Intelligent warehousing system optimization method based on production equipment linkage

Abstract

The invention relates to the technical field of intelligent warehousing management, in particular to an intelligent warehousing system optimization method based on production equipment linkage. The method comprises the following steps: carrying out multi-source data acquisition on a sales system and a production system, and carrying out dynamic demand prediction to obtain a predicted demand quantity; obtaining a product attribute list; performing safe inventory calculation according to the predicted demand quantity and the product attribute list to obtain a safe inventory benchmark; according to the predicted demand quantity, the product attribute list and the safety inventory benchmark, carrying out maximum inventory quantity rough calculation to obtain a maximum inventory limit; performing dynamic threshold adjustment according to the highest inventory limit to obtain a dynamic inventory threshold; and carrying out production demand decomposition on the predicted demand quantity and the dynamic inventory threshold value to obtain a to-be-produced list. According to the method, collaborative optimization of warehouse management and production plans is realized through data driving, prediction driving and intelligent optimization, and the warehouse operation efficiency and enterprise benefits are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent warehousing management, and in particular to an intelligent warehousing system optimization method based on production equipment linkage. Background Art

[0002] In the context of intelligent manufacturing, enterprises are paying more and more attention to the deep integration of production and logistics to improve efficiency, reduce costs and enhance market responsiveness. Early automated warehouses focused on improving the efficiency of warehousing operations, such as using automated stereoscopic warehouses (AS / RS), conveyor lines, stackers and other equipment to replace manual handling and storage. Later, enterprises began to use software such as warehouse management systems (WMS) and manufacturing execution systems (MES) to manage warehousing and production activities. However, the existing warehouse management methods still have the disadvantages of information islands and passive responsive management, which are specifically reflected as follows: Information islands: Traditional warehouse management systems and production execution systems are often independent of each other, making it difficult to share and exchange data, forming information islands. For example, the warehouse department is not clear about the real-time output and production plan of the production line, and the production department is not aware of the real-time inventory situation of the warehouse. This information asymmetry can lead to a disconnect between production plans and inventory, delayed replenishment decisions, and irrational resource allocation.

[0003] Passive responsive management: Traditional warehouse management systems are mostly passive responsive, that is, replenishment and delivery are carried out according to current inventory levels and order requirements. This passive management model is difficult to cope with fluctuations in market demand and changes in production plans, resulting in poor warehouse management ability to cope with changes, insufficient forecasting capabilities, and limited optimization space. Summary of the invention

[0004] Based on this, it is necessary to provide an intelligent warehousing system optimization method based on production equipment linkage to solve at least one of the above technical problems.

[0005] To achieve the above purpose, an intelligent warehousing system optimization method based on production equipment linkage includes the following steps: Preferably, the method comprises the following steps: Step S1: collect multi-source data from the sales system and the production system, and perform dynamic demand forecasting to obtain forecasted demand; Step S2: Obtain a product attribute list; calculate the safety stock quantity according to the predicted demand and the product attribute list to obtain a safety stock benchmark; estimate the maximum inventory quantity according to the predicted demand, the product attribute list and the safety stock benchmark to obtain a maximum inventory limit; adjust the dynamic threshold according to the maximum inventory limit to obtain a dynamic inventory threshold; Step S3: Decompose the production demand based on the predicted demand and the dynamic inventory threshold to obtain a to-be-produced list; optimize the production batches based on the to-be-produced list to obtain an optimized batch plan; arrange the production sequence for the optimized batch plan, and formulate a detailed production plan to obtain a production plan; Step S4: Obtain the latest inventory snapshot of the current inventory status; identify replenishment needs based on the latest inventory snapshot and dynamic inventory thresholds, and perform production plan verification to obtain a to-be-replenished list and production matching results; automatically generate replenishment instructions based on the production plan, to-be-replenished list, and production matching results to obtain inventory replenishment instructions; perform intelligent storage allocation based on the inventory replenishment instructions and the product attribute list, and issue warehousing tasks to obtain warehousing operation instructions; Step S5: Acquire actual sales data; execute warehouse system instructions according to warehouse operation instructions to obtain warehouse operation data; calculate performance indicators according to the latest inventory snapshot, production plan, actual sales data and warehouse operation data to obtain quantitative performance indicators; use quantitative performance indicators to perform performance deviation analysis on predicted demand, dynamic inventory thresholds and production plans to obtain a deviation analysis report; optimize performance according to the deviation analysis report, and monitor the optimization effect to achieve the optimization task of the intelligent warehouse system.

[0006] The present invention can more accurately predict future demand by integrating multi-source data and using multiple forecasting models, providing a reliable basis for subsequent inventory management and production planning, and avoiding out-of-stock or overstock caused by inaccurate forecasting. Taking into account factors such as forecasted demand, product attributes (such as turnover rate, storage space, shelf life), warehouse capacity and market activities, the inventory threshold is dynamically set to achieve refined inventory control, which can effectively reduce inventory costs and ensure the stability of the supply chain. According to the forecasted demand and inventory threshold, an accurate production plan is formulated, and by optimizing production batches and production sequences, equipment utilization is maximized, production costs are minimized, production efficiency is improved, and market demand can be better met. Automated inventory replenishment is achieved, and coordinated with production plans, inventory can be replenished in time to avoid out-of-stock, while avoiding inventory backlogs caused by excessive replenishment, and storage space utilization is optimized through intelligent cargo space allocation. By continuously monitoring and evaluating system performance and using reinforcement learning models for parameter optimization, the forecasting model, inventory strategy and production plan can be continuously improved to achieve system adaptation and continuous optimization, so as to better adapt to the changing market environment. Therefore, the present invention provides an intelligent warehousing system optimization method based on the linkage of production equipment. Through data-driven, prediction-driven and intelligent optimization, it effectively solves the drawbacks of information islands and passive response management, realizes the coordinated optimization of warehousing management and production plans, and significantly improves warehousing operation efficiency and enterprise benefits.

[0007] Preferably, step S1 comprises the following steps: Step S11: Collect and fuse multi-source data from the sales system and the production system to obtain an integrated data set; Step S12: extracting data features from the integrated data set to obtain a feature data set; Step S13: performing time series pattern analysis on the feature data set to obtain a time series prediction result; Step S14: performing product sales prediction based on the LSTM neural network model according to the feature data set to obtain product sales prediction results; Step S15: Fusing the time series prediction results and the product sales prediction results to obtain a fused prediction result; Step S16: Acquire real-time production data; calibrate the fusion prediction results according to the real-time production data to obtain the predicted demand.

[0008] The present invention eliminates information islands by integrating sales, production and market data, provides a complete data foundation for subsequent accurate demand forecasting, and helps to improve the accuracy and reliability of forecasting. Extracting and constructing richer features, such as time features, sales features, promotional features and production features, can more comprehensively reflect various factors affecting demand, thereby improving the learning ability and forecasting accuracy of the forecasting model. The ARIMA model is used to capture time series patterns in historical sales data, such as trend, seasonality and periodicity, and can effectively predict demand fluctuations in the short term, providing a basis for the subsequent fusion of forecast results. The LSTM neural network model is used to learn complex demand patterns and nonlinear relationships, and can effectively process long sequence data, thereby improving the accuracy of medium- and long-term demand forecasting and making up for the shortcomings of traditional time series models. The short-term forecast results of the ARIMA model and the long-term forecast results of the LSTM model are integrated, combining the advantages of the two models, and can more accurately predict demand for a period of time in the future, improving the robustness of the forecast. By considering real-time production data, such as equipment capacity and production plan, the forecast results are calibrated to ensure that the generated forecast demand is within the actual production capacity, avoiding the generation of unexecutable production plans, and improving the feasibility of the plan.

[0009] Preferably, step S2 comprises the following steps: Step S21: Analyze the demand characteristics of the predicted demand to obtain demand characteristic parameters; Step S22: using the sales system and the production system to evaluate product attributes and obtain a product attribute list, wherein the product attribute list includes a target turnover rate, a single product storage space, and a product shelf life; Step S23: Calculate the safety stock quantity according to the demand characteristic parameters and the product attribute list to obtain the safety stock benchmark; Step S24: setting the minimum inventory quantity according to the safety inventory benchmark and the demand characteristic parameters to obtain the minimum inventory threshold; Step S25: Obtain warehouse capacity information and market activity forecast data; Step S26: Calculate the maximum inventory quantity based on the predicted demand, product attribute list and warehouse capacity information to obtain the maximum inventory limit; Step S27: Adjust the dynamic threshold according to the safety stock benchmark, the minimum stock threshold, the maximum stock limit and the market activity forecast data to obtain the dynamic stock threshold.

[0010] The present invention can more accurately grasp the volatility of demand by analyzing the statistical characteristics of the predicted demand, such as the mean, standard deviation and coefficient of variation, and provide data support for the subsequent setting of more refined inventory thresholds, avoiding one-size-fits-all extensive management. Collecting product attribute information provides an important reference basis for the subsequent inventory threshold setting, so that the inventory strategy can better adapt to the characteristics of different products, such as high turnover products and perishable products. The inventory management strategy will be different. Considering demand fluctuations and service levels, and combining the shelf life of the product, the safety stock is calculated, which can effectively prevent out-of-stock caused by demand fluctuations or supply chain disruptions, and avoid inventory backlogs and expiration losses caused by excessive safety stock settings, especially for products with short shelf life. By setting the minimum inventory threshold, the replenishment process can be triggered in time to ensure that the inventory level is always kept within a safe range, avoid the risk of out-of-stock, and provide a reference for the formulation of production plans. Obtaining warehouse capacity information and market activity forecast data provides real-time information for dynamically adjusting the inventory threshold, so that the inventory strategy can be adjusted according to the actual situation. For example, during promotional activities, the inventory level can be appropriately increased to cope with the expected demand growth. Taking into account the forecast demand, product attributes and warehouse capacity limitations, the maximum inventory limit is calculated, which can effectively control the total inventory, avoid inventory backlogs and excessive storage resources, and ensure sufficient inventory to meet market demand. Dynamically adjusting the inventory threshold according to market activity forecast data can flexibly respond to market changes. For example, on the eve of a promotion, the maximum inventory limit can be increased in advance to prepare for the promotion; in periods of low market demand, the inventory threshold can be lowered to reduce inventory holding costs. This makes inventory management more flexible and efficient.

[0011] Preferably, step S26 comprises the following steps: Step S261: Calculate the ideal maximum inventory based on the predicted demand and the target turnover rate to obtain an ideal maximum inventory list; Step S262: Evaluate the ideal inventory space demand for the ideal maximum inventory list and the single product storage space to obtain the ideal total inventory space demand; Step S263: applying warehouse capacity constraints to the ideal total inventory space demand according to the warehouse capacity information to obtain a capacity constraint evaluation result; Step S264: adjusting the highest inventory based on priority according to the capacity constraint evaluation result to obtain the highest inventory after priority adjustment; Step S265: Perform shelf life constraint verification according to the product shelf life and the highest inventory after priority adjustment to obtain the highest inventory after shelf life verification; Step S266: Determine the final maximum inventory limit based on the maximum inventory after priority adjustment, the maximum inventory after shelf life verification, and the capacity constraint evaluation results to obtain the maximum inventory limit.

[0012] The present invention calculates the ideal maximum inventory by combining the predicted demand and the target turnover rate, provides an initial reference value for subsequent inventory management, and helps to achieve the enterprise's inventory turnover target. By calculating the total storage space required for the ideal maximum inventory of all SKUs, the required storage resources can be estimated, providing a basis for warehouse capacity planning. The ideal inventory space demand is compared with the actual warehouse capacity to determine whether it exceeds the warehouse capacity limit, providing a decision basis for subsequent inventory adjustments. If the ideal inventory space demand exceeds the warehouse capacity, it is adjusted according to the turnover priority of the SKU, and the inventory of the high-turnover SKU is prioritized, thereby maximizing inventory efficiency within a limited warehouse space. The adjusted maximum inventory is checked for shelf life to ensure that the inventory will not cause product expiration, avoiding potential inventory losses, which is particularly important for commodities with a short shelf life. Taking into account the warehouse capacity limit and the shelf life constraint, the final maximum inventory limit is determined to ensure that the final inventory strategy not only meets the warehouse capacity limit, but also avoids product expiration, thereby achieving refined and efficient inventory management.

[0013] Preferably, step S3 comprises the following steps: Step S31: Acquire real-time equipment status data and product process information; Step S32: Decomposing the predicted demand and the dynamic inventory threshold into production demand to obtain a to-be-produced list; Step S33: Perform production resource evaluation based on the to-be-produced list, real-time equipment status data, and product process information to obtain an equipment capacity list; Step S34: Optimize the production batches according to the to-be-produced list and product process information to obtain an optimized batch plan; Step S35: Arrange the production sequence of the optimized batch plan according to the equipment capacity list and product process information to obtain a production sequence plan; Step S36: Formulate a detailed production plan according to the production sequence plan to obtain a production plan.

[0014] The present invention provides basic data for the formulation of subsequent production plans by acquiring real-time equipment status data and product process information, ensuring that the formulation of production plans is based on the latest equipment status and product process requirements, and improving the executability of the plan. The predicted demand and dynamic inventory threshold are converted into specific quantities to be produced, the production target is clarified, and guidance is provided for subsequent production planning arrangements, avoiding blind production and inventory backlogs. By evaluating the availability and capacity of production resources, production tasks can be arranged more reasonably, equipment utilization can be improved, and production delays due to insufficient resources can be avoided. By optimizing production batches, equipment switching time and cost can be minimized, production efficiency can be improved, and production costs can be reduced. Considering equipment capacity and product process information, sorting production batches can optimize the production process, maximize equipment utilization, and minimize production delays, thereby improving overall production efficiency. The production sequence plan is converted into a detailed production plan, accurate to the start and end time of each batch, and the equipment used, providing clear guidance for production execution and ensuring the smooth execution of the production plan.

[0015] Preferably, step S34 includes the following steps: Step S341: extracting production tasks to be optimized from the production list to obtain production tasks to be optimized; constructing a switching cost matrix for product process information to obtain a switching cost matrix; Step S342: generating a representation of a solution based on an optimization algorithm for the production task to be optimized, and obtaining a batch plan coding scheme; Step S343: constructing an objective function according to the switching cost matrix and the batch plan coding scheme to obtain a batch plan evaluation index; Step S344: Constraint condition modeling is performed according to the production task to be optimized and the product process information to obtain production constraint rules; Step S345: Execute the optimization algorithm according to the batch plan coding scheme, batch plan evaluation indicators and production constraint rules to obtain optimization process data; Step S346: Analyze and output the optimization process data to obtain an optimized batch plan.

[0016] The present invention provides necessary input data for the subsequent batch optimization algorithm by extracting the production tasks to be optimized and constructing the switching cost matrix, and clarifies the optimization objectives and constraints. The batch plan is encoded as a chromosome, and a bridge is established between the optimization problem and the genetic algorithm, so that the genetic algorithm can be used to search for the optimal batch plan. The objective function (total switching time) is constructed to provide the genetic algorithm with a standard for evaluating the advantages and disadvantages of different batch plans, and guide the algorithm to search towards the goal of minimizing the switching time. By establishing constraints, such as demand constraints and equipment compatibility constraints, it is ensured that the generated batch plan is feasible and meets the constraints of actual production. The genetic algorithm is executed to continuously optimize the batch plan through iterative search, and finally the plan with the minimum total switching time is found, thereby improving production efficiency.

[0017] Preferably, step S35 includes the following steps: Step S351: performing batch readiness status evaluation on the optimized batch plan according to the equipment capability list to obtain a batch readiness status list; Step S352: setting target weights according to preset production targets to obtain optimized target weights; Step S353: construct a multi-objective optimization function for the optimization target weights to obtain a comprehensive evaluation function; Step S354: Execute equipment-dependent production sequence according to product process information, optimized batch plan and equipment capacity list; Step S355: using a comprehensive evaluation function to optimize the equipment dependent production sequence to obtain a preliminary production sequence; Step S356: fine-tune the preliminary production sequence according to the equipment capability list, and perform time allocation to obtain a production sequence plan.

[0018] The present invention can avoid arranging unready batches into the production plan by evaluating the readiness of the batches, thereby reducing production delays and improving the executability of the plan. By setting weights according to preset production goals, the production plan can be more in line with the overall goals of the enterprise. For example, if the current goal is to maximize equipment utilization, the weight of equipment utilization can be increased. By constructing a comprehensive evaluation function, multiple production goals, such as equipment utilization, production delays, and low inventory priority, can be comprehensively considered, and balanced according to weights to find a better production sequence. Considering product process information and equipment capabilities, it can ensure that the production sequence meets the process flow requirements and equipment capacity limitations, and avoid production interruptions caused by process flow errors or equipment mismatches. By using a comprehensive evaluation function to evaluate and sort all possible production sequences, the optimal production sequence can be found, thereby achieving the best balance of multiple production goals. Fine-tuning the preliminary production sequence and allocating specific time and equipment can further optimize the production plan, improve equipment utilization, and reduce production time, and ultimately generate an executable production sequence plan.

[0019] Preferably, step S4 comprises the following steps: Step S41: obtaining current inventory status data; performing real-time inventory monitoring on the current inventory status data to obtain the latest inventory snapshot; Step S42: Identify replenishment demand based on the latest inventory snapshot and dynamic inventory threshold to obtain a to-be-replenished list; Step S43: traverse the replenishment list and check the production plan to obtain a production matching result; Step S44: Generate automatic replenishment instructions based on the to-be-replenished list, production matching results, and production plan to obtain inventory replenishment instructions; Step S45: Perform intelligent cargo space allocation according to the inventory replenishment instruction, the latest inventory snapshot and the product attribute list to obtain a cargo space allocation plan; Step S46: issuing storage tasks for the cargo space allocation plan to obtain storage operation instructions.

[0020] The present invention can timely grasp inventory changes by real-time monitoring of inventory status, provide accurate data support for subsequent replenishment decisions, and avoid decision-making errors caused by data lag. By comparing with the dynamic inventory threshold, the SKU that needs to be replenished can be automatically identified, reducing manual intervention, improving replenishment efficiency, and avoiding out-of-stock or overstocking caused by human negligence. By checking the list of to-be-replenished goods with the production plan, it can be understood which SKUs are already in the production plan to avoid duplicate production, and the arrival time of replenishment can be estimated according to the completion time of the production plan. Replenishment instructions are automatically generated according to the production plan and inventory status, which realizes the automation of the replenishment process, improves the replenishment efficiency, and reduces human errors. At the same time, emergency replenishment needs are considered to ensure the supply of key materials. Intelligent cargo space allocation is performed according to the attributes of the product and the real-time status of the warehouse, which can optimize the warehouse layout, improve the utilization rate of the warehouse space, and facilitate subsequent picking and outbound operations. The cargo space allocation plan is converted into a specific warehousing operation instruction to guide the automated equipment to perform the warehousing operation, realize the automation of the warehousing operation, improve the operation efficiency, and reduce the labor cost.

[0021] Preferably, step S44 includes the following steps: Step S441: Generate a planned replenishment instruction according to the to-be-replenished list, the production matching result and the production plan to obtain a planned replenishment instruction; Step S442: evaluating the unplanned replenishment demand of the to-be-replenished list according to the production matching result, and obtaining an unplanned replenishment demand list; Step S443: generating temporary replenishment instructions for the unplanned replenishment demand list according to the dynamic inventory threshold to obtain temporary replenishment instructions; Step S444: merge and optimize the planned replenishment instructions and the temporary replenishment instructions to obtain the inventory replenishment instructions.

[0022] The present invention can achieve seamless connection between production and replenishment by generating replenishment instructions for SKUs that are already in the production plan, thereby improving the overall efficiency of the supply chain and reducing unnecessary communication and coordination costs. Identifying unplanned replenishment needs can timely discover potential out-of-stock risks and provide a decision-making basis for subsequent emergency replenishment, thereby avoiding production interruptions or delayed delivery due to unplanned needs. Generating temporary replenishment instructions for urgent unplanned replenishment needs can quickly respond to market changes or emergencies, ensure the timely supply of materials in emergency situations, and avoid causing greater losses. Merging and optimizing all replenishment instructions can avoid duplicate or conflicting instructions, simplify warehousing operation processes, and improve operating efficiency.

[0023] Preferably, step S5 comprises the following steps: Step S51: Execute the warehouse system instruction according to the warehouse operation instruction to obtain the warehouse operation data; obtain the actual sales data from the sales system to obtain the actual sales data; Step S52: collecting key indicators of the latest inventory snapshot, production plan, warehouse operation data and actual sales data to obtain comprehensive performance data; Step S53: Calculate the performance index based on the comprehensive performance data to obtain a quantitative performance index; Step S54: using the quantitative performance indicators to perform performance deviation analysis on the forecast demand, dynamic inventory threshold and production plan to obtain a deviation analysis report; Step S55: Obtain historical parameter configuration; perform intelligent parameter optimization evaluation based on the reinforcement learning model according to the historical parameter configuration, deviation analysis report and comprehensive performance data to obtain parameter optimization evaluation results; Step S56: using the parameter optimization evaluation results to adjust the optimization parameters and obtain the optimization parameters; Step S57: Monitor the optimization effect of the optimization parameters to achieve the optimization task of the intelligent warehousing system.

[0024] The present invention provides actual operation data for subsequent performance evaluation and optimization by executing warehouse operation instructions and obtaining actual sales data, which is the basis of data-driven optimization. Various key indicator data, such as inventory level, production output, sales volume and warehouse operation efficiency, are collected to provide a data basis for comprehensive evaluation of system performance. The collected data are converted into quantitative performance indicators, such as inventory turnover rate, out-of-stock rate, equipment utilization rate and prediction accuracy, etc., which can more intuitively evaluate system performance and provide a basis for subsequent deviation analysis. By analyzing the deviation between actual performance and plan, problems in the system can be identified, such as inaccurate prediction, unreasonable inventory strategy or inefficient production plan, etc., to provide direction for subsequent optimization. Using reinforcement learning model for intelligent parameter optimization evaluation, system parameters, such as prediction model parameters, safety stock coefficient and replenishment lead time, etc., can be automatically learned and adjusted, so as to continuously improve system performance. Applying the optimized parameters to the system can actually improve system operation, and save the optimized parameters to facilitate subsequent continuous optimization. Continuous monitoring of optimization results can promptly identify new problems that arise during the optimization process and make adjustments based on actual conditions to form a closed-loop optimization process, thereby continuously improving the overall performance of the intelligent warehousing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic diagram of the steps of an optimization method for an intelligent warehousing system based on production equipment linkage; Figure 2 Detailed implementation flow chart of step S2 in the present invention; Figure 3 It is a schematic diagram of the detailed implementation steps of step S3 in the present invention.

[0026] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0027] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0028] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0029] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0030] To achieve this, please refer to Figures 1 to 3 , an intelligent warehousing system optimization method based on production equipment linkage, comprising the following steps: Step S1: collect multi-source data from the sales system and the production system, and perform dynamic demand forecasting to obtain forecasted demand; Step S2: Obtain a product attribute list; calculate the safety stock quantity according to the predicted demand and the product attribute list to obtain a safety stock benchmark; estimate the maximum inventory quantity according to the predicted demand, the product attribute list and the safety stock benchmark to obtain a maximum inventory limit; adjust the dynamic threshold according to the maximum inventory limit to obtain a dynamic inventory threshold; Step S3: Decompose the production demand based on the predicted demand and the dynamic inventory threshold to obtain a to-be-produced list; optimize the production batches based on the to-be-produced list to obtain an optimized batch plan; arrange the production sequence for the optimized batch plan, and formulate a detailed production plan to obtain a production plan; Step S4: Obtain the latest inventory snapshot of the current inventory status; identify replenishment needs based on the latest inventory snapshot and dynamic inventory thresholds, and perform production plan verification to obtain a to-be-replenished list and production matching results; automatically generate replenishment instructions based on the production plan, to-be-replenished list, and production matching results to obtain inventory replenishment instructions; perform intelligent storage allocation based on the inventory replenishment instructions and the product attribute list, and issue warehousing tasks to obtain warehousing operation instructions; Step S5: Acquire actual sales data; execute warehouse system instructions according to warehouse operation instructions to obtain warehouse operation data; calculate performance indicators according to the latest inventory snapshot, production plan, actual sales data and warehouse operation data to obtain quantitative performance indicators; use quantitative performance indicators to perform performance deviation analysis on predicted demand, dynamic inventory thresholds and production plans to obtain a deviation analysis report; optimize performance according to the deviation analysis report, and monitor the optimization effect to achieve the optimization task of the intelligent warehouse system.

[0031] In the embodiment of the present invention, reference Figure 1 As shown, it is a schematic diagram of the steps of the intelligent warehousing system optimization method based on production equipment linkage of the present invention. In this example, the intelligent warehousing system optimization method based on production equipment linkage includes the following steps: Step S1: collect multi-source data from the sales system and the production system, and perform dynamic demand forecasting to obtain forecasted demand; In the embodiment of the present invention, the sales system, production system and market research data are integrated to extract time features, sales features, promotion features and production features to construct a feature data set. The ARIMA model is used to perform time series pattern analysis to obtain time series forecast results, and the LSTM neural network model is used to perform product sales forecast results to obtain product sales forecast results. The two forecast results are integrated and calibrated according to real-time production data to finally obtain the predicted demand; Step S2: Obtain a product attribute list; calculate the safety stock quantity according to the predicted demand and the product attribute list to obtain a safety stock benchmark; estimate the maximum inventory quantity according to the predicted demand, the product attribute list and the safety stock benchmark to obtain a maximum inventory limit; adjust the dynamic threshold according to the maximum inventory limit to obtain a dynamic inventory threshold; In an embodiment of the present invention, a product attribute list is obtained, including a target turnover rate, a single product storage space, and a product shelf life. A safety stock benchmark is calculated based on the predicted demand, the product attribute list, and the preset service level. A minimum inventory threshold is set in combination with the safety stock benchmark, demand characteristic parameters, and replenishment lead time. The maximum inventory limit is calculated by comprehensively considering the predicted demand, the product attribute list, warehouse capacity information, the target turnover rate, and the shelf life constraints. Finally, the safety stock benchmark, the minimum inventory threshold, and the maximum inventory limit are dynamically adjusted according to the market activity forecast data to obtain the final dynamic inventory threshold; Step S3: Decompose the production demand based on the predicted demand and the dynamic inventory threshold to obtain a to-be-produced list; optimize the production batches based on the to-be-produced list to obtain an optimized batch plan; arrange the production sequence for the optimized batch plan, and formulate a detailed production plan to obtain a production plan; In an embodiment of the present invention, the quantity to be produced is calculated based on the predicted demand, dynamic inventory threshold and current inventory, and a to-be-produced list is generated. Real-time equipment status data and product process information are obtained, production resources are evaluated, and an equipment capacity list is generated. Based on the to-be-produced list and product process information, a genetic algorithm is used to optimize production batches, minimize equipment switching time, and obtain an optimized batch plan. Considering the equipment capacity list, product process information and multi-objective optimization function (equipment utilization, production delays, low inventory priority), the optimized batch plan is arranged in production sequence to obtain a production sequence plan. Finally, a detailed production plan is formulated according to the production sequence plan, including the SKU, quantity, production equipment, start time and end time of each batch; Step S4: Obtain the latest inventory snapshot of the current inventory status; identify replenishment needs based on the latest inventory snapshot and dynamic inventory thresholds, and perform production plan verification to obtain a to-be-replenished list and production matching results; automatically generate replenishment instructions based on the production plan, to-be-replenished list, and production matching results to obtain inventory replenishment instructions; perform intelligent storage allocation based on the inventory replenishment instructions and the product attribute list, and issue warehousing tasks to obtain warehousing operation instructions; In an embodiment of the present invention, real-time inventory data is acquired and the latest inventory snapshot is generated. Replenishment needs are identified based on the latest inventory snapshot and dynamic inventory thresholds, and a to-be-replenished list is generated. The to-be-replenished list is compared with the production plan to obtain a production matching result. Automatic replenishment instructions are generated based on the to-be-replenished list, the production matching result, and the production plan, including planned replenishment instructions and temporary replenishment instructions generated based on emergency thresholds, and the two are combined and optimized to obtain the final inventory replenishment instructions. Finally, intelligent storage location allocation is performed based on the inventory replenishment instructions, the product attribute list, and the real-time storage location information of the warehouse, and warehousing operation instructions are issued; Step S5: Acquire actual sales data; execute warehouse system instructions according to warehouse operation instructions to obtain warehouse operation data; calculate performance indicators according to the latest inventory snapshot, production plan, actual sales data and warehouse operation data to obtain quantitative performance indicators; use quantitative performance indicators to perform performance deviation analysis on predicted demand, dynamic inventory threshold and production plan to obtain a deviation analysis report; optimize performance according to the deviation analysis report and monitor the optimization effect to achieve the optimization task of the intelligent warehouse system; In an embodiment of the present invention, actual sales data and warehouse operation data are obtained. The latest inventory snapshot, production plan, actual sales data and warehouse operation data are collected, and various performance indicators such as inventory turnover rate, out-of-stock rate, order picking accuracy, equipment utilization rate and forecast accuracy are calculated. The actual performance indicators are compared with the planned values, the causes of deviations are analyzed, and a deviation analysis report is generated. Using historical parameter configurations, deviation analysis reports and comprehensive performance data, intelligent parameter optimization evaluation is performed based on the reinforcement learning model, and optimization parameter adjustments are implemented. Finally, the optimization effect is monitored, and continuous improvement and optimization are performed as needed.

[0032] Preferably, step S1 comprises the following steps: Step S11: Collect and fuse multi-source data from the sales system and the production system to obtain an integrated data set; Step S12: extracting data features from the integrated data set to obtain a feature data set; Step S13: performing time series pattern analysis on the feature data set to obtain a time series prediction result; Step S14: performing product sales prediction based on the LSTM neural network model according to the feature data set to obtain product sales prediction results; Step S15: Fusing the time series prediction results and the product sales prediction results to obtain a fused prediction result; Step S16: Acquire real-time production data; calibrate the fusion prediction results according to the real-time production data to obtain the predicted demand.

[0033] In an embodiment of the present invention, through a preset API interface, the sales records of spicy snack foods for the past 365 days are automatically extracted from the sales system database every hour, including information such as sales time, SKU code, sales quantity, sales channel, and sales area. At the same time, the production records of the past 365 days are extracted from the production system database every hour, including information such as production time, SKU code, production quantity, production line number, and equipment status. In addition, market trend data for the past 365 days are extracted from the market research database, including information such as consumer preferences, competitor analysis, and market size. The data from the above three data sources are aligned and merged according to timestamps and SKU codes, and duplicate data and missing values ​​are removed, finally forming an integrated data set containing information such as time, SKU, sales volume, production volume, and market trends, which is stored in the distributed database Hadoop.

[0034] Use Python language and Pandas library to perform feature engineering on the integrated dataset. First, extract time features, such as converting the sales date into year, month, day, day of the week, quarter, whether it is a holiday, etc. Second, calculate sales features, such as the moving average sales volume in the past 7 days, 30 days, and 90 days, the year-on-year and month-on-month growth rates of sales, and the standard deviation of sales. Third, perform unique hot encoding on promotional activities, such as converting different types of promotional activities (full reduction, discount, buy-one-get-one-free, etc.) into 0 / 1 variables. Finally, extract production features, such as the average output, equipment utilization, and production line downtime in the past 7 days, 30 days, and 90 days. Merge the extracted features with the original data to form a feature dataset and store it as a CSV file.

[0035] Use the ARIMA model in the Statsmodels library to perform time series analysis on the historical sales data of each SKU. First, determine the p, d, and q parameters of the ARIMA model based on the ACF and PACF graphs. Then, use the historical sales data in the feature dataset to train the ARIMA model. Finally, use the trained ARIMA model to predict the sales volume for the next 7 days and save the prediction results in the database as time series prediction results.

[0036] Use TensorFlow and Keras libraries to build an LSTM neural network model. Divide the feature dataset into a training set and a test set with a ratio of 8:2. Use the training set to train the LSTM model, and use the test set to evaluate the model's prediction performance, such as using indicators such as root mean square error (RMSE) and mean absolute percentage error (MAPE). Use the trained LSTM model to predict sales for the next 7 days, and save the prediction results in the database as product sales prediction results.

[0037] Read the time series prediction results of the ARIMA model and the product sales prediction results of the LSTM model from the database. Use the weighted average method to fuse the two prediction results. According to the RMSE index of the two models on the test set, dynamically adjust the weight. The smaller the RMSE, the larger the weight. Save the fused prediction results to the database as the fused prediction results.

[0038] The production data of each production line is collected in real time through the industrial Internet of Things platform, including equipment status, output, production speed and other information, and the maximum possible output in the next 7 days is calculated. The fusion prediction result is compared with the maximum possible output. If the fusion prediction result exceeds the maximum possible output, the predicted demand is adjusted to the maximum possible output; otherwise, the predicted demand is equal to the fusion prediction result. The final predicted demand is saved in the database for use in subsequent steps.

[0039] Preferably, step S2 comprises the following steps: Step S21: Analyze the demand characteristics of the predicted demand to obtain demand characteristic parameters; Step S22: using the sales system and the production system to evaluate product attributes and obtain a product attribute list, wherein the product attribute list includes a target turnover rate, a single product storage space, and a product shelf life; Step S23: Calculate the safety stock quantity according to the demand characteristic parameters and the product attribute list to obtain the safety stock benchmark; Step S24: setting the minimum inventory quantity according to the safety inventory benchmark and the demand characteristic parameters to obtain the minimum inventory threshold; Step S25: Obtain warehouse capacity information and market activity forecast data; Step S26: Calculate the maximum inventory quantity based on the predicted demand, product attribute list and warehouse capacity information to obtain the maximum inventory limit; Step S27: Adjust the dynamic threshold according to the safety stock benchmark, the minimum stock threshold, the maximum stock limit and the market activity forecast data to obtain the dynamic stock threshold.

[0040] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: Analyze the demand characteristics of the predicted demand to obtain demand characteristic parameters; In this embodiment of the present invention, the forecast demand data for each SKU in the next 7 days generated in step S1 is read from the database. The mean, standard deviation, coefficient of variation, maximum and minimum values ​​of the forecast demand for each SKU are calculated using Python language and NumPy library. The calculation results are stored in a new data table as demand characteristic parameters for subsequent calculation of safety stock and minimum stock; Step S22: using the sales system and the production system to evaluate product attributes and obtain a product attribute list, wherein the product attribute list includes a target turnover rate, a single product storage space, and a product shelf life; In the embodiment of the present invention, the target turnover rate of each SKU is read from the product database (for example, set to 4 times per month), in units of times / month. The single product storage space of each SKU is read from the warehouse management system (for example, 0.01 cubic meters). The product shelf life of each SKU is read from the production system database (for example, 180 days). This information is integrated into a product attribute list data table, which includes fields such as SKU code, target turnover rate, single product storage space, and product shelf life; Step S23: Calculate the safety stock quantity according to the demand characteristic parameters and the product attribute list to obtain the safety stock benchmark; In the embodiment of the present invention, the standard deviation of demand for each SKU in the demand characteristic parameters and the shelf life of the product in the product attribute list are read. The service level is set to 95%, and the corresponding Z value is 1.65. The formula used is: safety stock = Z value * standard deviation of demand * sqrt (average lead time), where the average lead time is set to 3 days. In order to avoid the impact of the shelf life, if the calculated safety stock is greater than the sales volume corresponding to half the shelf life (i.e., shelf life / 2*average daily demand), the safety stock will be adjusted to the sales volume corresponding to half the shelf life. The calculation results are stored in the safety stock benchmark data table; Step S24: setting the minimum inventory quantity according to the safety inventory benchmark and the demand characteristic parameters to obtain the minimum inventory threshold; In the embodiment of the present invention, the safety stock benchmark data table and the demand characteristic parameter data table are read. The formula used is: minimum inventory = safety stock + average daily demand * replenishment lead time, where the replenishment lead time is set to 2 days, and the average daily demand comes from the demand characteristic parameter calculated in step S21. The calculation result is stored in the minimum inventory threshold data table; Step S25: Obtain warehouse capacity information and market activity forecast data; In the embodiment of the present invention, the available capacity information of the warehouse is obtained in real time from the warehouse management system, such as the available storage capacity, the number of available cargo spaces, etc. The market activity forecast data for the next 7 days is obtained from the marketing department, such as the promotion activity plan, the expected sales increase ratio, etc., and stored in JSON format; Step S26: Calculate the maximum inventory quantity based on the predicted demand, product attribute list and warehouse capacity information to obtain the maximum inventory limit; In the embodiment of the present invention, the ideal maximum inventory is calculated based on the predicted demand and the target turnover rate. Then, the total warehouse space required for these ideal inventories is evaluated. If the total space required exceeds the available warehouse capacity, the maximum inventory is adjusted according to the target turnover rate of the SKU (high turnover rate is preferred) until the total inventory space meets the warehouse capacity limit. Finally, the adjusted maximum inventory is checked for shelf life to ensure that even if the maximum inventory is stored, it will not exceed the shelf life limit, and finally the maximum inventory limit for each SKU is obtained; Step S27: adjusting the dynamic threshold according to the safety stock benchmark, the minimum stock threshold, the maximum stock limit and the market activity forecast data to obtain the dynamic stock threshold; In the embodiment of the present invention, the safety stock benchmark, the minimum stock threshold, the maximum stock limit and the market activity forecast data are read. If the market activity forecast data shows that there will be a large-scale promotion in the next 7 days, and the sales volume is expected to increase by 50%, the safety stock and the maximum stock limit will be increased by 50% accordingly. Ensure that the adjusted maximum stock limit does not exceed the available capacity of the warehouse. The adjusted safety stock, minimum stock and maximum stock are stored in the dynamic inventory threshold data table as the final dynamic inventory threshold.

[0041] Preferably, step S26 comprises the following steps: Step S261: Calculate the ideal maximum inventory based on the predicted demand and the target turnover rate to obtain an ideal maximum inventory list; Step S262: Evaluate the ideal inventory space demand for the ideal maximum inventory list and the single product storage space to obtain the ideal total inventory space demand; Step S263: applying warehouse capacity constraints to the ideal total inventory space demand according to the warehouse capacity information to obtain a capacity constraint evaluation result; Step S264: adjusting the highest inventory based on priority according to the capacity constraint evaluation result to obtain the highest inventory after priority adjustment; Step S265: Perform shelf life constraint verification according to the product shelf life and the highest inventory after priority adjustment to obtain the highest inventory after shelf life verification; Step S266: Determine the final maximum inventory limit based on the maximum inventory after priority adjustment, the maximum inventory after shelf life verification, and the capacity constraint evaluation results to obtain the maximum inventory limit.

[0042] In this embodiment of the present invention, the forecast demand data for each SKU in the next 7 days generated in step S1 and the target turnover rate data in the product attribute list obtained in step S22 are read from the database. Use Python script to traverse each SKU, and the calculation formula is: ideal maximum inventory = sum of forecast demand in the next 7 days / (target turnover rate / 30*7). The calculation result is retained to two decimal places and stored in a database table named "Ideal Maximum Inventory List", which contains two fields: SKU code and ideal maximum inventory.

[0043] Read the ideal maximum inventory of each SKU from the "Ideal Maximum Inventory List" database table, and read the single product storage space of each SKU from the product attribute list obtained in step S22. Use the Python script to traverse each SKU, multiply the ideal maximum inventory by the single product storage space, and get the ideal space occupied by each SKU. Add up the ideal space occupied by all SKUs to get the total ideal inventory space requirement in cubic meters, with two decimal places. Store the calculation result in a variable named "Ideal Total Inventory Space Requirement".

[0044] Read the available capacity information of the current warehouse from the warehouse management system database in cubic meters. Compare the value of the variable "Ideal Inventory Space Total Demand" with the warehouse available capacity. If the ideal inventory space total demand is less than or equal to the warehouse available capacity, set the capacity constraint evaluation result to "No Adjustment Required", indicating that the ideal maximum inventory quantity does not need to be adjusted; otherwise, set the capacity constraint evaluation result to "Need Adjustment", indicating that the ideal maximum inventory quantity needs to be adjusted. Store the evaluation result in a variable named "Capacity Constraint Evaluation Result".

[0045] If the "Capacity Constraint Evaluation Result" is "Needs Adjustment", perform the following operations: Read the target turnover rate for each SKU from the product attribute list, and sort the SKUs from high to low according to the target turnover rate. Read the ideal maximum inventory quantity for each SKU from the "Ideal Maximum Inventory List". Use a Python script to calculate the adjusted maximum inventory quantity for each SKU in the sorted SKU order. Initially, the adjusted maximum inventory quantity is equal to the ideal maximum inventory quantity. Reduce the maximum inventory quantity of low-turnover SKUs one by one (for example, reduce it by 1 unit at a time), and recalculate the total inventory space requirement until the total inventory space requirement is less than or equal to the available capacity of the warehouse. Store the adjusted maximum inventory quantity in a database table named "Priority Adjusted Maximum Inventory", which contains two fields: SKU code and adjusted maximum inventory quantity.

[0046] Read the shelf life of each SKU from the product attribute list, and read the adjusted maximum inventory of each SKU from the "Priority Adjusted Maximum Inventory" database table. Use a Python script to traverse each SKU and calculate the inventory turnover days corresponding to the adjusted maximum inventory: Turnover Days = Adjusted Maximum Inventory / Average Daily Demand. If the turnover days are greater than 80% of the shelf life, adjust the adjusted maximum inventory to shelf life * 80% / average daily demand, and round the result down. Store the verified maximum inventory in a database table named "Maximum Inventory after Shelf Life Verification", which contains two fields: SKU code and verified maximum inventory.

[0047] If the "Capacity Constraint Evaluation Result" is "No Adjustment Required", copy the data in the "Ideal Maximum Inventory List" to the "Maximum Inventory Limit" data table. If the "Capacity Constraint Evaluation Result" is "Need Adjustment", copy the data in the "Maximum Inventory after Shelf Life Verification" to the "Maximum Inventory Limit" data table. The "Maximum Inventory Limit" data table contains two fields: SKU Code and Maximum Inventory Limit.

[0048] Preferably, step S3 comprises the following steps: Step S31: Acquire real-time equipment status data and product process information; Step S32: Decomposing the predicted demand and the dynamic inventory threshold into production demand to obtain a to-be-produced list; Step S33: Perform production resource evaluation based on the to-be-produced list, real-time equipment status data, and product process information to obtain an equipment capacity list; Step S34: Optimize the production batches according to the to-be-produced list and product process information to obtain an optimized batch plan; Step S35: Arrange the production sequence of the optimized batch plan according to the equipment capacity list and product process information to obtain a production sequence plan; Step S36: Formulate a detailed production plan according to the production sequence plan to obtain a production plan.

[0049] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes: Step S31: Acquire real-time equipment status data and product process information; In the embodiment of the present invention, through the sensors and data acquisition system deployed on the production line, the equipment status data is collected once a minute, including the equipment ID, equipment status (running, idle, faulty), the currently produced SKU, production speed, output and other information, and the data is stored in the real-time database InfluxDB. The product process information is read from the production management system database, including the production process of each SKU, the required equipment, the standard production time, the equipment switching time between different SKUs and other information, and the data is stored in the relational database MySQL; Step S32: Decomposing the predicted demand and the dynamic inventory threshold into production demand to obtain a to-be-produced list; In the embodiment of the present invention, the predicted demand generated in step S1 and the dynamic inventory threshold (including safety stock, minimum inventory and maximum inventory) generated in step S2 are read from the database. Use Python script to calculate the number of items to be produced for each SKU: number of items to be produced = max(0, safety stock - current inventory + predicted demand in the next 7 days - maximum inventory). If the number of items to be produced is less than 0, it is set to 0. The calculation results are stored in a database table named "List of items to be produced", which contains fields such as SKU code, number of items to be produced and expected completion time (in the next 7 days); Step S33: Perform production resource evaluation based on the to-be-produced list, real-time equipment status data, and product process information to obtain an equipment capacity list; In the embodiment of the present invention, real-time equipment status data is read from the InfluxDB database, product process information is read from the MySQL database, and the list of SKUs to be produced is read from the "List to be Produced" database table. Use a Python script to traverse each device and calculate its available capacity for producing each SKU based on its current status and the available time window for the next 7 days (taking into account the planned downtime and maintenance time). The calculation results are stored in a database table called "Equipment Capacity List", which contains information such as the equipment ID, available time window, list of SKUs that can be produced, and the estimated maximum output of each SKU; Step S34: Optimize the production batches according to the to-be-produced list and product process information to obtain an optimized batch plan; In the embodiment of the present invention, the production tasks to be optimized and the SKU switching cost matrix are extracted. Then, a genetic algorithm is used for optimization. Possible batch planning schemes are encoded as chromosomes, the objective function (total switching time) is defined, and constraints (demand, equipment compatibility) are set. The genetic algorithm is executed and iteratively evolved to finally obtain the optimal batch plan, which includes the SKU combination and planned production quantity of each production batch; Step S35: Arrange the production sequence of the optimized batch plan according to the equipment capacity list and product process information to obtain a production sequence plan; In an embodiment of the present invention, the goal is to determine the optimal production sequence for different production batches to maximize equipment utilization. First, evaluate the readiness of each batch. Then, set the weights of the optimization objectives (such as equipment utilization, production delays, and low inventory priority) based on the production goals. Construct a comprehensive evaluation function that combines multiple optimization objectives and weights. Next, consider the equipment dependencies and use an optimization algorithm (such as a genetic algorithm) to sort the batches and find the production sequence that optimizes the comprehensive evaluation function. Finally, fine-tune the production sequence and assign specific start and end times and the equipment to be used for each batch to generate the final production sequence plan; Step S36: Formulate a detailed production plan according to the production sequence plan to obtain a production plan; In an embodiment of the present invention, the production sequence plan generated in step S35 is read from the database, which contains the SKU, quantity, production order, planned start time, planned end time, and allocated equipment of each production batch. Using a Python script, the production sequence plan is converted into a detailed production plan, accurate to the start and end time of each production batch, and taking into account factors such as equipment switching time and material preparation time. For example, if two adjacent production batches need to be produced on the same equipment, and equipment cleaning and parameter adjustment are required, a period of equipment switching time is inserted between the two batches. The final generated production plan contains detailed information such as the SKU, quantity, production line, equipment, start time, end time, expected output, equipment switching time, etc. of each production batch, and the results are stored in a database table named "Production Plan".

[0050] Preferably, step S34 includes the following steps: Step S341: extracting production tasks to be optimized from the production list to obtain production tasks to be optimized; constructing a switching cost matrix for product process information to obtain a switching cost matrix; Step S342: generating a representation of a solution based on an optimization algorithm for the production task to be optimized, and obtaining a batch plan coding scheme; Step S343: constructing an objective function according to the switching cost matrix and the batch plan coding scheme to obtain a batch plan evaluation index; Step S344: performing constraint modeling according to the production task to be optimized and the product process information to obtain production constraint rules; Step S345: executing the optimization algorithm according to the batch plan coding scheme, the batch plan evaluation index and the production constraint rules to obtain optimization process data; Step S346: Analyze and output the optimization process data to obtain an optimized batch plan.

[0051] In an embodiment of the present invention, a list of SKUs to be produced and their corresponding required quantities are extracted from the "list to be produced" database table to generate a list of production tasks to be optimized, including SKU codes and quantities to be produced. Product process information is read from the MySQL database to extract the equipment information required for the production of each SKU, as well as the switching time required to switch between different SKUs on the same equipment. A switching cost matrix is ​​constructed using a Python script, in which the rows and columns of the matrix represent different SKUs, and the values ​​of the matrix elements represent the time required to switch from one SKU to another on the same equipment. If the two SKUs are not produced on the same equipment, the switching cost is set to infinity.

[0052] Using a genetic algorithm, a batch plan is encoded into a chromosome. The chromosome consists of genes, and each gene represents a production batch. Each gene contains two parts: SKU code and production quantity. For example, gene (SKU_A, 100) represents a batch of 100 SKU_A products. A chromosome is, for example, [(SKU_A, 100), (SKU_B, 200), (SKU_A, 50)], which means that 100 SKU_A are produced first, then 200 SKU_B, and finally 50 SKU_A, forming three batches.

[0053] The objective function is defined as the total switching time of the batch plan. According to the batch plan encoding scheme of step S342, each gene in the chromosome is traversed to calculate the switching time between adjacent genes on the same device. For example, switching from (SKU_A, 100) to (SKU_B, 200), if SKU_A and SKU_B are produced on the same device, then find the switching time from SKU_A to SKU_B in the switching cost matrix; if they are not produced on the same device, the switching time is 0. The switching time between all adjacent batches is accumulated to obtain the total switching time corresponding to the chromosome, which is the batch plan evaluation index.

[0054] Define two constraint rules: 1. Demand constraint: The total production quantity of each SKU must be equal to the demand quantity specified in the to-be-produced list. 2. Equipment compatibility constraint: The SKUs in each batch must be able to be produced on the assigned equipment. During the iteration process of the genetic algorithm, if the generated chromosome violates these constraint rules, the chromosome is considered an invalid solution.

[0055] The genetic algorithm is implemented using Python's DEAP library. The population size is set to 100, the crossover probability is 0.8, and the mutation probability is 0.2. The number of iterations is set to 1000. In each iteration, the fitness of each chromosome (batch plan) is evaluated according to the objective function (total switching time), and selection, crossover, and mutation operations are performed to generate a new population. The best fitness value and the corresponding chromosome of each iteration are recorded as optimization process data.

[0056] The best chromosome of the final iteration is extracted from the optimization process data and decoded into a production batch plan. The plan contains the SKU, planned production quantity, and batch sequence of each production batch. The results are stored in a database table called "Optimized Batch Plan", which contains fields such as batch ID, SKU code, planned production quantity, and batch sequence.

[0057] Preferably, step S35 includes the following steps: Step S351: performing batch readiness status evaluation on the optimized batch plan according to the equipment capability list to obtain a batch readiness status list; Step S352: setting target weights according to preset production targets to obtain optimized target weights; Step S353: construct a multi-objective optimization function for the optimization target weights to obtain a comprehensive evaluation function; Step S354: Execute equipment-dependent production sequence according to product process information, optimized batch plan and equipment capacity list; Step S355: using a comprehensive evaluation function to optimize the equipment dependent production sequence to obtain a preliminary production sequence; Step S356: fine-tune the preliminary production sequence according to the equipment capability list, and perform time allocation to obtain a production sequence plan.

[0058] In the embodiment of the present invention, the available time window and the list of producible SKUs of each device are read from the "Device Capability List" database table. The SKU and planned production quantity of each batch are read from the "Optimized Batch Plan" database table. Using a Python script, each batch is traversed to check whether the required equipment is idle within the available time window and whether the equipment can produce the SKU of the batch. If the conditions are met, the readiness status of the batch is marked as "ready"; otherwise, it is marked as "not ready". The results are stored in the "Batch Ready Status List" data table, including the batch ID and the ready status.

[0059] The preset production goals include maximizing equipment utilization, minimizing production delays, and prioritizing the production of low inventory SKUs. Set weights for each of these three goals, for example, the weight of equipment utilization is 0.5, the weight of production delays is 0.3, and the weight of low inventory priority is 0.2. The weight values ​​are stored in the configuration file and can be adjusted according to actual production needs.

[0060] Based on the weights set in step S352, a comprehensive evaluation function is constructed. The function input is a production sequence, and the output is the comprehensive evaluation value of the sequence. The equipment utilization index is calculated as the ratio of the total working time of all equipment to the total available time. The production delay index is calculated as the sum of the differences between the completion time of all batches and the expected completion time. The low inventory priority index is calculated as the weighted sum of the number of days that the low inventory SKU batches are completed in advance before their expected completion time. The comprehensive evaluation function performs weighted summation of these three indicators according to the preset weights to obtain the final evaluation value.

[0061] Read product process information from the MySQL database to determine the production process and equipment dependencies of different SKUs. For example, the production of some SKUs requires the production of other SKUs to be completed first. Based on these dependencies, a directed acyclic graph (DAG) is constructed, where the nodes represent production batches and the edges represent the dependencies between batches.

[0062] Using the topological sorting algorithm, all possible production sequences are generated according to the DAG constructed in step S354. For each production sequence, its evaluation value is calculated using the comprehensive evaluation function constructed in step S353. The production sequence with the highest evaluation value is selected as the preliminary production sequence.

[0063] Read the available time window for each device from the "Equipment Capacity List" database table. Traverse each batch in the preliminary production sequence, and assign specific production equipment and time periods to each batch based on its SKU and planned production quantity, combined with the equipment capacity list. Consider the equipment switching time and reserve necessary switching time between adjacent batches. If a batch can be produced on multiple devices, select the device with the most appropriate available time window. Store the final production sequence, including the SKU, quantity, production equipment, start time and end time of each batch, in a database table called "Production Sequence Plan".

[0064] Preferably, step S4 comprises the following steps: Step S41: obtaining current inventory status data; performing real-time inventory monitoring on the current inventory status data to obtain the latest inventory snapshot; Step S42: Identify replenishment demand based on the latest inventory snapshot and dynamic inventory threshold to obtain a to-be-replenished list; Step S43: traverse the replenishment list and check the production plan to obtain a production matching result; Step S44: Generate automatic replenishment instructions based on the to-be-replenished list, production matching results, and production plan to obtain inventory replenishment instructions; Step S45: Perform intelligent cargo space allocation according to the inventory replenishment instruction, the latest inventory snapshot and the product attribute list to obtain a cargo space allocation plan; Step S46: issuing storage tasks for the cargo space allocation plan to obtain storage operation instructions.

[0065] In the embodiment of the present invention, the current inventory status data is collected every minute through RFID readers, sensors, cameras and other devices deployed in the warehouse, including SKU code, inventory quantity, shelf information, batch information, production date, etc. The collected data is transmitted to the warehouse management system database, and the latest inventory snapshot is generated and stored in a data table named "Latest Inventory Snapshot".

[0066] Read the current inventory quantity of each SKU from the "Latest Inventory Snapshot" data table. Read the minimum inventory threshold of each SKU from the "Dynamic Inventory Threshold" data table. Use a Python script to iterate over each SKU, and if the current inventory quantity is lower than the minimum inventory threshold, add the SKU to the to-be-replenished list. The to-be-replenished list contains the SKU code, the current inventory quantity, the minimum inventory threshold, and the quantity to be replenished (minimum inventory threshold - current inventory quantity).

[0067] Traverse each SKU in the replenishment list and search for the future production plan of the SKU in the "Production Plan" database table. If a matching production plan is found, record the planned production quantity and estimated completion time of the SKU; if no matching production plan is found, record the SKU as "unmatched". Store the results in the "Production Matching Result" data table, including SKU code, planned production quantity, estimated completion time and matching status.

[0068] For SKUs that are already in the production plan, generate planned replenishment instructions, specifying the source and expected entry time. Then, evaluate the replenishment needs of SKUs that are not in the production plan and generate a list of unplanned replenishment needs. For unplanned replenishment needs where the inventory is below the emergency threshold, generate urgent temporary replenishment instructions. Finally, merge and optimize all generated replenishment instructions (planned and temporary) (for example, merge instructions for the same SKU and sort by expected entry time) to generate the final inventory replenishment instructions.

[0069] Read the SKUs and quantities that need to be replenished from the "Inventory Replenishment Instructions" data table. Read the current shelf occupancy from the "Latest Inventory Snapshot" data table. Read the storage requirements of each SKU from the product attribute list, such as temperature, humidity, stacking height restrictions, etc. Use Python scripts and shelf allocation algorithms (such as rule-based algorithms or optimization-based algorithms) to allocate appropriate shelves for each SKU based on the storage requirements of the SKU and the available shelves in the warehouse, and generate a shelf allocation plan. The plan includes information such as SKU code, quantity, target shelf, etc.

[0070] The cargo allocation plan is converted into specific warehouse operation instructions, such as the driving path of AGV (Automated Guided Vehicle), the operation instructions of the robotic arm, etc. The instructions are sent to the corresponding automated equipment, such as AGV, RGV (Rail Guided Vehicle), stacker, etc. through the WMS (Warehouse Management System) system to perform the warehousing operation, and the execution results are fed back to the WMS system.

[0071] Preferably, step S44 includes the following steps: Step S441: Generate a planned replenishment instruction according to the to-be-replenished list, the production matching result and the production plan to obtain a planned replenishment instruction; Step S442: evaluating the unplanned replenishment demand of the to-be-replenished list according to the production matching result, and obtaining an unplanned replenishment demand list; Step S443: generating temporary replenishment instructions for the unplanned replenishment demand list according to the dynamic inventory threshold to obtain temporary replenishment instructions; Step S444: merge and optimize the planned replenishment instructions and the temporary replenishment instructions to obtain the inventory replenishment instructions.

[0072] In an embodiment of the present invention, the list and quantity of SKUs that need to be replenished are read from the "list to be replenished". The planned production quantity and estimated completion time of the matched SKU are read from the "production matching result". The production batch information of the corresponding SKU, including the production line and batch number, is read from the "production plan". A Python script is used to generate a planned replenishment instruction for each matched SKU. The instruction contains the SKU code, the quantity to be replenished, the source (production line and batch number), and the estimated warehousing time (completion time of the production plan). The generated instruction is stored in the "planned replenishment instruction" data table.

[0073] Read all SKUs that need to be replenished from the "To be replenished list". Read the matched SKU list from the "Production matching results". Use Python scripts to mark SKUs in the to be replenished list that do not appear in the matched SKU list as unplanned replenishment needs and add them to the unplanned replenishment needs list. The list contains SKU codes, quantities that need to be replenished, and priorities (determined by inventory and urgency of demand, for example, the lower the inventory and the more urgent the demand, the higher the priority).

[0074] Read the SKU and the quantity that needs to be replenished from the "Unplanned Replenishment Requirement List". Read the safety stock of each SKU from the "Dynamic Inventory Threshold". Using a Python script, if the current inventory of the SKU is less than half of the safety stock (set as the emergency replenishment threshold), a temporary replenishment instruction is generated. The instruction contains the SKU code, the quantity that needs to be replenished (the larger value of the difference between the demand and the safety stock and the minimum order quantity, where the minimum order quantity is preset to 50 pieces), the source (marked as "urgent replenishment", which requires manual designation of the production line), the estimated time of entry into the warehouse (as soon as possible, such as within 24 hours), and the status ("urgent"). The generated instructions are stored in the "Temporary Replenishment Instructions" data table, and an alarm is triggered to notify relevant personnel to handle the emergency replenishment needs.

[0075] All replenishment instructions are read from the "Planned Replenishment Instructions" and "Temporary Replenishment Instructions" data tables. Using Python scripts, replenishment instructions for the same SKU are first merged and the replenishment quantities are accumulated. Then, the merged instructions are sorted according to the expected entry time. Finally, the merged and sorted instructions are stored in the "Inventory Replenishment Instructions" data table as the final inventory replenishment instructions. This table contains information such as SKU code, total quantity to be replenished, source, expected entry time, and status.

[0076] Preferably, step S5 comprises the following steps: Step S51: Execute the warehouse system instruction according to the warehouse operation instruction to obtain the warehouse operation data; obtain the actual sales data from the sales system to obtain the actual sales data; Step S52: collecting key indicators of the latest inventory snapshot, production plan, warehouse operation data and actual sales data to obtain comprehensive performance data; Step S53: Calculate the performance index based on the comprehensive performance data to obtain a quantitative performance index; Step S54: using the quantitative performance indicators to perform performance deviation analysis on the forecast demand, dynamic inventory threshold and production plan to obtain a deviation analysis report; Step S55: Obtain historical parameter configuration; perform intelligent parameter optimization evaluation based on the reinforcement learning model according to the historical parameter configuration, deviation analysis report and comprehensive performance data to obtain parameter optimization evaluation results; Step S56: using the parameter optimization evaluation results to adjust the optimization parameters and obtain the optimization parameters; Step S57: monitoring the optimization effect of the optimization parameters to achieve the optimization task of the intelligent warehousing system; In the embodiment of the present invention, the warehousing operation instruction generated in step S46 is sent to the warehouse control system (WCS). WCS controls the automated equipment (such as AGV, RGV, stacker, etc.) according to the instruction to perform specific operations such as warehousing, warehousing, and transportation. The execution time, equipment status, material information, and other data of each operation are recorded to generate warehousing operation data and store it in the WMS database. At the same time, actual sales data is extracted from the sales system database every hour, including sales time, SKU code, sales quantity, sales channel, and other information.

[0077] Extract the actual inventory level from the "Latest Inventory Snapshot" data table. Extract the planned output and planned completion time from the "Production Plan" data table. Extract warehouse operation data from the WMS database, such as inbound quantity, outbound quantity, shelf utilization, order picking efficiency, etc. Extract actual sales data from the sales system database, such as sales volume, sales amount, etc. Integrate the above data together to form comprehensive performance data and store it in a data table called "Comprehensive Performance Data".

[0078] Use Python scripts to calculate various quantitative performance indicators based on comprehensive performance data. For example, inventory turnover rate = sales volume / average inventory volume; out-of-stock rate = number of out-of-stock times / total number of orders; order picking accuracy rate = number of correctly picked orders / total number of orders; equipment utilization rate = equipment operating time / total equipment available time; forecast accuracy rate = 1-mean absolute percentage error (MAPE). Store the calculation results in a data table named "Quantitative Performance Indicators".

[0079] Compare the actual value in the "Quantitative Performance Indicator" with the forecasted demand in step S1, the dynamic inventory threshold in step S2, and the production plan in step S3 to calculate the deviation value. For example, forecasted demand deviation = actual sales volume - forecasted demand; safety stock deviation = actual safety stock - dynamic safety stock threshold. Analyze the causes of the deviation, such as data fluctuations, unreasonable model parameter settings, external factors, etc., and record the analysis results in the deviation analysis report.

[0080] Read historical parameter configurations from the parameter configuration database, including forecasting model parameters, safety stock coefficients, replenishment lead times, etc. Use a reinforcement learning model (such as the Q-learning algorithm) to take historical parameter configurations, deviation analysis reports, and comprehensive performance data as input to evaluate system performance under different parameter configurations. The goal of the reinforcement learning model is to find the best parameter configuration to maximize long-term cumulative rewards (such as a weighted combination of indicators such as inventory turnover, out-of-stock rate, and cost). Train the reinforcement learning model to obtain parameter optimization evaluation results, including recommended parameter configurations and corresponding performance evaluation values.

[0081] The recommended parameter configuration obtained in step S55 is applied to the system, for example, updating the parameters of the forecasting model, adjusting the safety stock coefficient, modifying the replenishment lead time, etc. The new parameter configuration is stored in the parameter configuration database as the optimization parameter.

[0082] Continuously monitor the key performance indicators of the system, such as inventory turnover rate, out-of-stock rate, cost, etc. Compare the optimized performance indicators with the performance indicators before optimization to evaluate the optimization effect. If the optimization effect is not significant or new problems arise, return to step S54 and re-perform deviation analysis and parameter optimization to form a closed-loop optimization process.

[0083] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0084] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent warehousing system optimization method based on production equipment linkage, characterized in that: The following steps are involved: Step S1: collect multi-source data from the sales system and the production system, and perform dynamic demand forecasting to obtain forecasted demand; Step S2: Obtain a product attribute list; calculate the safety stock quantity based on the predicted demand and the product attribute list to obtain a safety stock benchmark; Calculate the maximum inventory quantity based on the forecast demand, product attribute list and safety stock benchmark to obtain the maximum inventory limit; Adjust the dynamic threshold according to the maximum inventory limit to obtain a dynamic inventory threshold; Step S3: Decompose the production demand based on the predicted demand and the dynamic inventory threshold to obtain a to-be-produced list; Optimize production batches according to the list to be produced to obtain an optimized batch plan; Arrange the production sequence for the optimized batch plan and make a detailed production plan to obtain the production plan; Step S4: Get the latest inventory snapshot of the current inventory status; Replenishment demand is identified based on the latest inventory snapshot and dynamic inventory threshold, and the production plan is checked to obtain the to-be-replenished list and production matching results; automatic replenishment instructions are generated based on the production plan, to-be-replenished list and production matching results to obtain inventory replenishment instructions; intelligent storage location allocation is performed based on the inventory replenishment instructions and product attribute list, and warehousing tasks are issued to obtain warehousing operation instructions; Step S5: Acquire actual sales data; execute warehouse system instructions according to warehouse operation instructions to obtain warehouse operation data; calculate performance indicators according to the latest inventory snapshot, production plan, actual sales data and warehouse operation data to obtain quantitative performance indicators; use quantitative performance indicators to perform performance deviation analysis on predicted demand, dynamic inventory threshold and production plan to obtain a deviation analysis report; Perform performance optimization based on the deviation analysis report and monitor the optimization effect to achieve the optimization task of the intelligent warehousing system.

2. The intelligent warehousing system optimization method based on production equipment linkage according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Collect and fuse multi-source data from the sales system and the production system to obtain an integrated data set; Step S12: extracting data features from the integrated data set to obtain a feature data set; Step S13: performing time series pattern analysis on the feature data set to obtain a time series prediction result; Step S14: performing product sales prediction based on the LSTM neural network model according to the feature data set to obtain product sales prediction results; Step S15: Fusing the time series prediction results and the product sales prediction results to obtain a fused prediction result; Step S16: Acquire real-time production data; calibrate the fusion prediction results according to the real-time production data to obtain the predicted demand.

3. The intelligent warehousing system optimization method based on production equipment linkage according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: Analyze the demand characteristics of the predicted demand to obtain demand characteristic parameters; Step S22: using the sales system and the production system to evaluate product attributes and obtain a product attribute list, wherein the product attribute list includes a target turnover rate, a single product storage space, and a product shelf life; Step S23: Calculate the safety stock quantity according to the demand characteristic parameters and the product attribute list to obtain the safety stock benchmark; Step S24: setting the minimum inventory quantity according to the safety inventory benchmark and the demand characteristic parameters to obtain the minimum inventory threshold; Step S25: Obtain warehouse capacity information and market activity forecast data; Step S26: Calculate the maximum inventory quantity based on the predicted demand, product attribute list and warehouse capacity information to obtain the maximum inventory limit; Step S27: Adjust the dynamic threshold according to the safety stock benchmark, the minimum stock threshold, the maximum stock limit and the market activity forecast data to obtain the dynamic stock threshold.

4. The intelligent warehousing system optimization method based on production equipment linkage according to claim 3 is characterized in that: Step S26 includes the following steps: Step S261: Calculate the ideal maximum inventory based on the predicted demand and the target turnover rate to obtain an ideal maximum inventory list; Step S262: Evaluate the ideal inventory space demand for the ideal maximum inventory list and the single product storage space to obtain the ideal total inventory space demand; Step S263: applying warehouse capacity constraints to the ideal total inventory space demand according to the warehouse capacity information to obtain a capacity constraint evaluation result; Step S264: adjusting the highest inventory based on priority according to the capacity constraint evaluation result to obtain the highest inventory after priority adjustment; Step S265: Perform shelf life constraint verification according to the product shelf life and the highest inventory after priority adjustment to obtain the highest inventory after shelf life verification; Step S266: Determine the final maximum inventory limit based on the maximum inventory after priority adjustment, the maximum inventory after shelf life verification, and the capacity constraint evaluation results to obtain the maximum inventory limit.

5. The intelligent warehousing system optimization method based on production equipment linkage according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: Acquire real-time equipment status data and product process information; Step S32: Decomposing the predicted demand and the dynamic inventory threshold into production demand to obtain a to-be-produced list; Step S33: Perform production resource evaluation based on the to-be-produced list, real-time equipment status data, and product process information to obtain an equipment capacity list; Step S34: Optimize the production batches according to the to-be-produced list and product process information to obtain an optimized batch plan; Step S35: Arrange the production sequence of the optimized batch plan according to the equipment capacity list and product process information to obtain a production sequence plan; Step S36: Formulate a detailed production plan according to the production sequence plan to obtain a production plan.

6. The intelligent warehousing system optimization method based on production equipment linkage according to claim 5 is characterized in that: Step S34 includes the following steps: Step S341: extracting production tasks to be optimized from the production list to obtain production tasks to be optimized; constructing a switching cost matrix for product process information to obtain a switching cost matrix; Step S342: generating a representation of a solution based on an optimization algorithm for the production task to be optimized, and obtaining a batch plan coding scheme; Step S343: constructing an objective function according to the switching cost matrix and the batch plan coding scheme to obtain a batch plan evaluation index; Step S344: performing constraint modeling according to the production task to be optimized and the product process information to obtain production constraint rules; Step S345: executing the optimization algorithm according to the batch plan coding scheme, the batch plan evaluation index and the production constraint rules to obtain optimization process data; Step S346: Analyze and output the optimization process data to obtain an optimized batch plan.

7. The intelligent warehousing system optimization method based on production equipment linkage according to claim 5 is characterized in that: Step S35 includes the following steps: Step S351: performing batch readiness status evaluation on the optimized batch plan according to the equipment capability list to obtain a batch readiness status list; Step S352: setting target weights according to preset production targets to obtain optimized target weights; Step S353: construct a multi-objective optimization function for the optimization target weights to obtain a comprehensive evaluation function; Step S354: Execute equipment-dependent production sequence according to product process information, optimized batch plan and equipment capacity list; Step S355: using a comprehensive evaluation function to optimize the equipment dependent production sequence to obtain a preliminary production sequence; Step S356: fine-tune the preliminary production sequence according to the equipment capability list, and perform time allocation to obtain a production sequence plan.

8. The intelligent warehousing system optimization method based on production equipment linkage according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: obtaining current inventory status data; performing real-time inventory monitoring on the current inventory status data to obtain the latest inventory snapshot; Step S42: Identify replenishment demand based on the latest inventory snapshot and dynamic inventory threshold to obtain a to-be-replenished list; Step S43: traverse the replenishment list and check the production plan to obtain a production matching result; Step S44: Generate automatic replenishment instructions based on the to-be-replenished list, production matching results, and production plan to obtain inventory replenishment instructions; Step S45: Perform intelligent cargo space allocation according to the inventory replenishment instruction, the latest inventory snapshot and the product attribute list to obtain a cargo space allocation plan; Step S46: issuing storage tasks for the cargo space allocation plan to obtain storage operation instructions.

9. The intelligent warehousing system optimization method based on production equipment linkage according to claim 8 is characterized in that: Step S44 includes the following steps: Step S441: Generate a planned replenishment instruction according to the to-be-replenished list, the production matching result and the production plan to obtain a planned replenishment instruction; Step S442: evaluating the unplanned replenishment demand of the to-be-replenished list according to the production matching result, and obtaining an unplanned replenishment demand list; Step S443: generating temporary replenishment instructions for the unplanned replenishment demand list according to the dynamic inventory threshold to obtain temporary replenishment instructions; Step S444: merge and optimize the planned replenishment instructions and the temporary replenishment instructions to obtain the inventory replenishment instructions.

10. The intelligent warehousing system optimization method based on production equipment linkage according to claim 1 is characterized in that: Step S5 includes the following steps: Step S51: Execute the warehouse system instruction according to the warehouse operation instruction to obtain the warehouse operation data; obtain the actual sales data from the sales system to obtain the actual sales data; Step S52: collecting key indicators of the latest inventory snapshot, production plan, warehouse operation data and actual sales data to obtain comprehensive performance data; Step S53: Calculate the performance index based on the comprehensive performance data to obtain a quantitative performance index; Step S54: using the quantitative performance indicators to perform performance deviation analysis on the forecast demand, dynamic inventory threshold and production plan to obtain a deviation analysis report; Step S55: Obtain historical parameter configuration; perform intelligent parameter optimization evaluation based on the reinforcement learning model according to the historical parameter configuration, deviation analysis report and comprehensive performance data to obtain parameter optimization evaluation results; Step S56: using the parameter optimization evaluation results to adjust the optimization parameters and obtain the optimization parameters; Step S57: Monitor the optimization effect of the optimization parameters to achieve the optimization task of the intelligent warehousing system.

Images