining the final prediction result
Wherein
The required time series of results are output for use of the SARIMA time series model, excluding the available SARIMAX time series model.
8. The intelligent supply chain-based steel multi-variety demand forecasting method as claimed in claim 3, wherein in the step 12), the data of the top 90% of the data quantity ranking of a complete cycle is selected as the available data source according to the rule that twelve months are taken as a complete cycle.
9. The intelligent supply chain-based steel multi-item demand forecasting method according to claim 4, wherein in step 22), the standard deviation σ between the seasonal feature component and the random feature component is calculated by:
in the formula, Cs and Cr are seasonal characteristic components and random characteristic components which are automatically decomposed from a demand data time sequence respectively;
and (3) obtaining upper and lower limits of the demand data time sequence value before each prediction by using the trend characteristic component Ct +/-standard deviation, and adjusting all abnormal values of the demand data time sequence within a program set interval, namely [ Ct-sigma, Ct + sigma ].
10. The intelligent supply chain-based steel multi-variety demand forecasting method as claimed in claim 6, wherein in step 42), the correlation coefficient W is selected to be not less than 0.7bIn is WaPreceding data of (2), Wa、WbThe two sets of (0, 1) range mapped data obtained in step 41).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111049815.5A CN113780655A (en) | 2021-09-08 | 2021-09-08 | Steel multi-variety demand prediction method based on intelligent supply chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111049815.5A CN113780655A (en) | 2021-09-08 | 2021-09-08 | Steel multi-variety demand prediction method based on intelligent supply chain |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113780655A true CN113780655A (en) | 2021-12-10 |
Family
ID=78841798
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111049815.5A Pending CN113780655A (en) | 2021-09-08 | 2021-09-08 | Steel multi-variety demand prediction method based on intelligent supply chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113780655A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116720630A (en) * | 2023-08-08 | 2023-09-08 | 浪潮云洲工业互联网有限公司 | Coal mine raw coal yield prediction method, equipment and medium based on time sequence |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040102937A1 (en) * | 2002-11-21 | 2004-05-27 | Honeywell International Inc. | Energy forecasting using model parameter estimation |
| US20090327033A1 (en) * | 2008-06-26 | 2009-12-31 | Xerox Corporation | Methods and systems for forecasting inventory levels in a production environment |
| US20150227859A1 (en) * | 2014-02-12 | 2015-08-13 | The Procter & Gamble Company | Systems and methods for creating a forecast utilizing an ensemble forecast model |
| US20180302291A1 (en) * | 2017-04-14 | 2018-10-18 | Accenture Global Solutions Limited | Comparative multi-forecasting analytics service stack for cloud computing resource allocation |
| US20200134640A1 (en) * | 2018-10-26 | 2020-04-30 | Target Brands, Inc. | Method and system for generating ensemble demand forecasts |
| KR20210014432A (en) * | 2019-07-30 | 2021-02-09 | 주식회사 마크에이트 | Method for accurate estimating demand of Raw Materials Using System for estimating demand of Raw Materials |
| KR20210014431A (en) * | 2019-07-30 | 2021-02-09 | 주식회사 마크에이트 | System for determining orders ranking of Raw Materials using demand estimate and Method thereof |
| CN112669065A (en) * | 2020-12-24 | 2021-04-16 | 润联软件系统(深圳)有限公司 | Demand prediction method and device for fast-moving products, computer equipment and storage medium |
| US20210224677A1 (en) * | 2020-01-21 | 2021-07-22 | International Business Machines Corporation | Forecasting model generation for time series data with change point and seasonality |
-
2021
- 2021-09-08 CN CN202111049815.5A patent/CN113780655A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040102937A1 (en) * | 2002-11-21 | 2004-05-27 | Honeywell International Inc. | Energy forecasting using model parameter estimation |
| US20090327033A1 (en) * | 2008-06-26 | 2009-12-31 | Xerox Corporation | Methods and systems for forecasting inventory levels in a production environment |
| US20150227859A1 (en) * | 2014-02-12 | 2015-08-13 | The Procter & Gamble Company | Systems and methods for creating a forecast utilizing an ensemble forecast model |
| US20180302291A1 (en) * | 2017-04-14 | 2018-10-18 | Accenture Global Solutions Limited | Comparative multi-forecasting analytics service stack for cloud computing resource allocation |
| US20200134640A1 (en) * | 2018-10-26 | 2020-04-30 | Target Brands, Inc. | Method and system for generating ensemble demand forecasts |
| KR20210014432A (en) * | 2019-07-30 | 2021-02-09 | 주식회사 마크에이트 | Method for accurate estimating demand of Raw Materials Using System for estimating demand of Raw Materials |
| KR20210014431A (en) * | 2019-07-30 | 2021-02-09 | 주식회사 마크에이트 | System for determining orders ranking of Raw Materials using demand estimate and Method thereof |
| US20210224677A1 (en) * | 2020-01-21 | 2021-07-22 | International Business Machines Corporation | Forecasting model generation for time series data with change point and seasonality |
| CN112669065A (en) * | 2020-12-24 | 2021-04-16 | 润联软件系统(深圳)有限公司 | Demand prediction method and device for fast-moving products, computer equipment and storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116720630A (en) * | 2023-08-08 | 2023-09-08 | 浪潮云洲工业互联网有限公司 | Coal mine raw coal yield prediction method, equipment and medium based on time sequence |
| CN116720630B (en) * | 2023-08-08 | 2023-12-22 | 浪潮云洲工业互联网有限公司 | Coal mine raw coal yield prediction method, equipment and medium based on time sequence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109376953B (en) | Medium-and-long-term electric quantity prediction method and system | |
| CN108846530B (en) | Short-term load prediction method based on clustering-regression model | |
| CN107146035B (en) | Method for calculating batch coefficients in large-goods production of knitted clothes | |
| CN106529704A (en) | Monthly maximum power load forecasting method and apparatus | |
| CN112801388B (en) | Power load prediction method and system based on nonlinear time series algorithm | |
| CN110689195A (en) | Power daily load prediction method | |
| CN106980910B (en) | Medium-and-long-term power load measuring and calculating system and method | |
| CN110909958A (en) | Short-term load prediction method considering photovoltaic grid-connected power | |
| CN113506121A (en) | Analysis method and device for price influence factors | |
| CN110852496A (en) | Natural gas load prediction method based on LSTM recurrent neural network | |
| CN115860797A (en) | Electric quantity demand prediction method suitable for new electricity price reform situation | |
| CN110866658A (en) | Method for predicting medium and long term load of urban power grid | |
| CN105184078A (en) | Technology maturity evaluation method based on patent relative-quantity analysis | |
| CN114626640A (en) | Natural gas load prediction method and system based on characteristic engineering and LSTM neural network | |
| CN110533249B (en) | Metallurgical enterprise energy consumption prediction method based on integrated long-term and short-term memory network | |
| CN113780655A (en) | Steel multi-variety demand prediction method based on intelligent supply chain | |
| CN114266408A (en) | Power distribution network engineering operation and maintenance cost optimization method and system | |
| CN112330030B (en) | System and method for predicting requirements of expansion materials | |
| CN113673842A (en) | Production efficiency diagnosis method, device, equipment and storage medium based on big data | |
| CN117277312A (en) | Gray correlation analysis-based power load influence factor method and equipment | |
| CN112241804A (en) | Macroscopic economy leading index construction method and system for energy power data | |
| CN112115620B (en) | Learning curve product unit cost evaluation method based on fractal thought | |
| CN112734159B (en) | Method and system for calculating reworking/reproduction rate of enterprise | |
| Kharkyanen et al. | Development of information technology for supporting the process of adjustment of the food enterprise assortment | |
| CN114529072A (en) | Regional electric quantity prediction method based on time series |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |