An intelligent management system and method for material warehousing data based on the Internet of Things

Abstract

The present invention discloses an intelligent management system and method for material warehousing data based on the Internet of Things, which relates to the field of data processing and analysis technology. The method includes the following steps: Step 1: Divide the warehouse into corresponding areas and record the capacity of each area; collect historical material entry and exit records and calculate the historical material usage frequency; Step 2: Combine the entry and exit records and date characteristics as input features to construct a time series model to predict future material demand; Step 3: Define regional standardized access costs; Based on the predicted demand and usage frequency, define the objective function and constraints; Step 4: Use a greedy heuristic algorithm to solve and plan the material storage allocation plan; Record and provide feedback when allocation is impossible. The present invention can effectively improve the situation in the existing technology where warehousing storage allocation lacks dynamic optimization and comprehensive consideration of multi-dimensional costs.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing and analysis, and in particular to an intelligent management system and method for material storage data based on the Internet of Things. Background Art

[0002] In modern logistics and supply chain systems, material warehouse management is a critical component in ensuring efficient production and distribution. With the development of technologies such as the Internet of Things and big data, traditional warehouse management models are gradually transforming towards intelligent, data-driven approaches. The core objectives of warehouse management are to optimize material storage layout, improve access efficiency, reduce operating costs, and ensure inventory accuracy and timeliness. Material storage allocation is a core management task. A sound storage strategy can reduce handling time and improve space utilization, directly impacting supply chain responsiveness and operational efficiency. The application of IoT technology has brought new breakthroughs to warehouse management. Using sensors, RFID tags, and positioning systems, data such as material location, quantity, and in / out times within the warehouse can be collected in real time, providing a foundation for intelligent decision-making. Furthermore, time series forecasting models can leverage historical data to identify demand patterns and accurately predict future material demand, incorporating factors such as holidays and promotional events.

[0003] Traditional material warehousing management methods rely heavily on manual experience and static rules, which present numerous limitations. For one thing, storage layouts lack the ability to dynamically adjust, making it difficult to adapt to fluctuations in material demand. For example, some companies employ fixed-area storage models, failing to consider differences in material usage frequency. This results in high-frequency materials being stored in remote locations, increasing transportation costs and operation time. Furthermore, demand forecasting accuracy is insufficient, often employing simple statistical methods that fail to fully incorporate dynamic factors such as date characteristics and market changes, easily leading to inventory backlogs or shortages. Furthermore, existing storage allocation schemes often only consider a single factor (such as storage space or transportation distance) when calculating costs, lacking a comprehensive assessment of multi-dimensional costs such as operation time and equipment utilization. This makes it difficult to achieve a globally optimal optimization result. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent management system and method for material storage data based on the Internet of Things to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a material storage data intelligent management method based on the Internet of Things, the method comprising the following steps:

[0006] Step 1: Divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records and calculate the historical frequency of material use;

[0007] Step 2: Combine the inbound and outbound records and date features as input features to build a time series model to predict future material demand;

[0008] Step 3: Define the standardized access cost for the region; based on the predicted demand and usage frequency, define the objective function and constraints;

[0009] Step 4: Use the greedy heuristic algorithm to solve and plan the material storage allocation plan; record and provide feedback when allocation is impossible.

[0010] In step 1, the warehouse is divided into n areas, denoted as: R1, R2, ..., R n ; The storage capacity corresponding to each area is expressed as: C1, C2, ..., C n ;

[0011] Material categories are represented as: T1, T2,…, T m ; m represents the number of material categories, T a represents the a-th material category, a∈{1,2,…,m};

[0012] Region R i The capacity is recorded as C i , represents the maximum storage capacity of the area; i∈{1,2,…,n};

[0013] Get the inbound and outbound records within the historical time period 1 to T, including the following fields: a ,timestamp,outbound / inbound,quantity};

[0014] For material T a , count the outbound quantity r in each time period a (t) and the incoming quantity s a (t); where t∈{1,2,…,T}; r a (t) represents the outbound quantity in the tth time period, s a (t) represents the incoming quantity in the tth time period; the outgoing quantity is marked as negative and the incoming quantity is marked as positive;

[0015] Calculate material T a Historical usage frequency F a :

[0016] .

[0017] Based on the warehouse layout and storage conditions, the warehouse space is divided into several areas, each with a fixed storage capacity limit. For example, some areas are suitable for storing bulky materials, while others are suitable for storing materials with high turnover. Next, historical material inbound and outbound data is retrieved, which includes information such as material type, inbound and outbound time, and quantity. By calculating the total number of inbound and outbound times and quantities for each material over a historical period, it is possible to reflect the frequency of use of that material in daily operations. Materials with high frequency of use should be prioritized in easily accessible areas during subsequent storage allocation, thereby improving overall warehouse operation efficiency.

[0018] In step 2, for material T a , taking r in the past T time periods a (t), s a (t), as well as holiday / promotion day labels and seasonal indicators as input features; an LSTM time series model is constructed for training;

[0019] The length of the forecast period is 1 time period; two sets of forecast sequences r are obtained a (t) T+1 and s a (t) T+1 , and summarized as the overall forecast demand D a =r a (t) T+1 +s a (t) T+1 ;

[0020] Among them, D a Indicates material T a Forecast demand;

[0021] Forecast each type of material to obtain the corresponding predicted demand.

[0022] By leveraging historical inbound and outbound records and incorporating date characteristics such as holidays, promotions, and seasonal changes, a long-short-term memory (LSTM) model is constructed to predict the demand quantity of each material over a specific time period, providing forward-looking data support for storage allocation. The model can predict the demand quantity of each material in the next time period, helping warehouse managers plan storage arrangements in advance and avoid inventory backlogs or shortages.

[0023] In step 3, for region R i , combined with region R i The average operation time per unit storage volume and the area R i The average transport distance to the exit, defining the area R i Standardized access cost e i ;

[0024] Through IoT devices such as sensors and RFID, real-time warehouse operation data is collected and recorded. i The time of each material access operation within the period P, the statistical area R i Total time T for material storage and retrieval operations t,i , and the total storage capacity S of the access materials in this area t,i ; Calculate the area R from this i Average operation time per unit storage volume t i :t i =T t,i / S t,i ;

[0025] Using the positioning technology of the Internet of Things (such as the positioning base station in the warehouse, the AGV car positioning system, etc.), obtain the area R i The distance from each material transport to the warehouse exit. Also in the time period P, the statistical area R i The total distance D from material handling to the exit t,i , and the number of transports N i , and then calculate the region R i Average transport distance to the exit d i :d i =D t,i / N i ;

[0026] In order to combine the average operation time and average transportation distance per unit storage volume, different weight coefficients are assigned to the two, and the weighted sum of the standardized average operation time and average transportation distance is obtained. i Standardized access cost e i :

[0027] Based on the forecast demand D a and frequency of use F a , define decision variables and objectives;

[0028] Decision variable: When material T a Stored in area R i When x ia =1; otherwise, x ia =0;

[0029] The objective function is to minimize the weighted access cost:

[0030] ;

[0031] Among them, F' a represents the normalized usage frequency; λ represents the frequency adjustment factor;

[0032] Constraints:

[0033] Regional capacity limit: The total forecast demand for materials allocated to each region does not exceed its maximum capacity;

[0034] ;

[0035] The inequality represents the region R i Store all allocated materials T a Total demand D a The sum of the two shall not exceed its upper capacity limit C i ;

[0036] Material full coverage constraint: Each material must and can only be assigned to one area:

[0037] .

[0038] In terms of optimization goals, the principle is to minimize overall storage and access costs. While also considering the frequency of material use, frequently used materials are stored in low-cost areas. Regarding constraints, the total amount of materials allocated to each area does not exceed its storage capacity, and each material can only be stored in one area, ensuring the feasibility and rationality of the allocation plan.

[0039] In step 4, a greedy heuristic algorithm is used to solve:

[0040] Prioritize materials: Calculate material priority score: S a =D a ·λF' a ; Press S a Sort from large to small and get the sequence [T (1) ,T (2) ,…,T (m) ];

[0041] Sort by regional cost: sort by regional access cost e i Sort from small to large and get the sequence [R (1) ,R (2) ,…,R (n) ];

[0042] Iterative allocation: Initially record the capacity C of each region i As the regional surplus capacity;

[0043] According to [T (1) ,T (2) ,…,T (m) ]Sequentially traverse the materials: for each material T (k) , according to the sorted region list [R (1) ,R (2) ,…,R (n) ] to traverse; when C (q) ≥D(k) and D (k) >0, allocate T (k) to R (q) , at this time, x (q)(k) =1; Update remaining capacity: set C (q) Update the value to C (q) -D (k) value; jump out of the current cycle and process the next material;

[0044] If all areas are traversed and still cannot be allocated, an alarm is triggered, a list of unallocated materials is recorded, and feedback is provided;

[0045] Among them, T (k) ∈{T (1) ,T (2) ,…,T (m)}, D (k) Indicates material T (k) The corresponding forecast demand; R (q) ∈{R (1) ,R (2) ,…,R (n)}, C (q) Indicates region R (q) The corresponding regional capacity.

[0046] Taking into account the projected demand and frequency of use of materials, a priority score is calculated for each material. Materials with higher priorities are those in high demand or frequently used and are therefore allocated first. Warehouse areas are also sorted by access cost, from lowest to highest. During the allocation process, starting with the highest-priority materials, attempts are made to place them in the areas with the lowest cost and sufficient capacity. If a material still cannot find suitable storage space after searching all areas, the material is recorded and an alert is issued for timely management action.

[0047] An intelligent management system for material warehousing data based on the Internet of Things, the system includes a warehousing division module, a demand forecasting module, an optimization modeling module, and a distribution scheduling module;

[0048] The storage division module is used to divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records, and calculate the historical frequency of material use;

[0049] The demand forecasting module is used to combine the inventory records and date features as input features to build a time series model to predict future material demand;

[0050] The optimization modeling module is used to define the standardized access cost of the region; based on the predicted demand and usage frequency, define the objective function and constraints;

[0051] The allocation and scheduling module is used to use a greedy heuristic algorithm to solve and plan material storage allocation plans; when allocation is impossible, it records and provides feedback.

[0052] The storage division module includes an area division unit and a frequency calculation unit;

[0053] The area division unit is used to physically divide the warehouse area and record the capacity parameters of each area;

[0054] The frequency calculation unit is used to collect statistics on historical in-and-out data and calculate the historical frequency of material use.

[0055] The demand forecasting module includes a data preprocessing unit and a model forecasting unit;

[0056] The data preprocessing unit is used to integrate historical in-and-out records and date features to construct a time series data set;

[0057] The model prediction unit is used to run the LSTM model to generate prediction requirements.

[0058] The optimization modeling module includes a cost definition unit and a constraint setting unit;

[0059] The cost definition unit is used to calculate the regional standardized access cost and normalized frequency;

[0060] The constraint setting unit is used to define the objective function and constraint conditions.

[0061] The allocation and scheduling module includes a sorting and allocation unit and an alarm feedback unit;

[0062] The sorting and allocation unit is used to execute a greedy algorithm to complete material allocation in order of priority and cost;

[0063] The alarm feedback unit is used to detect unallocated materials, trigger an alarm and provide feedback.

[0064] Compared with the existing technology, the beneficial effects of the present invention are as follows: the present invention can allocate materials to appropriate areas according to their actual demand and frequency of use through accurate demand forecasting and reasonable storage allocation plans, thereby avoiding waste of warehouse space and improving the storage capacity and space utilization of the warehouse; the optimized storage allocation plan of the present invention takes into account the access cost, and allocates materials with high frequency of use to areas with low access cost, reducing the time and labor costs in the material handling process, while also reducing inventory backlogs and loss costs caused by unreasonable storage, thereby reducing the cost of the entire warehousing operation; the present invention is based on real-time data collection and intelligent storage allocation decisions of Internet of Things devices, which can enable warehouse staff to find the required materials quickly and accurately, reduce search and handling time, improve the efficiency of material entry and exit operations, and thus improve the response speed of the entire supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a schematic diagram of the steps of an intelligent management method for material storage data based on the Internet of Things of the present invention;

[0066] Figure 2 This is a flow chart of an intelligent management system for material warehousing data based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0068] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a material storage data intelligent management method based on the Internet of Things, the method comprising the following steps:

[0069] Step 1: Divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records and calculate the historical frequency of material use;

[0070] Step 2: Combine the inbound and outbound records and date features as input features to build a time series model to predict future material demand;

[0071] Step 3: Define the standardized access cost for the region; based on the predicted demand and usage frequency, define the objective function and constraints;

[0072] Step 4: Use the greedy heuristic algorithm to solve and plan the material storage allocation plan; record and provide feedback when allocation is impossible.

[0073] In step 1, the warehouse is divided into n areas, denoted as: R1, R2, ..., R n ; The storage capacity corresponding to each area is expressed as: C1, C2, ..., C n ;

[0074] Material categories are represented as: T1, T2,…, T m ; m represents the number of material categories, T a represents the a-th material category, a∈{1,2,…,m};

[0075] Region R i The capacity is recorded as C i , represents the maximum storage capacity of the area; i∈{1,2,…,n};

[0076] Get the inbound and outbound records within the historical time period 1 to T, including the following fields: a ,timestamp,outbound / inbound,quantity};

[0077] For material T a , count the outbound quantity r in each time period a (t) and the incoming quantity s a (t); where t∈{1,2,…,T}; r a (t) represents the outbound quantity in the tth time period, s a (t) represents the incoming quantity in the tth time period; the outgoing quantity is marked as negative and the incoming quantity is marked as positive;

[0078] Calculate material T a Historical usage frequency F a :

[0079] .

[0080] Based on the warehouse layout and storage conditions, the warehouse space is divided into several areas, each with a fixed storage capacity limit. For example, some areas are suitable for storing bulky materials, while others are suitable for storing materials with high turnover. Next, historical material inbound and outbound data is retrieved, which includes information such as material type, inbound and outbound time, and quantity. By calculating the total number of inbound and outbound times and quantities for each material over a historical period, it is possible to reflect the frequency of use of that material in daily operations. Materials with high frequency of use should be prioritized in easily accessible areas during subsequent storage allocation, thereby improving overall warehouse operation efficiency.

[0081] In step 2, for material T a , taking r in the past T time periods a (t), s a(t), as well as holiday / promotion day labels and seasonal indicators as input features; an LSTM time series model is constructed for training;

[0082] The length of the forecast period is 1 time period; two sets of forecast sequences r are obtained a (t) T+1 and s a (t) T+1 , and summarized as the overall forecast demand D a =r a (t) T+1 +s a (t) T+1 ;

[0083] Among them, D a Indicates material T a Forecast demand;

[0084] Forecast each type of material to obtain the corresponding predicted demand.

[0085] By leveraging historical inbound and outbound records and incorporating date characteristics such as holidays, promotions, and seasonal changes, a long-short-term memory (LSTM) model is constructed to predict the demand quantity of each material over a specific time period, providing forward-looking data support for storage allocation. The model can predict the demand quantity of each material in the next time period, helping warehouse managers plan storage arrangements in advance and avoid inventory backlogs or shortages.

[0086] In step 3, for region R i , combined with region R i The average operation time per unit storage volume and the area R i The average transport distance to the exit, defining the area R i Standardized access cost e i ;

[0087] Through IoT devices such as sensors and RFID, real-time warehouse operation data is collected and recorded. i The time of each material access operation within the period P, the statistical area R i Total time T for material storage and retrieval operations t,i , and the total storage capacity S of the access materials in this area t,i ; Calculate the area R from this i Average operation time per unit storage volume t i :t i =T t,i / S t,i ;

[0088] Using the positioning technology of the Internet of Things (such as the positioning base station in the warehouse, the AGV car positioning system, etc.), obtain the area Ri The distance from each material transport to the warehouse exit. Also in the time period P, the statistical area R i The total distance D from material handling to the exit t,i , and the number of transports N i , and then calculate the region R i Average transport distance to the exit d i :d i =D t,i / N i ;

[0089] In order to combine the average operation time and average transportation distance per unit storage volume, different weight coefficients are assigned to the two, and the weighted sum of the standardized average operation time and average transportation distance is obtained. i Standardized access cost e i :

[0090] Based on the forecast demand D a and frequency of use F a , define decision variables and objectives;

[0091] Decision variable: When material T a Stored in area R i When x ia =1; otherwise, x ia =0;

[0092] The objective function is to minimize the weighted access cost:

[0093] ;

[0094] Among them, F' a represents the normalized usage frequency; λ represents the frequency adjustment factor;

[0095] Constraints:

[0096] Regional capacity limit: The total forecast demand for materials allocated to each region does not exceed its maximum capacity;

[0097] ;

[0098] The inequality represents the region R i Store all allocated materials T a Total demand D a The sum of the two shall not exceed its upper capacity limit C i ;

[0099] Material full coverage constraint: Each material must and can only be assigned to one area:

[0100] .

[0101] In terms of optimization goals, the principle is to minimize overall storage and access costs. While also considering the frequency of material use, frequently used materials are stored in low-cost areas. Regarding constraints, the total amount of materials allocated to each area does not exceed its storage capacity, and each material can only be stored in one area, ensuring the feasibility and rationality of the allocation plan.

[0102] In step 4, a greedy heuristic algorithm is used to solve:

[0103] Prioritize materials: Calculate material priority score: S a =D a ·λF' a ; Press S a Sort from large to small and get the sequence [T (1) ,T (2) ,…,T (m) ];

[0104] Sort by regional cost: sort by regional access cost e i Sort from small to large and get the sequence [R (1) ,R (2) ,…,R (n) ];

[0105] Iterative allocation: Initially record the capacity C of each region i As the regional surplus capacity;

[0106] According to [T (1) ,T (2) ,…,T (m) ]Sequentially traverse the materials: for each material T (k) , according to the sorted region list [R (1) ,R (2) ,…,R (n) ] to traverse; when C (q) ≥D (k) and D (k) >0, allocate T (k) to R (q) , at this time, x (q)(k) =1; Update remaining capacity: set C (q) Update the value to C (q) -D (k) value; jump out of the current cycle and process the next material;

[0107] If all areas are traversed and still cannot be allocated, an alarm is triggered, a list of unallocated materials is recorded, and feedback is provided;

[0108] Among them, T (k) ∈{T (1) ,T (2) ,…,T(m)}, D (k) Indicates material T (k) The corresponding forecast demand; R (q) ∈{R (1) ,R (2) ,…,R (n)}, C (q) Indicates region R (q) The corresponding regional capacity.

[0109] Taking into account the projected demand and frequency of use of materials, a priority score is calculated for each material. Materials with higher priorities are those in high demand or frequently used and are therefore allocated first. Warehouse areas are also sorted by access cost, from lowest to highest. During the allocation process, starting with the highest-priority materials, attempts are made to place them in the areas with the lowest cost and sufficient capacity. If a material still cannot find suitable storage space after searching all areas, the material is recorded and an alert is issued for timely management action.

[0110] An intelligent management system for material warehousing data based on the Internet of Things, the system includes a warehousing division module, a demand forecasting module, an optimization modeling module, and a distribution scheduling module;

[0111] The storage division module is used to divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records, and calculate the historical frequency of material use;

[0112] The demand forecasting module is used to combine the inventory records and date features as input features to build a time series model to predict future material demand;

[0113] The optimization modeling module is used to define the standardized access cost of the region; based on the predicted demand and usage frequency, define the objective function and constraints;

[0114] The allocation and scheduling module is used to use a greedy heuristic algorithm to solve and plan material storage allocation plans; when allocation is impossible, it records and provides feedback.

[0115] The storage division module includes an area division unit and a frequency calculation unit;

[0116] The area division unit is used to physically divide the warehouse area and record the capacity parameters of each area;

[0117] The frequency calculation unit is used to collect statistics on historical in-and-out data and calculate the historical frequency of material use.

[0118] The demand forecasting module includes a data preprocessing unit and a model forecasting unit;

[0119] The data preprocessing unit is used to integrate historical in-and-out records and date features to construct a time series data set;

[0120] The model prediction unit is used to run the LSTM model to generate prediction requirements.

[0121] The optimization modeling module includes a cost definition unit and a constraint setting unit;

[0122] The cost definition unit is used to calculate the regional standardized access cost and normalized frequency;

[0123] The constraint setting unit is used to define the objective function and constraint conditions.

[0124] The allocation and scheduling module includes a sorting and allocation unit and an alarm feedback unit;

[0125] The sorting and allocation unit is used to execute a greedy algorithm to complete material allocation in order of priority and cost;

[0126] The alarm feedback unit is used to detect unallocated materials, trigger an alarm and provide feedback.

[0127] In this example, an electronic product manufacturer's production process involves a variety of electronic components, assembly parts, and other materials. The warehouse is large, encompassing 10 distinct storage areas for different types of materials. The warehouse holds a wide variety of materials, totaling 50 types, and daily inbound and outbound operations are frequent. With the rapid growth of the business, traditional warehouse management methods are unable to meet the needs of efficient and accurate material storage and scheduling. Therefore, an IoT-based intelligent material storage data management method is introduced to optimize material storage allocation and reduce storage costs.

[0128] Step 1: Divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records and calculate the historical frequency of material use;

[0129] The warehouse is divided into 10 areas, denoted as R1, R2, ..., R 10 The storage capacity of each area is C1=1000 units, C2=800 units, C3=1200 units, C4=900 units, C5=1100 units, C6=700 units, C7=1300 units, C8=1000 units, C9=950 units, C 10 =850 units;

[0130] Material categories are represented by T1, T2, ..., T 50 , for example, T1 is a resistor, T2 is a capacitor, etc. Get the inbound and outbound records in the past 360 days (ie T=360), each record contains the fields: {Material Category T a,timestamp,outbound / inbound,quantity}.

[0131] Taking material T3 (a certain type of integrated circuit) as an example, its daily outbound quantity r3(t) and inbound quantity s3(t) are counted, where t∈{1,2,...,360}; r3(1)=-50 (indicating 50 units were outbound on the first day), s3(1)=80 (indicating 80 units were inbound on the first day), and the data for 360 days are counted in sequence.

[0132] Calculate the historical usage frequency F3 of material T3 using the formula. Perform similar statistics and calculations on all 50 materials to obtain the historical usage frequency of each material.

[0133] Step 2: Combine the inventory records and date features as input features to build a time series model to predict future material demand;

[0134] For material T3, the input features are r3(t), s3(t) in the past 360 days, holiday / promotion day marks (special periods such as the Spring Festival and Double Eleven are marked as 1, and ordinary periods are marked as 0), and seasonal indicators (different values ​​are set according to the peak and off-peak seasons of electronic product production).

[0135] Construct an LSTM time series model, set appropriate hyperparameters, such as the number of hidden layer neurons and learning rate, and use the collected data for training. The forecast period is set to 1 day, and the predicted outbound quantity sequence r3(13) and inbound quantity sequence s3(13) are obtained, which are summarized as the overall forecast demand D3=r3(13)+s3(13).

[0136] Similarly, the LSTM model is trained and predicted for the remaining 49 materials to obtain the predicted demand D1, D2, ..., D for each type of material. 50 ;

[0137] Step 3: Define the regional standardized access cost; based on the predicted demand and usage frequency, define the objective function and constraints;

[0138] For area R1, real-time warehouse operation data is collected through sensors and RFID in IoT devices. In the past day (period P), the total time T for material storage and retrieval operations in area R1 is calculated. t,1 =500 minutes, the total storage capacity of the material stored and retrieved in this area is S t,1 =2000 units, thus calculating the average operation time per unit storage volume in area R1, t1 = T t,1 / S t,1 =0.25 minutes / unit.

[0139] Using the positioning base station and AGV positioning system in the warehouse, the distance from each material transport to the exit in area R1 is obtained. On the same day, the total distance D of material transport from area R1 to the exit is calculated. t,1 =3000 meters, number of transports N1=150 times, and then calculate the average transport distance d1=D from area R1 to the exit. t,1 / N1=20 meters.

[0140] Assign a weight coefficient of 0.6 to the average operation time per unit storage volume and a weight coefficient of 0.4 to the average transportation distance. Take the weighted sum of the standardized average operation time and average transportation distance to obtain the standardized access cost e1 of region R1.

[0141] Similarly, calculate the standardized access costs e2, e3, ..., e of the remaining 9 regions. 10 .

[0142] Define decision variables: When material T a Stored in area R i When x ia =1; otherwise, x ia =0.

[0143] The historical usage frequency F of each material a Normalize and get F a ', frequency adjustment factor λ = 0.5. The objective function is to minimize the weighted access cost:

[0144] ;

[0145] The constraints are as follows:

[0146] Regional capacity restrictions: For region R1,

[0147] ,

[0148] That is, area R1 stores all allocated materials T a Total demand D a The sum of the total capacity of the area does not exceed its upper capacity limit C1. The same applies to other areas.

[0149] Material full coverage constraint: For material T3, it must and can only be allocated to one area, that is,

[0150] .

[0151] All materials must meet this condition.

[0152] Step 4: Use the greedy heuristic algorithm to solve and plan the material storage allocation plan; record and provide feedback when allocation is impossible;

[0153] Calculate the material priority score: For material T3, S3=D3λF'3, calculate the priority score for all 50 materials, and press S a Sort from large to small and get the sequence [T (1) ,T (2) ,...,T (50) ].

[0154] Access cost by region i Sort from small to large and get the sequence [R (1) ,R (2) ,...,R (10) ].

[0155] Initially record the capacity C of each area i As the remaining capacity of the area. According to [T (1) ,T (2) ,...,T (50) ]Sequentially traverse the materials, for material T (k) , according to the sorted region list [R (1) ,R (2) ,...,R (10) ] to traverse. Assume that the current material T (5) , its predicted demand D (5) = 100 units, when traversing to area R (3) When C (3) =1200 units, meeting C (3) >D (5) and D (5) >0, then assign T (5) to R (3) , at this time x (3)(5) =1, update R (3) The remaining capacity is C (3) -D (5) =1200-100=1100 units, then jump out of the current loop and process the next material.

[0156] If a material cannot be allocated after traversing all areas, such as material T (40) , the predicted demand is large and the remaining capacity of the existing area cannot meet it. At this time, an alarm is triggered and a list of unallocated materials (including T (40) ) and provide feedback to warehouse managers so that further measures can be taken, such as adjusting the warehouse layout or increasing storage capacity.

[0157] Through the implementation of the above-mentioned IoT-based intelligent management method for material warehousing data, the electronic product manufacturing company has achieved reasonable planning of material storage, effectively reduced warehousing costs, improved material storage and retrieval efficiency, and provided strong support for the company's efficient production and operation.

[0158] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A material storage data intelligent management method based on the Internet of Things, characterized by: The method comprises the following steps: Step 1: Divide the warehouse into corresponding areas and record the capacity of each area; collect historical material in and out records and calculate the historical frequency of material use; Step 2: Combine the inbound and outbound records and date features as input features to build a time series model to predict future material demand; Step 3: Define the standardized access cost for the region; based on the predicted demand and usage frequency, define the objective function and constraints; Step 4: Use the greedy heuristic algorithm to solve and plan the material storage allocation plan; record and provide feedback when allocation is impossible; In step 1, calculate material T a Historical usage frequency F a : ； Among them, T a represents the a-th material category; T represents the number of time periods; t∈{1,2,…,T} represents a certain time period; r a (t) represents the outbound quantity in the tth time period, s a (t) represents the incoming quantity in the tth time period; the outgoing quantity is marked as negative and the incoming quantity is marked as positive; In step 3, based on the forecast demand D a and frequency of use F a , define decision variables and objectives; decision variables: when material T a Stored in area R i When x ia =1; otherwise, x ia =0; The objective function is to minimize the weighted access cost: ； Among them, F'a represents the normalized usage frequency; λ represents the frequency adjustment factor; n represents the number of warehouse division areas; m represents the number of material categories; e i Indicates region R i Standardized access costs; Constraints: Regional capacity limit: The total forecast demand for materials allocated to each region does not exceed its maximum capacity; ； The inequality represents the region R i Store all allocated materials T a Total demand D a The sum of the two shall not exceed its upper capacity limit C i ; Material full coverage constraint: Each material must and can only be assigned to one area: ； In step 4, a greedy heuristic algorithm is used to solve: Prioritize materials: Calculate material priority score: S a =D a ·λF' a ; Press S a Sort from large to small and get the sequence [T (1) ,T (2) ,…,T (m) ]; Sort by regional cost: sort by regional access cost e i Sort from small to large and get the sequence [R (1) ,R (2) ,…,R (n) ]; Iterative allocation: Initially record the capacity C of each region i As the regional surplus capacity; According to [T (1) ,T (2) ,…,T (m) ]Sequentially traverse the materials: for each material T (k) , according to the sorted region list [R (1) ,R (2) ,…,R (n) ] to traverse; when C (q) ≥D (k) and D (k) >0, allocate T (k) to R (q) , at this time, x (q)(k) =1; Update remaining capacity: set C (q) Update the value to C (q) -D (k) value; jump out of the current loop and process the next material; if all areas are traversed and still cannot be allocated, trigger an alarm, record the list of unallocated materials and provide feedback; Among them, T (k) ∈{T (1) ,T (2) ,…,T (m) }, D (k) Indicates material T (k) The corresponding forecast demand; R (q) ∈{R (1) ,R (2) ,…,R (n) }, C (q) Indicates region R (q) The corresponding regional capacity.

2. The method for intelligent management of material warehousing data based on the Internet of Things according to claim 1, characterized in that: In step 1, the warehouse is divided into n areas, denoted as: R1, R2, ..., R n ; The storage capacity corresponding to each area is expressed as: C1, C2, ..., C n ; Material categories are represented as: T1, T2,…, T m ; m represents the number of material categories, a∈{1,2,…,m}; Region R i The capacity is recorded as C i , represents the maximum storage capacity of the area; i∈{1,2,…,n}; Get the inbound and outbound records within the historical time period 1 to T, including the following fields: a , timestamp, outbound / inbound, quantity}; for material T a , count the outbound quantity r in each time period a (t) and the incoming quantity s a (t).

3. The method for intelligent management of material warehousing data based on the Internet of Things according to claim 2, characterized in that: In step 2, for material T a , taking r in the past T time periods a (t), s a (t), as well as holiday / promotion day flags and seasonal indicators as input features; Build an LSTM time series model for training; The length of the forecast period is 1 time period; two sets of forecast sequences r are obtained a (t) T+1 and s a (t) T+1 , and summarized as the overall forecast demand D a =r a (t) T+1 +s a (t) T+1 Among them, D a Indicates material T a Forecast demand; Forecast each type of material to obtain the corresponding predicted demand.

4. The method for intelligent management of material storage data based on the Internet of Things according to claim 3, characterized in that: In step 3, for region R i , combined with region R i The average operation time per unit storage volume and the area R i The average transport distance to the exit, defining the area R i Standardized access cost e i .

5. An intelligent management system for material warehousing data based on the Internet of Things, applied to the intelligent management method for material warehousing data based on the Internet of Things according to any one of claims 1 to 4, characterized in that: The system includes a warehouse division module, a demand forecasting module, an optimization modeling module and a distribution scheduling module; The warehouse partitioning module is used to divide the warehouse into corresponding areas and record the capacity of each area; collect historical material entry and exit records and calculate the historical material usage frequency; the demand forecasting module is used to combine the entry and exit records and date features as input features to construct a time series model to predict future material demand; the optimization modeling module is used to define regional standardized access costs; based on the predicted demand and usage frequency, define the objective function and constraints; the allocation scheduling module is used to use a greedy heuristic algorithm to solve and plan the material storage allocation plan; and record and feedback are provided when allocation is impossible.

6. The material storage data intelligent management system based on the Internet of Things according to claim 5 is characterized by: The storage division module includes an area division unit and a frequency calculation unit; The area division unit is used to physically divide the warehouse area and record the capacity parameters of each area; the frequency calculation unit is used to count historical in-and-out warehouse data and calculate the historical usage frequency of materials.

7. The material storage data intelligent management system based on the Internet of Things according to claim 6 is characterized by: The demand forecasting module includes a data preprocessing unit and a model forecasting unit; The data preprocessing unit is used to integrate historical in-and-out warehouse records and date features to construct a time series data set; the model prediction unit is used to run the LSTM model to generate prediction requirements.

8. The material storage data intelligent management system based on the Internet of Things according to claim 7 is characterized by: The optimization modeling module includes a cost definition unit and a constraint setting unit; The cost definition unit is used to calculate the regional standardized access cost and normalized frequency; the constraint setting unit is used to define the objective function and constraint conditions.

9. The material storage data intelligent management system based on the Internet of Things according to claim 8, characterized in that: The allocation and scheduling module includes a sorting and allocation unit and an alarm feedback unit; The sorting and allocation unit is used to execute a greedy algorithm to complete material allocation in order of priority and cost; the alarm feedback unit is used to detect unallocated materials, trigger an alarm and provide feedback.

Images