Information system of cigarette logistics distribution center based on combined forecasting method

Abstract

The present invention discloses a cigarette logistics distribution center information system based on a combined forecasting method, a cigarette demand combination forecasting model: a cigarette combination demand model of a time series is constructed according to influencing factors to forecast cigarette regional logistics demand; a cigarette industrial and commercial finished product pre-warehouse based on a logistics alliance: a logistics alliance integrating industrial and commercial enterprises is established, and a mathematical model for site selection is established to determine the most optimal site longitude and latitude according to the principle of lowest cost and best service, and a pre-warehouse for cigarette finished products of industrial enterprises is built in the area where the most optimal site is located to store some cigarettes with high market demand, and commercial enterprises provide warehousing and delivery services for the industry; an intelligent warehouse management system: based on RFID technology, electronic tags are associated with cigarette information, and precise intelligent management of cigarettes of different industries in the warehouse is implemented to realize storage and warehousing integration, and seamless connection with the No. 1 engineering system. The present invention improves industrial transportation efficiency, reduces commercial shortages, and enhances the timeliness of sorting and distribution.

Description

Technical Field

[0001] The invention relates to the field of tobacco logistics distribution, and in particular to a cigarette logistics distribution center information system based on a combined forecasting method. Background Art

[0002] Due to the actual needs of building a modern tobacco economic system and promoting high-quality development of the tobacco industry, it is necessary to explore and promote the construction of regional logistics distribution centers and delivery centers. Therefore, promoting regional logistics construction and integrating industrial and commercial logistics business is an inevitable trend for the high-quality development of logistics in the tobacco industry. At present, due to the uneven distribution of cigarette market demand, the lack of scientific guidance in logistics work, and the lack of effective connection between industrial and commercial enterprises, the full load rate of cigarette transportation vehicles of industrial enterprises is low, the time in transit is long, and the transportation cost is high; commercial enterprises have low inventory turnover and high capital occupation costs. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a cigarette logistics distribution center information system based on a combined forecasting method to solve the problems of uneven distribution of cigarette market demand, lack of scientific guidance in logistics work, lack of effective connection between industrial and commercial enterprises, low full load rate, long transit time and high transportation cost of cigarette transportation vehicles of industrial enterprises; low inventory turnover and high capital occupation cost of commercial enterprises.

[0004] To achieve the above object, the technical solution adopted by the present invention is: a cigarette logistics distribution center information system based on a combined prediction method, comprising

[0005] Cigarette demand combination forecasting model: Construct a time series cigarette combination demand model based on influencing factors to forecast cigarette regional logistics demand;

[0006] Pre-warehouse for finished industrial and commercial cigarette products based on logistics alliance: Establish a logistics alliance integrating industrial and commercial enterprises, establish a mathematical model for site selection based on the principle of lowest cost and best service to determine the longitude and latitude of the most optimal site, and build a pre-warehouse for finished cigarette products of industrial enterprises in the most optimal site to store some cigarettes with high market demand, and commercial enterprises provide warehousing and delivery services for industry;

[0007] Intelligent warehouse management system: Based on RFID technology, electronic tags are associated with cigarette information, and precise intelligent management of cigarettes from different industries in the warehouse is implemented to achieve integrated warehousing and outbound operations, and seamlessly connect with the No. 1 Engineering System; the No. 1 Engineering System is specifically the National Bureau's No. 1 Engineering System.

[0008] Preferably, the influencing factors include historical order data, seasonal changes, holidays, and prices.

[0009] Preferably, the process of constructing the cigarette demand combination prediction model comprises the following steps:

[0010] S1, Cigarette market capacity and demand analysis: Focusing on the macro environment and cigarette market, the overall cigarette market capacity and other influencing factors are comprehensively analyzed; the overall cigarette market capacity includes the market capacity of the permanent population, the gift-giving capacity of the permanent population, and the purchasing capacity of the tourist population; the other influencing factors include seasons and holidays;

[0011] S2, establish seasonal ARIMA model: establish an autoregressive moving average model for historical cigarette ordering data, and then introduce quarters and holidays into the influencing factors to establish a seasonal ARIMA model;

[0012] S3, establish a cigarette demand combination prediction model based on machine learning: on the basis of historical cigarette ordering data, construct seasonal characteristic values, integrate the characteristics of machine learning algorithms, construct a semi-supervised machine learning cigarette demand prediction model, and use a rolling single-step cycle prediction method to predict the monthly cigarette ordering volume.

[0013] Preferably, the overall market capacity includes

[0014] Permanent population market capacity: smoking angle = permanent population × smoking rate × average daily smoking amount × number of days living in the province;

[0015] Gift-giving capacity of permanent population: Gift consumption = number of smokers of appropriate age in permanent population × gift-giving rate × annual gift-giving volume;

[0016] Consumption capacity of smokers of appropriate age: Consumption of smokers of appropriate age = Number of people of appropriate age by age group × proportion of smokers × average amount of cigarettes smoked by smokers;

[0017] Market capacity of tourist population: purchasing angle = number of tourists × cigarette purchasing rate × cigarette purchasing volume.

[0018] Preferably, the permanent population is set as having lived there for more than 6 months, the age of the smokers of appropriate age is set as 18-64 years old, and the tourist population is set as having lived there for less than 1 month.

[0019] Preferably, the construction process of the seasonal ARIMA model comprises the following steps:

[0020] S1, sequence difference: select the monthly cigarette order data of a certain period of time for model fitting, establish a time-order quantity sequence chart, and perform stationarity test and seasonal difference on the data;

[0021] S2, stationarity test: Perform a stationarity test on the original data. When the p-value of the test result is less than 0.005, it proves that the original data is stationary and purely random. If the stationarity test fails, perform a difference calculation. When the p-value of the original data after the first-order difference is less than 0.005, the stationarity test is passed.

[0022] S3, seasonal decomposition: perform periodic decomposition on the original data with a periodicity of 12, extract seasonal indicators, form a de-periodic time series, and perform second-order differences on it to obtain first-order difference & second-order seasonal difference series diagrams, and determine that the order of d is 1 and 2 respectively;

[0023] S4, p, q value determination: select the autoregressive order p and the number of moving average items q of the seasonal ARIMA model, and determine them through the first-order difference and second-order seasonal autocorrelation coefficient ACF diagram and the partial autocorrelation coefficient PACF diagram; through analysis, the p and q values ​​are both determined to be (1,1), and the p, d, q values ​​are all determined, and it can be concluded that the cigarette demand forecasting model is ARIMA (1,1,1) (1,2,1), which is a seasonal ARIMA model.

[0024] Preferably, the process of constructing the cigarette demand combination prediction model based on machine learning includes the following steps:

[0025] S1, time matrix conversion: train and fit the model. For time series, assume that at different time points The observed values ​​can be used To indicate that if The observed value at time can be expressed as

[0026]

[0027] in , then explain the The observation value at a certain moment is a function of the observation values ​​at the previous k time points, and is predicted in a rolling manner according to the time series.

[0028] S2, data set cutting: machine learning is performed to divide the data into test set and training set. The training set is used to train the model, and the test set is used to judge the final training effect based on the training results of the training set;

[0029] S2, prediction results: After 14 iterations, the prediction model test results are obtained, the results in the figure are converted into a table and the error is calculated.

[0030] Preferably, the process of constructing the pre-warehouse for finished industrial and commercial cigarette products based on the logistics alliance includes the following steps:

[0031] S1, establish a forward warehouse location evaluation model: determine the candidate location of the forward warehouse through a combination of qualitative and quantitative methods, with the goal of minimizing the sum of storage cost, variable cost, transportation cost and cargo damage cost and maximizing service reliability, convert it into a mathematical formula, and establish a forward warehouse location mathematical model;

[0032] S2, determine the optimal site selection point: Based on the established mathematical model for site selection of the front warehouse, use Python language to program and solve, get the optimal solution of the objective function, get the longitude and latitude of the area where the optimal solution is located, use the map for visual conversion, and know the location corresponding to the longitude and latitude.

[0033] Preferably, the construction process of the pre-stock site selection evaluation model includes the following steps:

[0034] S1, decision variables and parameter definitions:

[0035] L—the set of industrial enterprises, L=(1,2,……,p);

[0036] I—the set of candidate points in the front-end library, I=(1,2,……,m);

[0037] J—the set of demand points of the logistics center, J=(1,2,……,n);

[0038] —Unit price of transportation from factory l to warehouse i;

[0039] —Transport volume from factory l to warehouse i;

[0040] —The distance from factory l to warehouse i;

[0041] —Demand d of each logistics center demand point j;

[0042] —Set the fixed cost of warehouse candidate node i;

[0043] —The maximum supply of warehouse candidate node i;

[0044] —Transportation mileage from warehouse candidate node i to logistics center warehouse j;

[0045] —Unit price from warehouse candidate node i to logistics center j;

[0046] —The quantity of materials that the warehouse candidate node i provides to the logistics center j;

[0047] — indicates the cost of building a position z by selecting the i candidate node;

[0048] —Select warehouse candidate point i as the warehouse, yes, ;otherwise ;

[0049] The transportation cost of each industrial company delivering goods to the forward warehouse is the product of the transportation unit price, transportation volume and mileage. The total cost T is expressed as:

[0050] (1)

[0051] Similarly, the total transportation cost Q from the forward warehouse to each logistics center is expressed as:

[0052] (2)

[0053] S2, mathematical modeling: Based on the above conditions, establish mathematical expressions

[0054] (3)

[0055] (4)

[0056] Explain the constraint group (4)

[0057] 1) Demand satisfaction of demand point j;

[0058] 2) The quantity of products provided by the forward warehouse shall not exceed the maximum capacity of the facility;

[0059] 3) The total transportation volume of each industrial enterprise to the front warehouse i is less than the maximum capacity of the facility;

[0060] 4) The delivery quantity from warehouse i to demand point j is not negative;

[0061] 5) There must be at least one alternative warehouse location.

[0062] Preferably, the intelligent warehouse management system includes

[0063] Basic management module: As the basic layer, it is mainly responsible for realizing basic information maintenance, including the maintenance of information on cargo owners, warehouses, storage areas, shelves, cargo locations and strategies;

[0064] Operation management module: divided into warehouse entry management, inventory management and outbound management. The warehouse entry management is used for warehouse entry order management, warehouse entry location allocation, and warehouse entry confirmation. The inbound management is used for inventory management, loss and surplus management, and location adjustment. The outbound management is used for outbound order management, outbound location allocation, and outbound confirmation.

[0065] Comprehensive analysis module: including inventory analysis, in-and-out inventory analysis, and inventory age analysis. The inventory analysis is used for inventory data query and handheld PDA inventory display, the in-and-out inventory analysis is used for in-warehouse data query and out-warehouse data query, and the inventory age analysis is used for inventory age screening and inventory age data query.

[0066] The beneficial effects of the present invention are:

[0067] 1. The present invention is based on the research on cigarette time series demand forecasting based on the combined forecasting method. It uses advanced forecasting methods such as machine learning and ARIMA to scientifically forecast and analyze changes in the cigarette market, realize independent and controllable industry forecasting methods, improve the industry's intelligent forecasting level, and dynamically grasp market changes.

[0068] 2. The present invention realizes the dynamic transformation of the industrial and commercial attributes of the forward warehouse through logistics alliance, resource sharing, joint warehouse construction, and industrial and commercial sharing and co-management mode of the forward warehouse, thereby improving industrial transportation efficiency, reducing commercial shortages, and enhancing the timeliness of sorting and distribution.

[0069] 3. The intelligent warehouse management system of the present invention realizes the intelligent management of the whole process of cigarettes from warehousing management, cargo space allocation to outbound confirmation, etc., ensuring the high integration and information sharing inside and outside the management information system, and improving the resource reuse rate; the open interface design ensures the strong adaptability of future system expansion and secondary development.

[0070] 4. The intelligent warehouse management system model of the present invention adopts a set of client systems in the front warehouse to connect to the "1 to N" mode of management machines of multiple industrial enterprises, which greatly simplifies the connection work with various industrial enterprises and eliminates the need to deploy independent client software and hardware separately. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a date order quantity sequence diagram of the present invention;

[0072] Figure 2 It is the first-order difference & second-order seasonal difference sequence diagram of the present invention;

[0073] Figure 3 It is the first-order seasonal difference autocorrelation ACF diagram of the present invention;

[0074] Figure 4 It is the first-order seasonal partial autocorrelation PACF diagram of the present invention;

[0075] Figure 5 It is the ACF diagram of the second-order seasonal difference autocorrelation of the present invention;

[0076] Figure 6 It is the second-order seasonal partial autocorrelation PACF diagram of the present invention;

[0077] Figure 7 This is a graph showing the ARIMA cigarette demand forecasting results taking seasonal effects into account in the present invention;

[0078] Figure 8 This is a graph showing the non-seasonal ARIMA cigarette demand forecasting results of the present invention;

[0079] Fig. 9Output graph of the test results of the present invention;

[0080] Fig.10 This is a diagram of the distribution model of the pre-warehouse of the industrial and commercial association of the present invention;

[0081] Fig.11 It is a diagram showing the optimal solution of the Python objective function of the present invention;

[0082] Fig.12 This is a functional architecture diagram of the intelligent warehouse management system of the present invention;

[0083] Fig.13 It is a schematic diagram of the system functional structure of the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings.

[0085] Example

[0086] This embodiment takes Sichuan tobacco market as an example. Fig.13 .

[0087] 1. Research on cigarette time series demand forecasting based on combined forecasting method

[0088] 1.1 Analysis of Cigarette Market Capacity and Demand in Sichuan Province

[0089] The premise of logistics demand forecasting is to accurately grasp the development of the cigarette market. The present invention comprehensively analyzes the cigarette market and influencing factors at two levels: macro environment and cigarette market.

[0090] 1) Overall cigarette market capacity in Sichuan Province

[0091] The cigarette market capacity in Sichuan Province consists of three parts:

[0092] ① Market capacity of permanent population:

[0093] Smoking angle = population × smoking rate × average daily smoking amount × number of days living in the province

[0094] Definition of permanent population: residency for more than 6 months;

[0095] ② Gift-giving capacity of permanent population:

[0096] Gift consumption = population aged 18-64 × gift giving rate × annual gift giving volume

[0097] Definition of permanent population: residency for more than 6 months;

[0098] Among them, those aged 18-64 are smokers of appropriate age. The calculation method for smokers of appropriate age is as follows:

[0099] Consumption of smokers of appropriate age = Number of people of appropriate age by age group × proportion of smokers × average amount of cigarettes smoked by smokers

[0100] ③Purchase angle = number of tourists × cigarette purchase rate × cigarette purchase volume

[0101] Definition of tourist population: The population whose residence time is less than 1 month is called tourist population, including those coming for business trip, entertainment, inspection, short-term work, etc.

[0102] After calculation and market deduction, it was determined that the province's cigarette capacity in 2020 will be 2.578 million boxes, with an error range of 2.565-2.591 million boxes. Among them, the cigarette capacity of the permanent population aged 18-64 is 2.127 million boxes, the cigarette capacity of the permanent population aged 64 and above is 329,000 boxes, and the cigarette consumption of the tourist population is 95,000 boxes.

[0103] 2) Other factors

[0104] Since cigarette sales may fluctuate significantly due to the proximity of holidays such as the Spring Festival and National Day, when forecasting market demand, in addition to considering market capacity, it is also necessary to focus on factors such as seasons and holidays.

[0105] 1.2 Cigarette demand forecast analysis based on time series forecasting - ARIMA model

[0106] The present invention analyzes the historical ordering data of cigarettes in Sichuan Province and finds that the data is a time series with linear characteristics. For this type of data, an autoregressive moving average model (ARIMA) is established, which has good fitting and prediction accuracy. Therefore, the present invention chooses to establish an ARIMA model to predict the cigarette demand market. However, since the historical ordering data of cigarettes in Sichuan Province has a strong seasonal characteristic, it is necessary to introduce seasonal influencing factors when modeling and establish a seasonal ARIMA model to enhance the model fitting and improve its prediction accuracy.

[0107] 1) Serial difference

[0108] The monthly cigarette ordering data from October 2006 to June 2021 in Sichuan Province is selected for model fitting to predict future cigarette ordering volume. First, the time-ordering volume sequence is established as follows: Figure 1 As shown in the figure, it can be seen that the sequence diagram contains obvious seasonal components and there are obvious changes. According to the above characteristics, the data are subjected to stationarity test and seasonal difference test respectively.

[0109] 2) Stationarity test

[0110] The ARIMA model requires that the sequence is a stationary sequence, so the original data must be tested for stationarity (white noise test). p When the value is less than 0.005, it proves that the data is stable and purely random. p The value is 0.01573855, which fails the stability test, so a difference calculation is required. After calculation, the order of the data after the first-order difference is p The value is 6.53619543×10 -14 , which can pass the stationarity test.

[0111] 3) Seasonal decomposition

[0112] The original data is decomposed into a periodicity of 12, and seasonal indicators are extracted to form a de-periodic time series. The second-order difference is performed on it to obtain the first-order difference & second-order seasonal difference series diagram as shown below: Figure 2 shown.

[0113] Therefore, it can be determined d The orders are 1 and 2 respectively.

[0114] 4) p,q Value determination

[0115] Choosing the autoregressive order for the seasonal ARIMA model p and the number of moving average items q , determined by the 1st-order difference and 2nd-order seasonal autocorrelation coefficient ACF diagram and partial autocorrelation coefficient PACF diagram, such as Figure 3 —6.

[0116] As can be seen from the figure, the graphs are all tailing, and through analysis p,q The values ​​are all set to (1,1). p, d, q The values ​​are all determined, and it can be concluded that the cigarette demand forecasting model is ARIMA (1,1,1) (1,2,1).

[0117] 5) Comparison of prediction results

[0118] Compare the forecast results of the ARIMA (1,1,1) (1,2,1) model taking into account seasonal factors with the ARIMA (1,1,1) model:

[0119] Table 1 Comparison of prediction results

[0120]

[0121] refer to Figure 7 —8. It can be clearly seen from the prediction results that the average absolute error of the ARIMA prediction model considering seasonal influencing factors can be reduced to 7%, and the prediction results are more accurate and have a higher degree of fit.

[0122] 1.3 Analysis of cigarette demand combination prediction model based on machine learning

[0123] The ARIMA prediction model can essentially only capture the linear relationship of data but not the nonlinear relationship. Therefore, in order to further improve the prediction accuracy, the present invention adopts a machine learning method to construct 72 characteristic values ​​such as the Spring Festival, year, and month based on the historical cigarette ordering data, integrates the characteristics of multiple machine learning algorithms, and constructs a semi-supervised machine learning cigarette demand prediction model. The rolling single-step cycle prediction method is used to predict the monthly cigarette ordering volume.

[0124] 1) Time matrix conversion

[0125] In the time series prediction model based on machine learning, it is usually necessary to train and fit the model. For time series, assuming that at different time points The observed values ​​can be used To indicate that if The observed value at time can be expressed as:

[0126]

[0127] in , then explain the The observed value at time is the value before k The time series rolling prediction form is shown in Table 2. The present invention uses the cycle period k The value is set to 12.

[0128] Table 2 Time series rolling forecast table

[0129]

[0130] The prediction of cigarette order quantity can be understood as using the first 12 input feature sequences to solve the target value for the 13th month, that is, the cigarette order quantity for the 13th month.

[0131] 2) Dataset cutting

[0132] Machine learning is performed to divide the data into a test set and a training set. The training set is used to train the model, and the test set is used to judge the final training effect based on the training results of the training set. There are 177 pieces of historical cigarette ordering data in Sichuan Province from October 2006 to June 2021. Considering that the model should be fully trained as much as possible to meet the prediction requirements, the first 171 pieces of data are selected as the training set, and the last 6 pieces of data are selected as the test set.

[0133] Training set: ;Training set labels:

[0134] Test set: ;Test set label:

[0135] Among them, for the training data, the dimension of a single sample is 12, which means that the data at the 13th time point is predicted using the data at the previous 12 adjacent time points.

[0136] 3) Prediction results

[0137] After 14 iterations, the prediction model test results are as follows: Fig. 9 As shown, the results in the figure are converted into a table and the error is calculated as shown in Table 3:

[0138] Table 3 Comparison of actual and predicted results

[0139]

[0140] After adjustment, the average MAE of the model is 5.2%. Compared with the seasonal ARIMA model, the cigarette forecasting model based on a combination of multiple machine learning methods has further improved its prediction accuracy, making the tobacco industry more scientific and accurate when conducting cigarette market analysis.

[0141] 2. Construction of pre-warehouses for finished products of cigarette industry and commerce in Sichuan

[0142] 2.1 Pre-storage Site Selection Evaluation Model

[0143] 2.1.1 Factors affecting the location of the pre-warehouse and its objectives

[0144] 1 Influencing factors

[0145] The factors affecting the location of the forward warehouse can be divided into cost factors and non-cost factors based on the key indicator of cost, as shown in Table 4:

[0146] Table 4 Various factors affecting the location of the pre-warehouse

[0147]

[0148] The entire process of logistics and transportation in the tobacco supply chain must strictly comply with national tobacco transportation standards. Finished cigarettes are transported by industrial companies to the logistics centers of commercial enterprises for warehousing, sorting, and outbound operations. On the other hand, some storage centers in Sichuan Province are idle due to factors such as aging equipment, and a large amount of management fees are still incurred every year. Therefore, in addition to considering traditional influencing factors such as the geographical distribution of tobacco logistics centers in various cities (prefectures) and transportation costs, special factors such as revitalizing assets and upgrading and reusing old factories must also be considered.

[0149] 2 Site selection objectives

[0150] In view of the above factors, the present invention focuses on the following objectives when selecting a site:

[0151] ① Minimize costs.

[0152] In the site selection process, the most important goal is to minimize the total cost of transportation and facilities. The total cost of facilities is often fixed, and the transportation cost is determined by the transportation quantity, transportation distance and transportation unit price. Among them, in order to achieve economies of scale, the transportation quantity must reach the transportation batch, which is conducive to reducing transportation costs; the transportation distance is determined by the logistics forward warehouse. Reasonable logistics warehouse site selection can reduce the total transportation distance and transportation costs; and the transportation unit price depends on the transportation method and transportation batch.

[0153] ②Maximize logistics volume

[0154] Maximizing logistics volume is achieved on the basis of minimizing logistics costs. The service area of ​​Sichuan Province's cigarette forward warehouses covers all cities (prefectures) in Sichuan Province. The service area is wide, the southeastern region has a more prosperous economy, greater customer demand, and greater logistics volume. The larger logistics volume reflects the higher processing capacity and operation level of the forward warehouse.

[0155] ③Maximize service reliability

[0156] The location of the forward warehouse meets the optimization of delivery time, delivery distance, delivery speed and timeliness, reflecting a higher service level.

[0157] ④High development potential

[0158] The service time of the forward warehouse is long and the construction cost is high. It will not be easily relocated or changed after completion. Therefore, in the process of site selection, the present invention fully considers the development potential of the surrounding environment, the level of economic development and the expansion of its own services.

[0159] 2.1.2 Pre-selected address and mathematical modeling

[0160] 1 Preselected address

[0161] In order to solve the problems of low loading rate of industrial transport vehicles and "small, scattered and urgent" urban orders, a forward warehouse is established as the mid-end node of the Sichuan tobacco regional logistics distribution network. Each industrial enterprise selects some high-selling cigarettes in Sichuan and transports them to the forward warehouse with full loads. When urgent orders are generated in various cities (prefectures), the forward warehouse will transport the cigarettes in the warehouse to the corresponding commercial logistics center. Fig.10 .

[0162] The geographical locations of the logistics centers and some branches of tobacco companies in 38 cities (states) in Sichuan Province were collected, and the Baidu map interface was called through Python to convert the geographical location to longitude and latitude. The complete table is shown in Appendix 4.3, and the part shown in the text is shown in Table 5:

[0163] Table 5 Latitude and longitude of tobacco logistics centers in various cities (prefectures) in Sichuan Province

[0164]

[0165] 2 Mathematical modeling

[0166] Through a combination of qualitative and quantitative methods, the alternative locations of the forward warehouse are determined. With the goal of minimizing the sum of warehousing cost, variable cost, transportation cost and cargo damage cost and maximizing service reliability, they are converted into mathematical formulas to establish a mathematical model for the location of the forward warehouse.

[0167] ① Decision variables and parameter definitions

[0168] L — an aggregation of industrial enterprises, L = (1, 2, ..., p );

[0169] I —A collection of candidate points in the front library, I = (1, 2, ..., m )

[0170] J —Logistics center demand point collection, J =(1,2,……, n );

[0171] —From the factory l To the warehouse i The unit price of transportation;

[0172] —From the factory l To the warehouse i The transport volume;

[0173] —From the factory l To the warehouse i Mileage;

[0174] —Demand points of each logistics center j Demand d ;

[0175] —Set warehouse alternative nodes i fixed costs;

[0176] —Warehouse alternative node i The maximum supply of

[0177] —Select nodes from the warehousei To the logistics center j Transport mileage to warehouses;

[0178] —Select nodes from the warehouse i To logistics center j Unit price;

[0179] —Warehouse alternative node i Warehouse to logistics center j The quantity of supplies;

[0180] — indicates choice i Cost of building a position at an alternative node z ;

[0181] -Select warehouse alternatives i For warehouse, yes, ;otherwise .

[0182] The transportation cost of each industrial company delivering goods to the forward warehouse is the product of the transportation unit price, transportation volume and mileage, and the total cost is T The expression is:

[0183] (1)

[0184] Similarly, the total transportation cost Q from the forward warehouse to each logistics center is expressed as:

[0185] (2)

[0186] ②Mathematical modeling

[0187] Based on the above conditions, establish the mathematical expression:

[0188] (3)

[0189] (4)

[0190] Explanation for constraint group (4):

[0191] 1) Demand points j Needs satisfaction;

[0192] 2) The quantity of products provided by the forward warehouse shall not exceed the maximum capacity of the facility;

[0193] 3) All industrial enterprises go to the front warehouse i The total transport volume is less than the maximum capacity of the facility;

[0194] 4) Warehousei To the demand point j The delivery volume is not negative;

[0195] 5) There must be at least one alternative warehouse location.

[0196] 3 Determination of the optimal location

[0197] According to the established mathematical model for site selection of the pre-storage warehouse, the optimal solution of the objective function is obtained by programming in Python. Fig.11 As shown, the longitude and latitude of the optimal solution area are (103.78, 30.84).

[0198] The characteristics of the forward warehouse construction of the present invention: the present invention combines industrial enterprises with commercial companies, and the industry transports the planned cigarettes to the forward warehouse in a centralized manner according to the sales plan formulated with the commercial companies through warehouse transfer. At the same time, the logistics distribution center will communicate with the industrial enterprises in a timely manner according to the inventory warning situation in the warehouse, so as to initiate activities such as warehouse distribution, replenishment, and warehouse transfer;

[0199] While sharing industrial and commercial inventory information, the forward warehouse of the present invention can also implement precise management of cigarettes from different industries, complete the integrated work of warehousing and warehousing, and seamlessly connect with the National Bureau's No. 1 Engineering System, so as to improve operating efficiency, reduce operating costs, realize intelligent management, standardize processes, visualize inventory and trace business.

[0200] 3. Intelligent warehouse management system

[0201] The intelligent warehouse management system mode of the present invention adopts a set of client systems in the front warehouse to connect to the "1 to N" mode of management machines of multiple industrial enterprises, which greatly simplifies the connection work with various industrial enterprises and no longer needs to deploy independent client software and hardware respectively; the system is based on RFID technology and uses the background database to associate the electronic tag with the cigarette piece information on the pallet to which the electronic tag is attached, so that the electronic tag corresponds to the cigarette piece information on the pallet, and the barcode information of the entire pallet of cigarette pieces can be obtained by collecting the data of the electronic tag.

[0202] For the cigarettes arriving, the mobile handheld terminal PDA is used to scan the "No. 1 Project" barcode of each cigarette arriving, and then the barcode data is written into the electronic tag of the corresponding pallet through the UHF ultra-high frequency RFID fixed reading and writing device. For the cigarettes arriving in a whole pallet, the electronic tag is scanned by the RFID device to scan the whole pallet into the warehouse; in the warehousing stage, the RFID mobile device is used to adjust the pallet information; in the sorting and outbound stage, the barcode data of the cigarettes is read from the electronic tag through the RFID device to complete the sorting and outbound scanning.

[0203] Intelligent warehouse management system includes basic management, operation management and comprehensive analysis. Fig.12 .

[0204] 3.1 Basic Management Module

[0205] The basic management module is the foundation layer of the intelligent warehouse management system, and is mainly responsible for maintaining basic information such as cargo owners, warehouses, storage areas, shelves and cargo locations.

[0206] Cargo owner information maintenance, mainly used to maintain cargo owner type, cargo owner code and cargo owner name;

[0207] Warehouse information maintenance mainly maintains key information such as warehouse name, storage capacity, type, etc.

[0208] Storage area information maintenance, mainly maintaining storage area type, whether the storage area is established, etc.

[0209] Shelf information maintenance: maintain the number of rows and columns on the shelves according to actual conditions;

[0210] Storage location information maintenance, mainly maintaining the storage of cigarettes, storage type, and maximum storage capacity of the storage location;

[0211] Strategic information maintenance mainly maintains the inventory entry and exit strategies and cargo location adjustment strategies, and establishes allocation ideas for automatic cargo location allocation.

[0212] 3.2 Job Management Module

[0213] The operation management module is divided into warehousing management, in-stock management and outbound management. It uses RFID technology, PDA handheld mobile terminals, telescopic carts and other tools to realize the intelligent management of the entire process from cigarette incoming order management, cargo space allocation, inventory management to outbound confirmation, greatly improving the efficiency of cigarette operation management and accelerating the high-quality development of its intelligent management equipment.

[0214] ①Warehouse management

[0215] The warehouse management system is connected with the industrial and commercial in-transit system and the industrial logistics system to obtain real-time information on industrial vehicles in transit so as to make preparations for harvesting in advance.

[0216] When industrial cigarettes are put into storage, the warehouse management system is connected to the "No. 1 Project" system of the National Bureau to obtain the industrial cigarette storage documents in real time, and automatically generate storage tasks based on the storage documents. The warehouse management system automatically allocates the cargo space according to the pre-set storage strategy, which can be adjusted manually. After the allocation is completed, the cargo space allocation data is sent to the handheld PDA, and the handheld PDA puts the cigarettes on the shelves according to the storage space allocation results. After the shelving is completed, click the "Storage Completed" button to return the shelving completion information to the warehouse system, increase the inventory of the corresponding owner in the warehouse, and complete the storage operation.

[0217] ②In-stock management

[0218] When taking inventory, the inventory process can be initiated through the warehouse system, the platform computer can view the inventory list information, and the inventory personnel can conduct mobile inventory through handheld PDAs. After the inventory is completed, the actual data will be sent back to the warehouse system, and the warehouse system will generate a difference order.

[0219] The warehousing system can generate profit and loss documents based on inventory difference documents and initiate profit and loss processes. At the same time, it can also directly initiate profit and loss processes to complete profit and loss business.

[0220] Due to the storage needs of cigarettes in the warehouse, the cargo position adjustment business can be carried out through the cargo position adjustment function. After the adjustment document is generated, it will be sent to the handheld PDA. After the adjustment is completed, click the "Adjustment Completed" button on the handheld PDA to feedback the adjustment completion information to the warehouse system, and the adjustment order process is completed.

[0221] ③Outbound management

[0222] When industrial cigarettes are shipped out, the warehouse management system is connected with the National Bureau's No. 1 Engineering System in real time to obtain industrial cigarette shipment documents (including transfer shipment and sales shipment) in real time, and automatically generate shipment operation tasks based on the shipment documents.

[0223] The warehouse management system automatically allocates cargo space according to the pre-set outbound delivery strategy, which can be adjusted manually. After the allocation is completed, the cargo space allocation data is sent to the handheld PDA, and the handheld PDA puts the cigarettes on the shelves according to the cargo space allocation results. After the shelving is completed, click the "Outbound Delivery Completed" button to return the off-shelf completion information to the warehouse system, reduce the inventory of the corresponding cargo owners in the warehouse, and complete the outbound operation.

[0224] 3.3 Comprehensive analysis

[0225] Based on daily operations and warehousing operation requirements, the forward warehouse intelligent warehousing management system needs to integrate relevant data and store data to facilitate industrial and commercial enterprises to grasp the flow of goods in real time, query data, and conduct specific analysis of specific data, further realize the digitization and visualization of inventory management, and strengthen the dynamic interconnection between industrial and commercial enterprises in the tobacco industry.

[0226] Improve the accuracy of delivery data collection, inventory data query and handheld PDA inventory display, you can query the inventory information of cigarettes in the warehouse for each consignor; through the handheld PDA, the storage status of cigarettes in the warehouse can be displayed in real time;

[0227] Ensure the standardization of cigarette management in warehouse, query the in-and-out data, query the in-warehouse cigarette information of each owner in the warehouse within a period of time, pass the in-warehouse order to the storage system, and automatically and dynamically feedback relevant information to the system;

[0228] It can realize the whole process trace management of cigarettes from industry to commercial enterprises, inventory age screening and inventory age data query, query the inventory age information of cigarettes in the warehouse, conduct inventory age screening, analyze cigarette turnover rate, and prepare for the selection, purchase and transportation of the next batch of cigarette specifications.

[0229] The above embodiments are used to illustrate the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. Cigarette logistics distribution center information system based on combined forecasting method, including: Cigarette demand combination forecasting model, cigarette industrial and commercial finished product pre-warehouse based on logistics alliance, intelligent warehouse management system; The cigarette demand combination forecasting model is specifically as follows: constructing a time series cigarette combination demand model according to influencing factors to forecast cigarette regional logistics demand; The above-mentioned pre-warehouse for finished industrial and commercial cigarette products based on logistics alliance is specifically as follows: establish a logistics alliance integrating industrial and commercial enterprises, establish a mathematical model for site selection to determine the longitude and latitude of the most optimal site based on the principle of lowest cost and best service, build a pre-warehouse for finished cigarette products of industrial enterprises in the most optimal site area to store some cigarettes with high market demand, and have commercial enterprises provide warehousing and delivery services for industry; The process of constructing the pre-warehouse for finished products of the cigarette industry and commerce based on the logistics alliance includes establishing a pre-warehouse site selection evaluation model and determining the optimal site selection point; The establishment of the forward warehouse location selection evaluation model is specifically as follows: by combining qualitative and quantitative methods, the candidate points of the forward warehouse are determined, and the minimum sum of storage cost, variable cost, transportation cost and cargo damage cost and the maximum service reliability are taken as the goal, and converted into a mathematical formula to establish a forward warehouse location selection mathematical model; the construction process of the forward warehouse location selection evaluation model includes the following steps: S1. Decision variables and parameter definitions: L—the set of industrial enterprises, L=(1,2,……,p); I—the set of candidate points in the front-end library, I=(1,2,……,m); J—the set of demand points of the logistics center, J=(1,2,……,n); —Unit price of transportation from factory l to warehouse i; —Transport volume from factory l to warehouse i; —The distance from factory l to warehouse i; —Demand d of each logistics center demand point j; —Set the fixed cost of warehouse candidate node i; —The maximum supply of warehouse candidate node i; —Transportation mileage from warehouse candidate node i to logistics center warehouse j; —Unit price from warehouse candidate node i to logistics center j; —The quantity of materials that the warehouse candidate node i provides to the logistics center j; — indicates the cost of building a position z by selecting the i candidate node; —Select warehouse candidate point i as the warehouse, yes, ;otherwise ; The transportation cost of each industrial company delivering goods to the forward warehouse is the product of the transportation unit price, transportation volume and mileage. The total cost T is expressed as: （1） Similarly, the total transportation cost Q from the forward warehouse to each logistics center is expressed as: （2） S2. Mathematical modeling: Based on the above conditions, establish mathematical expressions （3） （4） Explain the constraint group (4) 1) Demand satisfaction of demand point j; 2) The quantity of products provided by the forward warehouse shall not exceed the maximum capacity of the facility; 3) The total transportation volume of each industrial enterprise to the front warehouse i is less than the maximum capacity of the facility; 4) The delivery quantity from warehouse i to demand point j is not negative; 5) There must be at least one alternative warehouse location; The optimal site selection point is determined by: using the Python language to program and solve the established mathematical model for site selection of the front warehouse, obtaining the optimal solution of the objective function, obtaining the longitude and latitude of the area where the optimal solution is located, and using a map for visual conversion to obtain the location corresponding to the longitude and latitude; The intelligent warehouse management system is specifically: based on RFID technology, the electronic tag is associated with the cigarette information, and precise intelligent management of cigarettes from different industries in the warehouse is implemented to achieve integrated warehousing and outbound operations, and seamlessly connect with the No. 1 Engineering System; the No. 1 Engineering System is specifically the No. 1 Engineering System of the National Bureau.

2. The cigarette logistics distribution center information system based on combined forecasting method according to claim 1 is characterized in that: The influencing factors include historical order data, seasonal changes, holidays, and prices.

3. The cigarette logistics distribution center information system based on combined forecasting method according to claim 2 is characterized in that: The construction process of the cigarette demand combination prediction model includes the following steps: S1, Cigarette market capacity and demand analysis: Focusing on the macro environment and cigarette market, the overall cigarette market capacity and other influencing factors are comprehensively analyzed; the overall cigarette market capacity includes the market capacity of the permanent population, the gift-giving capacity of the permanent population, and the purchasing capacity of the tourist population; the other influencing factors include seasons and holidays; S2, establish seasonal ARIMA model: establish an autoregressive moving average model for historical cigarette ordering data, and then introduce quarters and holidays into the influencing factors to establish a seasonal ARIMA model; S3, establish a cigarette demand combination prediction model based on machine learning: on the basis of historical cigarette ordering data, construct seasonal characteristic values, integrate the characteristics of machine learning algorithms, construct a semi-supervised machine learning cigarette demand prediction model, and use a rolling single-step cycle prediction method to predict the monthly cigarette ordering volume.

4. The cigarette logistics distribution center information system based on combined forecasting method according to claim 3 is characterized in that: The overall market volume includes Permanent population market capacity: smoking angle = permanent population × smoking rate × average daily smoking amount × number of days living in the province; Gift-giving capacity of permanent population: Gift consumption = number of smokers of appropriate age in permanent population × gift-giving rate × annual gift-giving volume; Consumption capacity of smokers of appropriate age: Consumption of smokers of appropriate age = Number of people of appropriate age by age group × proportion of smokers × average amount of cigarettes smoked by smokers; Market capacity of tourist population: purchasing angle = number of tourists × cigarette purchasing rate × cigarette purchasing volume.

5. The cigarette logistics distribution center information system based on combined forecasting method according to claim 4 is characterized in that: The permanent population is set as those who have lived there for more than 6 months, the age of smokers is set as 18-64 years old, and the tourist population is set as those who have lived there for less than 1 month.

6. The cigarette logistics distribution center information system based on combined forecasting method according to claim 3 is characterized in that: The construction process of the seasonal ARIMA model includes the following steps: S1, sequence difference: select the monthly cigarette order data of a certain period of time for model fitting, establish a time-order quantity sequence chart, and perform stationarity test and seasonal difference on the data; S2, stationarity test: Perform a stationarity test on the original data. When the p-value of the test result is less than 0.005, it proves that the original data is stationary and purely random; If the stability test fails, a difference calculation is performed. If the p-value of the original data after the first-order difference is less than 0.005, the stability test is passed; S3, seasonal decomposition: perform periodic decomposition on the original data with a periodicity of 12, extract seasonal indicators, form a de-periodic time series, and perform second-order differences on it to obtain first-order difference & second-order seasonal difference series diagrams, and determine that the order of d is 1 and 2 respectively; S4, p, q value determination: select the autoregressive order p and the number of moving average items q of the seasonal ARIMA model, and determine them through the first-order difference and second-order seasonal autocorrelation coefficient ACF diagram and the partial autocorrelation coefficient PACF diagram; through analysis, the p and q values ​​are both determined to be (1,1), and the p, d, q values ​​are all determined, and it can be concluded that the cigarette demand forecasting model is ARIMA (1,1,1) (1,2,1), which is a seasonal ARIMA model.

7. The cigarette logistics distribution center information system based on combined forecasting method according to claim 3 is characterized in that: The process of constructing the cigarette demand combination prediction model based on machine learning includes the following steps: S1, time matrix conversion: train and fit the model. For time series, assume that at different time points The observed values ​​can be used To indicate that if The observed value at time can be expressed as in , then explain the The observation value at a certain moment is a function of the observation values ​​at the previous k time points, and is predicted in a rolling manner according to the time series. S2, data set cutting: machine learning is performed to divide the data into test set and training set. The training set is used to train the model, and the test set is used to judge the final training effect based on the training results of the training set; S2, prediction results: After 14 iterations, the prediction model test results are obtained, the results in the figure are converted into a table and the error is calculated.

8. The cigarette logistics distribution center information system based on combined forecasting method according to claim 1 is characterized in that: The intelligent warehouse management system includes Basic management module: As the basic layer, it is mainly responsible for realizing basic information maintenance, including the maintenance of information on cargo owners, warehouses, storage areas, shelves, cargo locations and strategies; Operation management module: divided into warehouse entry management, inventory management and outbound management. The warehouse entry management is used for warehouse entry order management, warehouse entry location allocation, and warehouse entry confirmation. The inbound management is used for inventory management, loss and surplus management, and location adjustment. The outbound management is used for outbound order management, outbound location allocation, and outbound confirmation. Comprehensive analysis module: including inventory analysis, in-and-out inventory analysis, and inventory age analysis. The inventory analysis is used for inventory data query and handheld PDA inventory display, the in-and-out inventory analysis is used for in-warehouse data query and out-warehouse data query, and the inventory age analysis is used for inventory age screening and inventory age data query.

Images