Commodity wholesale data management method and device based on Internet of Things, and medium

Abstract

The invention discloses a commodity wholesale data management method and device based on the Internet of Things, and a medium, and relates to the technical field of commodity operation, and the method comprises the steps: binding a commodity quality attenuation coefficient with static information, generating a dynamic inventory list, collecting an Internet of Things sensing signal data packet of a commodity downstream distribution node, and storing the data packet in the dynamic inventory list; the method comprises the steps of obtaining a dynamic inventory list, performing association analysis with the dynamic inventory list, generating a differential replenishment demand instruction, taking the differential replenishment demand instruction as a scheduling target, constructing a single-target mathematical programming model in combination with the dynamic inventory list, solving the single-target mathematical programming model by adopting a meta-heuristic optimization algorithm, and obtaining a commodity scheduling scheme. After the commodity scheduling scheme is converted into a specific scheduling instruction, the specific scheduling instruction is issued to a corresponding warehouse management center for execution; according to the commodity wholesale management method driven by the data, the problems of stock overstock and stockout caused by quality monitoring deficiency and demand response lag in commodity wholesale management are solved, and the overall operation efficiency and the asset value keeping level of a supply chain are improved.

Description

Technical Field

[0001] This invention relates to the field of commodity operation technology, and in particular to a commodity wholesale data management method, device and medium based on the Internet of Things. Background Technology

[0002] With the rapid development of the Internet of Things (IoT), traditional wholesale management models are gradually evolving towards intelligent and data-driven approaches. In the modern supply chain system, the wholesale link, as a key node connecting production and retail, directly impacts overall logistics costs and the quality of commodity circulation due to its management efficiency. In recent years, IoT architectures based on RFID, sensor networks, edge computing, and big data analytics have been widely applied in warehousing and logistics monitoring systems, enabling real-time perception of commodity location, quantity, and environmental conditions. Some advanced warehouse management centers have introduced temperature and humidity monitoring, inventory early warning, and automated replenishment mechanisms. By digitizing data from the physical world, they have improved inventory visibility and response speed. Meanwhile, the application of artificial intelligence algorithms, especially machine learning models, in demand forecasting enables companies to conduct trend analysis based on historical sales data and optimize replenishment strategies.

[0003] Nevertheless, existing wholesale management methods still have room for improvement. Relying solely on fixed-period inventory checks fails to establish a quantitative model for the dynamic degradation of product quality due to storage conditions. It also ignores the cumulative damage to product value caused by temperature and humidity, leading to undetected losses of high-value, fragile goods and resulting in economic losses. Secondly, demand forecasting at downstream distribution nodes is mostly based on historical sales or manual experience, lacking the integration of real-time operational data to generate differentiated and dynamically responsive replenishment instructions, which can easily lead to inventory backlogs and stockouts. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an Internet of Things-based method for managing wholesale commodity data to solve the problems of commodity inventory backlog and stockouts.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a commodity wholesale data management method based on the Internet of Things, comprising,

[0008] Collect environmental parameters and static information of the microenvironment in which goods are stored;

[0009] The quality decay function is used to compare and integrate environmental parameters with stored thresholds in static information to obtain the product quality decay coefficient.

[0010] By binding the product quality decay coefficient with static information, a dynamic inventory list is generated.

[0011] Collect IoT sensing signal data packets from downstream distribution nodes of goods, and perform correlation analysis with dynamic inventory lists to generate differentiated replenishment demand instructions;

[0012] Using differentiated replenishment demand instructions as the scheduling objective, a single-objective mathematical programming model is constructed by combining a dynamic inventory list;

[0013] A metaheuristic optimization algorithm is used to solve a single-objective mathematical programming model to obtain a commodity scheduling scheme.

[0014] After the commodity scheduling plan is converted into specific scheduling instructions, it is sent to the corresponding warehouse management center for execution.

[0015] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the environmental parameters of the commodity storage microenvironment include temperature and humidity;

[0016] The static information includes product category identifier, warehousing time, warehouse location, product quantity, procurement cost, product identifier, and storage threshold.

[0017] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of using a quality decay function to compare and integrate environmental parameters with stored thresholds in static information to obtain a commodity quality decay coefficient specifically involves:

[0018] The environmental parameters are compared with the storage threshold to obtain the deviation between the environmental parameters and the storage threshold;

[0019] The deviation between environmental parameters and storage thresholds is integrated over time. The result of the integration is then substituted into the quality decay function as an exponential term to obtain the product quality decay coefficient.

[0020] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of generating a dynamic inventory list by binding the commodity quality decay coefficient with static information refers to creating independent dynamic inventory records for all commodities using commodity identifiers, aggregating the dynamic inventory records of all commodities, and forming a dynamic inventory list.

[0021] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of collecting IoT sensing signal data packets from downstream distribution nodes of the commodity and performing correlation analysis with a dynamic inventory list to generate differentiated replenishment demand instructions specifically includes:

[0022] Collect IoT sensing signal data packets from IoT devices at downstream distribution nodes of goods;

[0023] Extract signal features from IoT sensing signal data packets to obtain feature values ​​of truck entry and exit frequency and shelf inventory depth;

[0024] Input the product quality decay coefficient, truck entry and exit frequency, and shelf inventory depth feature values ​​from the dynamic inventory list into the GBDT model to predict demand and obtain differentiated replenishment demand instructions.

[0025] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of constructing a single-objective mathematical programming model by combining differentiated replenishment demand instructions as the scheduling target with a dynamic inventory list specifically includes:

[0026] Decision variables are defined based on differentiated replenishment demand instructions, and total logistics and transportation costs are calculated based on these decision variables.

[0027] The decision variables include the volume of various commodities transferred from the source warehouse to the target warehouse and the truck routes that carry out the transfer tasks;

[0028] The total inventory value preservation rate is obtained by weighted summing of the product quality decay coefficients in the dynamic inventory list.

[0029] Using differentiated replenishment demand instructions as constraints, and with the objectives of minimizing total logistics costs and maximizing the preservation of total inventory value, a single-objective mathematical programming model is established.

[0030] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of using a metaheuristic optimization algorithm to solve a single-objective mathematical programming model to obtain a commodity scheduling scheme specifically includes:

[0031] The decision variables of the single-objective mathematical programming model are encoded as chromosomes in the metaheuristic optimization algorithm, and the total logistics transportation cost and inventory value preservation of each chromosome are calculated.

[0032] For each chromosome, the total logistics transportation cost and inventory value preservation rate are used to calculate chromosome fitness using a non-dominated sorting method;

[0033] The tournament selection method was used to screen chromosomes for fitness, resulting in high-quality chromosomes.

[0034] The selected high-quality chromosomes are subjected to evolutionary operations using simulated binary crossover and polynomial mutation operators to generate a new generation of chromosome populations.

[0035] Using a new generation of chromosome population as input, chromosome evaluation and evolution operations are repeatedly performed until the maximum number of iterations is reached to obtain a commodity scheduling plan.

[0036] As a preferred embodiment of the IoT-based commodity wholesale data management method of the present invention, the step of converting the commodity scheduling plan into specific scheduling instructions and then issuing them to the corresponding warehouse management center for execution specifically includes:

[0037] The commodity dispatch plan converts the various commodity transfer quantities from the source warehouse to the target warehouse into specific warehouse transfer and outbound instructions. These instructions are then sent to the warehouse management center of the source warehouse via an application programming interface (API) to execute the corresponding picking, packaging, and outbound processes.

[0038] The truck route arrangement in the commodity scheduling plan is transformed into specific truck transportation task instructions. The truck transportation task instructions are sent to the transportation management platform through a message queue, and specific trucks are assigned to perform the transportation tasks.

[0039] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the Internet of Things-based commodity wholesale data management method described in the first aspect of the present invention.

[0040] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the Internet of Things-based commodity wholesale data management method described in the first aspect of the present invention.

[0041] The beneficial effects of this invention are as follows: By quantifying the cumulative damage to commodity value caused by environmental factors such as temperature and humidity through an integral quality decay function, the measurability of the quality status of inventory goods and the ability to preserve value are improved. By extracting operational features from IoT sensing signals of downstream distribution nodes, the GBDT model is driven to generate differentiated replenishment instructions, enhancing the real-time performance and accuracy of demand forecasting. The quality decay coefficient is incorporated into inventory value assessment and, together with logistics costs, a weighted single-objective optimization model is constructed. Combined with metaheuristic algorithms to solve the optimal scheduling scheme, the dual objectives of reducing transportation costs and prioritizing the scheduling of high-risk goods are achieved. This solves the problems of hidden losses, inventory backlogs, and stockouts caused by the lack of quality monitoring and delayed demand response in previous commodity wholesale management, thereby improving the overall operational efficiency and asset preservation level of the supply chain. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a commodity wholesale data management method based on the Internet of Things.

[0044] Figure 2 A flowchart for generating differentiated replenishment demand instructions.

[0045] Figure 3 A flowchart for generating a single-objective mathematical programming model.

[0046] Figure 4 This is a flowchart for calculating the product quality degradation coefficient. Detailed Implementation

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0048] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0049] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0050] Reference Figures 1-4 This is one embodiment of the present invention, which provides a commodity wholesale data management method based on the Internet of Things, including the following steps:

[0051] S1. Collect environmental parameters and static information of the microenvironment for product storage;

[0052] S1.1 It should be noted that temperature and humidity sensors and gas sensor arrays are deployed in the warehouse shelves and storage areas respectively to collect environmental parameters (including temperature and humidity) of the microenvironment for commodity storage; static information of the corresponding goods in the warehouse management database is obtained by querying the inventory master data table; specifically, the warehouse location code of the goods to be queried is extracted as a structured query statement; after the structured query statement is sent to the warehouse management database server, the query operation is executed on the inventory master data table, that is, an index field is created in the inventory master data table according to the goods location code for fast retrieval; after successful retrieval, the goods category identifier, entry time, warehouse location, goods quantity, purchase cost, goods identifier and storage threshold, i.e., static information, are read from the corresponding record in the inventory master data table.

[0053] It should also be noted that the establishment of the warehouse management database involves collecting information on product category, purchase batch, and production date through scanning devices during the product receiving process; establishing a binding relationship between products and storage locations by scanning the location codes and product identifiers through handheld terminals during the product shelving process; and using database management tools to store the product information collected by the scanning devices and the binding relationship between products and storage locations into the inventory master data table, location information table, and product information table created in the database management center, thus forming the warehouse management database.

[0054] S2. Using the quality decay function, compare and integrate the environmental parameters with the stored threshold in the static information to obtain the product quality decay coefficient.

[0055] S2.1 Compare the environmental parameters with the storage threshold to obtain the deviation between the environmental parameters and the storage threshold;

[0056] It should be noted that the storage threshold for product categories in static information is read, and the storage threshold includes an upper limit threshold and a lower limit threshold; the environmental parameters are arithmetically compared with the corresponding storage thresholds; when the environmental parameters exceed the upper limit threshold, the deviation calculation formula is:

[0057] ;

[0058] in, This indicates the degree of deviation when environmental parameters exceed the upper limit threshold. Indicates environmental parameters, Indicates the upper limit threshold;

[0059] When environmental parameters are below the lower threshold, the deviation calculation formula is:

[0060] ;

[0061] in, This indicates the degree of deviation when environmental parameters are below the lower threshold. Indicates the lower threshold. Indicates environmental parameters;

[0062] It should also be noted that the specific range of the upper limit threshold is determined by the chemical characteristics of the product, while the lower limit threshold is determined by the physical characteristics of the product. For example, for cold chain products, the upper limit threshold is determined by the protein denaturation temperature, which is usually no higher than -18°C; for electronic components, the lower limit threshold is determined by the material embrittlement temperature, such as silicon chips, which are no lower than -40°C.

[0063] S2.2. Perform an integral operation on the deviation between environmental parameters and storage threshold over time, and substitute the integral operation result as an exponential term into the quality decay function to obtain the product quality decay coefficient.

[0064] It should be noted that, within the specified time interval, the deviation between environmental parameters and the storage threshold is continuously sampled according to the method for calculating the deviation in step S2.1, forming a deviation time series. A definite integral operation is then performed on the deviation time series, i.e., the area under the deviation-time curve from the initial time to the current time is calculated. This area value is the cumulative deviation. The definite integral operation is expressed by the formula:

[0065] ;

[0066] in, Indicates the cumulative deviation. Indicates the current moment. Indicates the start time. Indicates in The deviation between environmental parameters at a given time point and the storage threshold. Indicates to Integrating at specific points in time. Represents a point in time, and is used in integration operations as a loop variable to iterate through and accumulate the effects of each instant;

[0067] The cumulative deviation is multiplied by the reaction rate constant corresponding to the product category to obtain the exponential term input value of the quality decay function. This exponential term input value is then substituted into the exponential function with the natural constant e as its base, and the product quality decay coefficient is calculated through exponential operations, expressed by the formula:

[0068] ;

[0069] in, Indicates the product quality degradation coefficient. Represents the natural constant. This represents the reaction rate constant corresponding to the product category. Indicates the current moment. Indicates the start time. Indicates in The deviation between the environmental parameters at any given time and the storage threshold. Indicates to Integrate points continuously;

[0070] The product quality degradation coefficient ranges from 0 to 1, where a value of 1 represents a product in good condition and a value of 0 represents a product that is completely ineffective.

[0071] S3. After binding the product quality decay coefficient with static information, a dynamic inventory list is generated;

[0072] S3.1 It should be noted that, based on the product identifiers collected for the products in step S1.1, an independent storage record is assigned to each product identifier; fixed fields are created in the storage record (including product identifier field, warehouse entry timestamp field, storage location code field, product quantity field, and product quality decay coefficient field); the product identifier field is filled with the product identifier characters; the warehouse entry timestamp field is filled with the time point when the product was completed in the warehouse; the storage location code field is filled with the physical storage location code where the product was actually stored; the product quantity field is filled with the actual quantity of the product in the warehouse; the product quality decay coefficient field is initialized to the value 1, indicating that the product is in good condition; after all fields are filled, a dynamic inventory record is formed.

[0073] All dynamic inventory records are grouped and sorted according to product category identifiers; within each product category identifier group, after being sorted in ascending order by the inbound timestamp field, a composite index based on product category identifiers and product quality decay coefficients is created in the dynamic inventory records; the indexed dynamic inventory records are stored as a dynamic inventory list; the dynamic inventory list supports fast retrieval and querying by product category identifiers and product quality decay coefficient thresholds.

[0074] S4. Collect IoT sensing signal data packets from downstream distribution nodes of goods, and perform correlation analysis with dynamic inventory list to generate differentiated replenishment demand instructions.

[0075] S4.1 Collect IoT sensing signal data packets from IoT devices at downstream distribution nodes of goods;

[0076] It should be noted that surveillance cameras are deployed at the entrance of the downstream distribution node warehouse to continuously collect video stream data (including time-series image frames of truck entry and exit activities); weight sensors are installed on the shelves of the downstream distribution nodes to periodically collect pressure data, which reflects the changes in the weight of the goods on the shelves; all collected video stream data and pressure data are encapsulated through an IoT gateway to form a standard format IoT sensing signal data packet.

[0077] S4.2 Extract signal features from IoT sensing signal data packets to obtain feature values ​​of truck entry and exit frequency and shelf inventory depth;

[0078] It should be noted that, based on the time-series image frames of truck entry and exit activities in the video stream data, the number of times trucks appear within a specified time window (e.g., 3 days) is counted; the quotient of the truck appearance count and the time window (number of days) is then calculated to obtain the truck entry and exit frequency. The pressure data is input into the prediction update stage of the Kalman filter. The Kalman filter performs recursive calculations based on the statistical characteristics of the pressure data and outputs the stabilized pressure value, expressed by the formula:

[0079] ;

[0080] in, This represents the pressure value after it has stabilized. This represents the predicted pressure value. Indicates Kalman gain, This represents the actual pressure measurement value in the pressure data;

[0081] The product of the stabilized pressure value and the linear transformation coefficient is used to obtain the product quantity sequence; the product quantity with the latest timestamp is selected from the product quantity sequence as the shelf inventory depth feature value.

[0082] It should also be noted that the linear conversion coefficient is obtained through calibration calculation. Specifically, after the warehouse shelves are installed, a known quantity of standard-weight goods is placed on the shelves in a gradient combination, and a pressure sensor is used to read the voltage readings. The least squares method is used to calculate the linear conversion coefficient by relating the voltage readings to the corresponding quantity of goods, which is expressed by the formula:

[0083] ;

[0084] in, Represents the linear transformation coefficients. Indicates the quantity of goods. Indicates the voltage reading;

[0085] S4.3 Input the product quality decay coefficient, truck entry and exit frequency and shelf inventory depth feature value from the dynamic inventory list into the GBDT model to predict demand and obtain differentiated replenishment demand instructions;

[0086] It should be noted that the historical product quality decay coefficient (collected according to the product quality decay coefficient calculation method in step S2.2), historical truck entry and exit frequency, historical shelf inventory depth feature value, and historical actual replenishment quantity are collected at each time point within a certain period. The historical product quality decay coefficient, historical truck entry and exit frequency, and historical shelf inventory depth feature value are used to form a training feature matrix, and the historical actual replenishment quantity is used as the training label. The training feature matrix and training label are input into the GBDT model. The GBDT model iteratively constructs multiple regression decision trees through the gradient boosting algorithm. Each regression decision tree learns the mapping relationship between the training feature matrix and the training label. During the training process, the GBDT model gradually optimizes the parameters of each regression decision tree in the GBDT model by calculating the gradient of the loss function between the predicted replenishment quantity and the historical actual replenishment quantity. When the GBDT model training reaches the maximum number of iterations, the trained GBDT model is obtained. The maximum number of iterations for the GBDT model is determined by the number of product categories. For example, if the number of product categories is less than 1000, the maximum number of iterations is set to 500; if the number of product categories is between 1000 and 5000, the maximum number of iterations is set to 1000; and if the number of product categories is greater than 5000, the maximum number of iterations is set to 2000.

[0087] The product quality decay coefficient, truck entry and exit frequency, and shelf inventory depth features from the dynamic inventory list are input into the trained GBDT model. The first decision tree of the GBDT model generates an initial prediction value according to the feature splitting rule. The initial prediction value is then input into subsequent decision trees, with each subsequent decision tree correcting the prediction residual of the previous decision tree. The prediction results of all decision trees in the GBDT model are weighted and summed to obtain the preliminary predicted demand for each product category. The preliminary predicted demand is then transformed using a sigmoid function to obtain the predicted demand for each product category. The predicted demand for each product category is represented in floating-point form, representing the expected quantity of products that need to be replenished within a specific future period.

[0088] By forecasting demand for each product category, a safety stock threshold is preset. The forecasted demand for each product category is compared with the safety stock threshold. When the forecasted demand is lower than the safety stock threshold, a replenishment suggestion is generated, and the replenishment quantity is the difference between the safety stock threshold and the forecasted demand. The replenishment suggestions are prioritized according to the product quality decay coefficient. For example, product categories with a quality decay coefficient below 0.7 are marked as the highest priority, product categories with a quality decay coefficient between 0.7 and 0.9 are marked as medium priority, and product categories with a quality decay coefficient above 0.9 are marked as ordinary priority. When the forecasted demand for a product category is greater than the safety stock threshold, no replenishment is required, ultimately forming a differentiated replenishment demand instruction.

[0089] It should also be noted that the safety stock threshold is measured in units of product quantity. The specific value is determined based on a quantitative analysis of the company's specific stockout costs, inventory holding costs, and demand uncertainty, rather than being an arbitrary or fixed number. For example, the safety stock threshold for fast-moving consumer goods (FMCG) is 7-30 days of average daily sales. The lower limit (7 days) is the survival line; below this level, inventory is highly susceptible to stockouts due to demand fluctuations or transportation delays, leading to sales losses and customer dissatisfaction. The upper limit (30 days) is the cost tolerance line; above this level, holding costs (capital and warehousing) will erode profits, and the risk of product expiration increases dramatically.

[0090] Product categories with a quality degradation coefficient below 0.7 are marked as the highest priority because the current value of these products is lower than their original value due to cumulative environmental damage. If they are not dealt with first, they are very likely to suffer a total loss in the short term due to spoilage, expiration, or other reasons.

[0091] Goods categories with a quality decay coefficient between 0.7 and 0.9 are marked as medium priority because the quality of the goods has begun to decay at an accelerated rate and the asset value is continuing to decline. Replenishing with medium priority aims to seize the opportunity to convert the goods into cash flow through sales or allocation before the value of the goods falls further significantly, thereby controlling the extent of value loss.

[0092] Goods with a quality degradation coefficient higher than 0.9 are marked as ordinary priority because the goods are almost intact and their asset value is close to their original value. Therefore, the urgency of replenishment is the lowest, and limited warehousing, capital and transportation resources can be prioritized for high-risk and medium-risk goods.

[0093] S5. Using differentiated replenishment demand instructions as the scheduling target, a single-objective mathematical programming model is constructed in conjunction with a dynamic inventory list;

[0094] S5.1 It should be noted that, based on differentiated replenishment demand instructions, decision variables (including the transfer volume of various commodities from the source warehouse to the target warehouse and the truck routes for executing the transfer tasks) are defined. The total logistics transportation cost is calculated based on the decision variables, expressed by the formula:

[0095]

[0096] in, This represents the total logistics and transportation cost. Indicates the total number of available trucks. This indicates the total number of warehouses (including source and target warehouses). Indicates the total number of product categories. Indicates from the source repository To the target warehouse The transportation distance Indicates truck Unit distance transportation cost Indicates goods unit weight, Indicates from the source warehouse To the target warehouse The unit weight transportation rate, Represents path decision variables. Indicates from the source warehouse Transferred to the target warehouse The number of product categories;

[0097] It should also be noted that path decision variables The value can only be 0 or 1. Specifically, when the truck... The driving route is from the source warehouse To the target warehouse hour, The value is 1; when the truck The travel route is not from the source warehouse. To the target warehouse hour, The value is 0;

[0098] S5.3. The weighted summation of the product quality decay coefficients in the dynamic inventory list yields the total inventory value preservation rate.

[0099] It should be noted that the process involves iterating through each dynamic inventory record in the dynamic inventory list, reading the product quality decay coefficient, product quantity, and purchase cost per unit from the dynamic inventory record; multiplying the product quality decay coefficient by the product quantity to obtain the product quality adjustment quantity; multiplying the product quality adjustment quantity by the purchase cost per unit to obtain the product preservation value; and summing the preservation values ​​of all products in the dynamic inventory list to obtain the total inventory value preservation level.

[0100] S5.4. Using differentiated replenishment demand instructions as constraints, and with the objective functions of minimizing total logistics costs and maximizing the preservation of total inventory value, establish a single-objective mathematical programming model.

[0101] It should be noted that the product categories, target warehouses, and minimum replenishment quantities specified in the differentiated replenishment demand instructions are used as mathematical constraints to ensure that the total amount of goods transferred from the source warehouse to the target warehouse meets the requirements of the differentiated replenishment demand instructions. The total logistics transportation cost calculation formula is used as the first objective function to minimize the total logistics transportation cost; the total inventory value preservation rate calculation formula is used as the second objective function to maximize the total inventory value preservation rate. The total logistics transportation cost objective function and the total inventory value preservation rate objective function are normalized to eliminate dimensional differences. Weighting coefficients are then assigned to the total logistics transportation cost objective function, for example, 0.8, and to the total inventory value preservation rate objective function, for example, 0.2. The objective function of total logistics transportation cost is multiplied by the corresponding weight coefficient (0.8), and the objective function of total inventory value preservation is multiplied by the corresponding weight coefficient (0.2) to obtain the weighted objective function values ​​of total logistics transportation cost and total inventory value preservation. The weighted objective function values ​​of total logistics transportation cost and total inventory value preservation are added together to obtain the comprehensive objective function value. The optimization objective is to minimize the comprehensive objective function value, which together with the mathematical constraints constitutes a single-objective mathematical programming model.

[0102] It should also be noted that the weighting coefficients of the total logistics transportation cost objective function and the total inventory value preservation objective function are set according to financial management principles and commodity characteristics (low value and high shelf life), indicating that the majority of efforts should be devoted to reducing cash costs.

[0103] S6. Use a metaheuristic optimization algorithm to solve the single-objective mathematical programming model and obtain a commodity scheduling scheme;

[0104] S6.1 It should be noted that an initial population consisting of randomly generated chromosomes is initialized. The chromosomes are constructed as follows: Unique gene positions are assigned to the commodity allocation quantity in the decision variables using integer encoding. The gene positions are determined by prioritizing the source warehouse number, target warehouse number, and commodity category number in ascending order. The integer value of the commodity allocation quantity (e.g., 150 boxes) is directly used as the equivalent integer and filled into the assigned gene positions to obtain integer gene segments of the chromosome. For the truck route arrangement for performing the allocation task in the decision variables, real-number encoding is used to encode the real-number sequence gene segments in the chromosome according to the path node order. The integer gene segments and the real-number sequence gene segments are then concatenated to form the chromosome.

[0105] The algorithm sequentially reads integer gene segments from each chromosome in the initial population, mapping the integer values ​​in these segments to the allocation quantities of various goods from the source warehouse to the target warehouse using decoding rules. It then analyzes real-number sequence gene segments in the chromosomes, converting their encoded values ​​into truck routes. The decoded allocation quantities of various goods and truck routes are then substituted into the total logistics transportation cost calculation formula to calculate the total logistics transportation cost of the chromosomes. Finally, the decoded allocation quantities of various goods are combined with the product quality decay coefficient, product quantity, and purchase cost unit price in the dynamic inventory list to calculate the chromosome inventory value preservation rate. Based on the calculated total logistics transportation cost and chromosome inventory value preservation rate, the chromosome fitness is calculated using a single-objective mathematical programming model, expressed by the following formula:

[0106] ;

[0107] in, Indicates the fitness of chromosomes. The weighting coefficients representing the total logistics and transportation costs of chromosomes. This represents the total logistics and transportation cost of chromosomes. Indicates the preservation of the chromosome inventory value;

[0108] It should also be noted that the weighting coefficient for the total logistics and transportation cost of chromosomes is 0.8, which is also set according to financial management principles and product characteristics (meaning that most of the effort is devoted to reducing cash costs).

[0109] S6.3. Use tournament selection to screen chromosome fitness and obtain high-quality chromosomes;

[0110] It should be noted that the tournament selection method sets a competition size parameter (e.g., 5), which determines the number of chromosomes participating in each selection; a tournament group is formed by randomly selecting a number of chromosomes specified by the competition size parameter from the initial population; the fitness of all chromosomes in the tournament group is compared, and the chromosome with the best fitness is selected as the winner; the winning chromosome is copied to the candidate pool of the next generation population; the operations of randomly selecting tournament groups and selecting winners are repeated until the number of chromosomes in the candidate pool of the next generation population reaches the preset population size (e.g., 80), thus obtaining high-quality chromosomes.

[0111] It should also be noted that the competition size parameter is set to 5 based on the selection pressure control principle in genetic algorithm theory, which can maintain the diversity of the population and avoid premature convergence.

[0112] The population size is preset to 80 because it can effectively cover multiple locally optimal regions in the logistics scheduling problem.

[0113] S6.4. Apply simulated binary crossover operator and polynomial mutation operator to perform evolutionary operations on selected high-quality chromosomes to generate a new generation of chromosome population;

[0114] It should be noted that the selected high-quality chromosomes are paired to form parent chromosome pairs. A random number uniformly distributed in the interval [0,1] is generated for each parent chromosome pair. The generated random number is compared with the preset crossover probability. If the random number is less than or equal to the crossover probability, a simulated binary crossover operation is performed on the parent chromosome pair, that is, the simulated binary crossover operator calculates the weighted average of the gene values ​​of the parent chromosome pair (i.e., the gene values ​​of the offspring chromosomes). If the random number is greater than the crossover probability, the parent chromosome pair is not crossovered and is directly copied into the offspring chromosome population. The crossover probability is usually between 0.7 and 0.9, which is set according to the genetic algorithm theory and is the empirically optimal range for maintaining healthy population evolution, avoiding premature convergence, and ensuring computational efficiency. A polynomial mutation operator is applied to the offspring chromosomes generated after crossover. The polynomial mutation operator randomly perturbs the gene values ​​of the offspring chromosomes based on the probability distribution. The offspring chromosomes generated after the evolutionary operations of the simulated binary crossover operator and the polynomial mutation operator, together with some high-quality chromosomes that did not participate in the evolution, constitute the new generation chromosome population.

[0115] Using the new generation of chromosome population as input, the chromosome evaluation and evolution operations are repeated until the iteration count counter reaches the preset maximum number of iterations (e.g., 200) to obtain the last generation of chromosome population. The chromosome with the best fitness in the last generation of chromosome population is decoded into the specific allocation of various commodities from the source warehouse to the target warehouse and the arrangement of truck routes, forming the final commodity scheduling plan.

[0116] It should also be noted that the maximum number of iterations of 200 is set according to convergence theory, ensuring that the algorithm has sufficient time to find a high-quality solution.

[0117] S7. After converting the commodity scheduling plan into specific scheduling instructions, send them to the corresponding warehouse management center for execution;

[0118] S7.1 It should be noted that the various commodity transfer quantities from the source warehouse to the target warehouse in the commodity scheduling plan are converted into specific warehouse transfer and outbound instructions. The warehouse transfer and outbound instructions are sent to the warehouse management center of the source warehouse through the application programming interface to execute the corresponding picking, packaging and outbound processes. The truck route arrangement in the commodity scheduling plan is converted into specific truck transportation task instructions. The truck transportation task instructions are sent to the transportation management platform through the message queue to allocate specific trucks to perform transportation tasks.

[0119] This embodiment also provides a computer device applicable to the Internet of Things (IoT)-based commodity wholesale data management method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the IoT-based commodity wholesale data management method proposed in the above embodiment.

[0120] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0121] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the IoT-based commodity wholesale data management method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0122] In summary, this invention quantifies the cumulative damage to commodity value caused by environmental factors such as temperature and humidity through an integral quality decay function, thereby improving the measurability of the quality status of inventory goods and their value preservation capabilities. It also utilizes IoT sensing signals from downstream distribution nodes to extract operational characteristics, driving the GBDT model to generate differentiated replenishment instructions, enhancing the real-time performance and accuracy of demand forecasting. Furthermore, it incorporates the quality decay coefficient into inventory value assessment and constructs a weighted single-objective optimization model together with logistics costs. Combined with a metaheuristic algorithm to solve the optimal scheduling scheme, this invention achieves the dual objectives of reducing transportation costs while prioritizing the scheduling of high-risk goods. This solves the problems of hidden losses, inventory backlogs, and stockouts caused by the lack of quality monitoring and delayed demand response in traditional commodity wholesale management, improving the overall operational efficiency and asset preservation level of the supply chain.

[0123] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for managing wholesale commodity data based on the Internet of Things, characterized in that: include, Collect environmental parameters and static information of the microenvironment in which goods are stored; The quality decay function is used to compare and integrate environmental parameters with stored thresholds in static information to obtain the product quality decay coefficient. By binding the product quality decay coefficient with static information, a dynamic inventory list is generated. Collect IoT sensing signal data packets from downstream distribution nodes of goods, and perform correlation analysis with dynamic inventory lists to generate differentiated replenishment demand instructions; Using differentiated replenishment demand instructions as the scheduling objective, a single-objective mathematical programming model is constructed by combining a dynamic inventory list; A metaheuristic optimization algorithm is used to solve a single-objective mathematical programming model to obtain a commodity scheduling scheme. After the commodity scheduling plan is converted into specific scheduling instructions, it is sent to the corresponding warehouse management center for execution.

2. The IoT-based commodity wholesale data management method as described in claim 1, characterized in that: The environmental parameters of the microenvironment for product storage include temperature and humidity; The static information includes product category identifier, warehousing time, warehouse location, product quantity, procurement cost, product identifier, and storage threshold.

3. The IoT-based commodity wholesale data management method as described in claim 2, characterized in that: The process of using a quality decay function to compare and integrate environmental parameters with stored thresholds in static information to obtain the product quality decay coefficient is as follows: The environmental parameters are compared with the storage threshold to obtain the deviation between the environmental parameters and the storage threshold; The deviation between environmental parameters and storage thresholds is integrated over time. The result of the integration is then substituted into the quality decay function as an exponential term to obtain the product quality decay coefficient.

4. The IoT-based commodity wholesale data management method as described in claim 3, characterized in that: The process of binding the product quality decay coefficient with static information to generate a dynamic inventory list refers to using product identifiers to create independent dynamic inventory records for all products, aggregating the dynamic inventory records of all products, and forming a dynamic inventory list.

5. The IoT-based commodity wholesale data management method as described in claim 4, characterized in that: The process involves collecting IoT sensing signal data packets from downstream distribution nodes of goods, performing correlation analysis with the dynamic inventory list, and generating differentiated replenishment demand instructions. Specifically: Collect IoT sensing signal data packets from IoT devices at downstream distribution nodes of goods; Extract signal features from IoT sensing signal data packets to obtain feature values ​​of truck entry and exit frequency and shelf inventory depth; Input the product quality decay coefficient, truck entry and exit frequency, and shelf inventory depth feature values ​​from the dynamic inventory list into the GBDT model to forecast demand and obtain differentiated replenishment demand instructions.

6. The IoT-based commodity wholesale data management method as described in claim 5, characterized in that: The method of constructing a single-objective mathematical programming model based on differentiated replenishment demand instructions as the scheduling target and combined with a dynamic inventory list is as follows: Decision variables are defined based on differentiated replenishment demand instructions, and total logistics and transportation costs are calculated based on these decision variables. The decision variables include the volume of various commodities transferred from the source warehouse to the target warehouse and the truck routes that carry out the transfer tasks; The total inventory value preservation rate is obtained by weighted summing of the product quality decay coefficients in the dynamic inventory list. Using differentiated replenishment demand instructions as constraints, and with the objectives of minimizing total logistics costs and maximizing the preservation of total inventory value, a single-objective mathematical programming model is established.

7. The IoT-based commodity wholesale data management method as described in claim 6, characterized in that: The method of using a metaheuristic optimization algorithm to solve the single-objective mathematical programming model to obtain a commodity scheduling scheme is as follows: The decision variables of the single-objective mathematical programming model are encoded as chromosomes in the metaheuristic optimization algorithm, and the total logistics transportation cost and inventory value preservation of each chromosome are calculated. For each chromosome, the total logistics transportation cost and inventory value preservation rate are used to calculate chromosome fitness using a non-dominated sorting method; The tournament selection method was used to screen chromosomes for fitness, resulting in high-quality chromosomes. The selected high-quality chromosomes are subjected to evolutionary operations using simulated binary crossover and polynomial mutation operators to generate a new generation of chromosome populations. Using a new generation of chromosome population as input, chromosome evaluation and evolution operations are repeatedly performed until the maximum number of iterations is reached to obtain a commodity scheduling plan.

8. The IoT-based commodity wholesale data management method as described in claim 7, characterized in that: The process of converting the commodity scheduling plan into specific scheduling instructions and then issuing them to the corresponding warehouse management center for execution is as follows: The commodity dispatch plan converts the various commodity transfer quantities from the source warehouse to the target warehouse into specific warehouse transfer and outbound instructions. These instructions are then sent to the warehouse management center of the source warehouse via an application programming interface (API) to execute the corresponding picking, packaging, and outbound processes. The truck route arrangement in the commodity scheduling plan is transformed into specific truck transportation task instructions. The truck transportation task instructions are sent to the transportation management platform through a message queue, and specific trucks are assigned to perform the transportation tasks.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the IoT-based commodity wholesale data management method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the IoT-based commodity wholesale data management method according to any one of claims 1 to 8.

Images