Intelligent warehouse location optimization method and system

Abstract

The invention provides an intelligent storage location optimization method and system, and belongs to the technical field of storage management, and the method comprises the steps: obtaining the data information of commodities in a storage system, constructing a storage location optimization model, comprising a comprehensive objective function taking maximization of space utilization rate and sorting efficiency and minimization of storage cost as objectives and constraint conditions based on physical rules and business rules; solving the storage location optimization model by using an optimization algorithm, and generating a storage location distribution scheme and prediction performance data; executing the storage location allocation scheme, and collecting actual storage location data and actual operation performance data; comparing the actual storage location data with the storage location allocation scheme to obtain a scheme execution coincidence rate; comparing the actual operation performance data with the predicted performance data to obtain a performance deviation rate; and based on the scheme execution coincidence rate and the performance deviation rate, adjusting the constraint condition of the storage location optimization model and / or the weight coefficient of the comprehensive objective function to obtain an updated storage location optimization model, and carrying out storage location allocation.

Description

Technical Field

[0001] This application relates to the technical field of warehouse management, and in particular to an intelligent warehouse location optimization method and system. Background Technology

[0002] In traditional warehouse location allocation, warehouse managers place goods in fixed areas based on attributes such as weight and volume. For example, heavier goods are placed near the ground or in locations with strong load-bearing capacity, while larger goods are stored in more spacious areas. Some warehouses also allocate locations based on sales frequency, storing best-selling items near the warehouse exit for easy and quick picking. Traditional location allocation methods only consider one or a few factors, failing to comprehensively consider multiple factors such as product attribute data, business data, and external environmental data. In actual operation, warehouse space utilization is low, picking efficiency is inefficient, warehousing costs remain high, and there is a lack of mechanisms to adjust location allocation schemes, making it impossible to optimize allocation in a timely manner based on actual operational conditions and adapt to the complex and ever-changing needs of warehousing operations.

[0003] Therefore, there is an urgent need for an intelligent warehouse location optimization method and system. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides an intelligent warehouse location optimization method and system.

[0005] A first aspect of this application provides an intelligent warehouse location optimization method, comprising: Obtain data information about goods in the warehousing system, including attribute data, business data, and external environment data; Based on the data information, a storage location optimization model is constructed. The storage location optimization model includes a comprehensive objective function that aims to maximize space utilization, maximize picking efficiency, and minimize storage costs, as well as constraints based on physical rules and business rules. The storage location optimization model is solved using an optimization algorithm to generate storage location allocation schemes and corresponding prediction performance data; Execute the aforementioned storage location allocation scheme and collect actual storage location data and actual operational performance data; The actual storage location data is compared with the storage location allocation plan to obtain the plan execution compliance rate; The actual operational performance data is compared with the predicted performance data to obtain the performance deviation rate; Based on the execution compliance rate and the performance deviation rate of the scheme, the constraints of the storage location optimization model and / or the weight coefficients of the comprehensive objective function are adjusted to obtain the updated storage location optimization model. Warehouse location allocation is based on the updated location optimization model.

[0006] A second aspect of this application provides an intelligent warehouse location optimization system, comprising: The data acquisition module is used to acquire data information of goods in the warehousing system, including attribute data, business data and external environment data; A model building module is used to build a storage location optimization model based on the data information. The storage location optimization model includes a comprehensive objective function aimed at maximizing space utilization, maximizing picking efficiency, and minimizing storage costs, as well as constraints based on physical rules and business rules. The scheme generation module is used to solve the storage location optimization model using optimization algorithms to generate storage location allocation schemes and corresponding prediction performance data. The execution plan module is used to execute the storage location allocation plan and collect actual storage location data and actual operational performance data. The scheme analysis module is used to compare the actual storage location data with the storage location allocation scheme to obtain the scheme execution compliance rate; The performance comparison module is used to compare the actual operational performance data with the predicted performance data to obtain the performance deviation rate; The model adjustment module is used to adjust the constraints of the storage location optimization model and / or the weight coefficients of the comprehensive objective function based on the execution compliance rate and the performance deviation rate of the scheme, so as to obtain an updated storage location optimization model. The warehouse location allocation module is used to allocate warehouse locations based on the updated warehouse location optimization model.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described intelligent warehouse location optimization method.

[0008] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described intelligent warehouse location optimization method.

[0009] The beneficial effects of the intelligent warehouse location optimization method and system provided in this application are as follows: This application can comprehensively understand the warehousing situation by acquiring multi-faceted data information of goods, construct a warehouse location optimization model including a comprehensive objective function and constraints, comprehensively consider space utilization, picking efficiency and warehousing costs, use optimization algorithms to solve and generate warehouse location allocation schemes and predict performance data to initially plan warehouse locations, execute the scheme and collect actual data, and analyze the scheme execution and performance by comparing the scheme execution compliance rate and performance deviation rate. Based on these ratios, the warehouse location optimization model is adjusted to obtain an updated warehouse location optimization model that is more in line with reality. Finally, warehouse location allocation based on the updated warehouse location optimization model can improve the rationality and effectiveness of warehouse location allocation and achieve warehouse management optimization. Attached Figure Description

[0010] Figure 1 A flowchart illustrating an intelligent warehouse location optimization method provided in an embodiment of this application; Figure 2 A structural block diagram of an intelligent warehouse location optimization system provided in an embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the purpose, technical solution, and advantages of this application clearer, the following will be described in conjunction with the appendix. Figure 1-3 The following is an explanation using specific examples.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the intelligent warehouse location optimization method provided in this application. The method includes: S101: Obtain product data information from the warehousing system. The data information includes attribute data, business data, and external environment data.

[0014] In this embodiment, attribute data refers to the inherent physical characteristics and inherent attributes of the product itself. This type of data is stable and does not change frequently as the warehousing business process progresses, serving as the basis for judging physical suitability in warehouse location allocation. Business data is generated throughout the entire warehousing operation process. It represents the circulation status and business relationships of the product. This type of data is updated as business activities unfold, serving as the basis for efficiency and cost optimization in warehouse location allocation. External environment data refers to the data related to the storage needs of the product and warehousing operations from outside the warehousing system. This type of data exists outside the warehousing scenario and serves as a supplementary basis for adaptability and foresight in warehouse location allocation.

[0015] S102: Construct a storage location optimization model based on data information. The storage location optimization model includes a comprehensive objective function aimed at maximizing space utilization, maximizing picking efficiency, and minimizing storage costs, as well as constraints based on physical rules and business rules.

[0016] In this embodiment, the warehouse location optimization model is a mathematical model used to calculate which warehouse location a product should be allocated to; it is a decision-making tool with multiple objectives and constraints. The warehouse location optimization model consists of a comprehensive objective function and constraints, and after being solved by an optimization algorithm, it outputs the optimal warehouse location allocation scheme. Specifically, the objectives of the warehouse location optimization model are to maximize space utilization, maximize picking efficiency, and minimize warehousing costs; the constraints of the warehouse location optimization model are physical rules and business rules; the inputs of the warehouse location optimization model are product attribute data, business data, and external environment data; the outputs of the warehouse location optimization model are the warehouse location allocation scheme and predicted performance data; the iterative optimization mechanism of the warehouse location optimization model adjusts the constraints or weight coefficients of the warehouse location optimization model through feedback of the scheme execution compliance rate and performance deviation rate.

[0017] The overall objective function, a mathematical expression aimed at maximizing space utilization, picking efficiency, and minimizing warehousing costs, comprises three objectives that require weighting coefficients to balance their priorities. Maximizing space utilization aims to fully utilize warehouse space and reduce idle capacity. This is represented by maximizing metrics such as the total volume of stored goods / total available warehouse space and the number of warehouse spaces occupied by goods / total number of warehouse spaces. Maximizing picking efficiency aims to shorten picking time and reduce picking paths. This is represented by optimizing metrics such as the number of picking orders per unit time, average picking path length, and the average distance from frequently picked goods to the outbound exit. Minimizing warehousing costs aims to reduce storage-related costs, including warehouse space rental costs (e.g., higher costs in prime locations), handling costs (e.g., placing heavy goods near entrances / exits reduces energy consumption), and spoilage costs (e.g., placing fragile items in stable areas reduces damage). Mathematically, this is expressed as minimizing the sum of all costs. The comprehensive objective function should combine the three objectives with weights. For example: Total objective = ω1 × space utilization rate + ω2 × picking efficiency score + ω3 × (1 / warehousing cost), where ω1, ω2, and ω3 are weight coefficients, and the sum is 1. Adjustments are made according to business priorities.

[0018] Constraints are rules that limit the feasibility of solutions in a warehouse location optimization model, ensuring that the warehouse location allocation scheme can be implemented in practice. They are divided into two categories: physical rules and business rules. Physical rules are hard restrictions based on physical characteristics, such as: warehouse location load-bearing constraints: the weight of the goods must be less than or equal to the maximum load-bearing capacity of the warehouse location; for example, a 10kg item cannot be placed on a shelf with a load-bearing capacity of 5kg. Size matching constraints: the volume of the goods must be less than or equal to the volume of the warehouse location; for example, a 1m³ warehouse location cannot be used for storage. 3 Goods should not be placed within 0.5m. 3 Storage location; Environmental adaptation constraints: Fresh produce must be placed in refrigerated storage locations, and flammable and explosive materials must be placed in explosion-proof areas. Business rules are soft constraints based on operational logic, such as: Relevance constraints: Frequently shipped goods need to be allocated to adjacent storage locations, such as toothbrushes and toothpaste; Category isolation constraints: Food and chemicals need to be stored in separate areas; Turnover rate rules: Goods with high picking frequency need to be placed in prime areas near the outbound exit, such as best-selling items; First-in, first-out constraints: Goods in the same batch are allocated storage locations according to their arrival time, so that the first to arrive is the first to leave.

[0019] S103: Solve the storage location optimization model using optimization algorithms to generate storage location allocation schemes and corresponding prediction performance data.

[0020] In this embodiment, the optimization algorithm is a computational method used to solve the warehouse location optimization model. Its purpose is to find the optimal solution for the overall objective function while satisfying the constraints. This embodiment uses the ant colony optimization algorithm, which simulates the pheromone mechanism of ants seeking paths, for complex combinatorial optimization problems. In this embodiment, the essence of warehouse location allocation is a matching and combination problem of goods and warehouse locations.

[0021] Specifically, the ant colony algorithm's hierarchical structure includes a bottom layer, a middle layer, and a top layer. The bottom layer is the data support layer, containing raw data such as products, storage locations, and business rules, forming the foundation for its operation. The middle layer is the core layer of the ant colony algorithm, including problem encoding, heuristic information, pheromone management, and ant search modules, which are the core of generating storage location allocation schemes. The top layer is the evaluation and output layer, including fitness evaluation and optimal solution storage modules, responsible for filtering storage location allocation schemes and outputting results. The optimization goals of the ant colony algorithm are to maximize space utilization, maximize picking efficiency, and minimize warehousing costs. The input to the ant colony algorithm is readily available entity data, such as products, storage locations, and business rules, rather than abstract variables. The output is a directly usable decision result, such as a storage location allocation scheme, accompanied by performance metrics. Constraint embedding uses scenario-specific rules as hard constraints for the ant colony algorithm, making the output feasible, such as promotional or seasonal constraints.

[0022] In this embodiment, solving the warehouse location optimization model involves transforming the model into a mathematical problem that an algorithm can handle. The optimization algorithm iteratively calculates to find the optimal or near-optimal product-warehouse location matching scheme. The solution process must balance computational efficiency with solution quality, i.e., whether the scheme is close to the theoretical optimum. The warehouse location allocation scheme is the final decision result output by the optimization algorithm. It is a clear list of product-warehouse location correspondences, including: the specific warehouse location number assigned to each product, such as shelf 101, level 3 in area A; the usage method of the warehouse location, such as whether mixed storage is allowed, stacking height restrictions, etc.; and the handling rules for special products, such as temporary warehouse location adjustments for promotional products.

[0023] In this embodiment, the predicted performance data is operational indicator data based on the warehouse location allocation scheme, used to evaluate the effectiveness of the scheme, including: space utilization prediction: for example, the total space utilization is expected to be 85%, an increase of 10% compared to the current level; picking efficiency prediction: for example, the average picking path length is expected to be shortened from 150 meters to 90 meters, and the daily picking order volume is expected to increase by 20%; warehousing cost prediction: for example, the total daily warehousing cost is expected to decrease by 1,500 yuan, due to high-frequency goods occupying low-cost warehouse locations; constraint satisfaction rate: for example, 100% satisfaction of physical constraints such as load-bearing capacity and size, and 95% satisfaction of business constraints such as product association.

[0024] S104: Execute the storage location allocation plan and collect actual storage location data and actual operational performance data.

[0025] In this embodiment, executing the warehouse location allocation scheme is the process of implementing the warehouse location allocation scheme generated by the optimization algorithm into actual warehouse operations. This involves transferring goods from the current warehouse location to the allocated target warehouse location according to the scheme's requirements, ensuring that operations such as handling and shelving comply with the rules in the scheme, such as stacking methods and mixed storage restrictions. The execution process involves the coordination of warehouse personnel and AGVs (Automated Guided Vehicles), and it is crucial to avoid deviations from the scheme due to operational errors.

[0026] In this embodiment, actual storage location data refers to the actual storage location information of goods collected by the warehouse management system, sensors, etc., after the storage location allocation plan is implemented. This includes: the storage location number of each product is ultimately stored in a location that is inconsistent with the storage location allocated by the storage location allocation plan, for example, due to temporary adjustments caused by the storage location being occupied; the actual placement status of goods in the storage location, such as whether they are stacked, the stacking height, or whether they are mixed; and the actual occupancy status of the storage location, such as whether it is fully utilized or whether there is idle space. Actual operational performance data refers to the performance indicators generated in actual operation after the plan is implemented, used to evaluate the true effect of the plan, and corresponding to the predicted performance data. This includes: actual space utilization rate: the total volume of actual stored goods / the total available volume of storage locations; actual picking efficiency: such as the actual average picking path length, the actual number of picking orders per unit time, and the picking error rate; actual warehousing costs: such as the actual storage location rental fees, handling energy costs, and goods loss costs; and plan execution efficiency: such as the actual time to complete the goods transfer and the manpower / equipment input during the transfer process.

[0027] S105: Compare the actual storage location data with the storage location allocation plan to obtain the plan execution compliance rate.

[0028] In this embodiment, the actual storage location data refers to the storage location information where the goods are actually stored after the storage location allocation plan is executed, such as the specific storage location number and placement status. The storage location allocation plan is a theoretical goods-storage location matching list generated by the optimization algorithm. For example, a certain goods should be allocated to storage location 101 in area A, which is the target benchmark for execution. The comparison involves checking the actual storage location data against the storage location allocation plan one by one, analyzing their consistency, including: whether the goods have been allocated to the storage location specified in the storage location allocation plan; whether the actual placement status of the goods meets the plan requirements, such as whether they are mixed or stacked; and whether the actual occupancy of the storage location is consistent with the storage location allocation plan, such as whether the space is being used as planned.

[0029] In this embodiment, the scheme execution compliance rate is an indicator used to represent the effectiveness of scheme execution. It is the ratio of the number of goods (or the number of storage locations) that meet the requirements of the storage location allocation scheme to the total number of goods (or the total number of storage locations), representing the degree of matching between the actual execution and the theoretical scheme. The calculation formula is adjusted according to the scenario, for example: Goods-level compliance rate = (Number of goods in actual storage locations that match the scheme ÷ Total number of goods) × 100%; Storage location-level compliance rate = (Number of storage locations in actual use that match the scheme ÷ Total number of storage locations) × 100%.

[0030] S106: Compare the actual operational performance data with the predicted performance data to obtain the performance deviation rate.

[0031] In this embodiment, actual operational performance data refers to the performance indicators generated during actual warehousing operations after the storage location allocation plan is implemented. It records the true effects, including actual space utilization, actual picking efficiency (e.g., average picking time, order processing volume), and actual warehousing costs (e.g., labor, energy, and loss costs). Predicted performance data are the expected performance indicators calculated when the storage location allocation plan is generated. They are estimates of the theoretical effects of the plan and correspond one-to-one with the actual operational performance data, such as predicted space utilization, predicted picking efficiency, and predicted warehousing costs. Comparison involves comparing the actual operational performance data with the corresponding predicted performance data to analyze the degree and direction of their differences (e.g., whether the actual value is greater than or less than the predicted value), clarifying the deviation of each performance indicator from the prediction. The performance deviation rate is an indicator used to represent the difference between the actual and predicted operational results, indicating the accuracy of the prediction. The calculation formula is: Performance deviation rate = (actual value - predicted value) ÷ predicted value × 100%. If the result is positive, it means that the actual value is greater than the predicted value. For example, if the cost deviation rate is positive, it means that the actual cost is higher than the predicted cost. If the result is negative, it means that the actual value is less than the predicted value. For example, if the picking efficiency deviation rate is negative, it means that the actual efficiency is higher than the predicted efficiency, i.e., better.

[0032] S107: Based on the scheme execution compliance rate and performance deviation rate, adjust the constraints and / or weight coefficients of the comprehensive objective function of the storage location optimization model to obtain the updated storage location optimization model.

[0033] In this embodiment, the scheme execution compliance rate refers to the degree of matching between the actual storage location data and the storage location allocation scheme. It is an indicator of the effectiveness of the storage location allocation scheme and reflects whether the constraints of the storage location optimization model fit the actual execution scenario. For example, overly strict constraints may lead to inability to execute, while overly loose constraints may lead to a chaotic scheme. The performance deviation rate is the degree of difference between the actual operational performance data and the predicted performance data. It reflects the predictive accuracy of the comprehensive objective function of the storage location optimization model, especially whether the weight coefficients of each objective match the actual business needs. For example, unreasonable weight settings may lead to a large deviation in a certain performance indicator. Constraints are the rules in the storage location optimization model that limit the feasibility of the scheme, including physical rules (e.g., load-bearing capacity, size) and business rules (e.g., product association, storage area restrictions). If the scheme execution compliance rate is low, it indicates that the constraints are out of touch with the actual scenario. For example, it may not consider the temporary unavailability of some storage locations, requiring adjustment of the constraints. The weight coefficients of the comprehensive objective function are parameters for the three objectives of maximizing space utilization, maximizing picking efficiency, and minimizing warehousing costs. The weight coefficients determine the priority of each objective. For example, ω1=0.3 indicates that the priority of space utilization is 30%. If the performance deviation rate is large, it means that the weights do not match the actual business needs. For example, the actual focus is on efficiency, but the efficiency weight is set too low, and the weights need to be adjusted.

[0034] In this embodiment, the updated storage location optimization model is an optimization model that is more in line with the actual scenario and business needs after adjusting the constraints or weight coefficients. The output storage location allocation scheme is easy to execute and has more accurate prediction performance.

[0035] S108: Allocate storage locations based on the updated storage location optimization model.

[0036] In this embodiment, the updated warehouse location optimization model is an adjusted model with constraints that better reflect actual execution scenarios. For example, infeasible warehouse location restrictions have been corrected, and the weight coefficients of the overall objective function are more aligned with business needs. For instance, the priority of picking efficiency has been increased based on performance deviation rate. Compared to the original warehouse location optimization model, it has higher practicality and predictive accuracy. Warehouse location allocation is a decision-making process that rationally allocates goods to specific warehouse locations based on the output of the warehouse location optimization model. The allocation result must simultaneously satisfy the constraints of the updated warehouse location optimization model (e.g., adjusted warehouse location availability rules) and the optimization objective (e.g., prioritizing picking efficiency with higher weight).

[0037] As can be seen from the above, this application can comprehensively understand the warehousing situation by acquiring multi-faceted data on goods. Constructing a warehouse location optimization model that includes a comprehensive objective function and constraints can comprehensively consider space utilization, picking efficiency, and warehousing costs. Using optimization algorithms to solve for warehouse location allocation schemes and predict performance data allows for preliminary warehouse location planning. Executing the scheme and collecting actual data, and comparing the scheme execution compliance rate and performance deviation rate, allows for analysis of scheme execution and performance. Adjusting the warehouse location optimization model based on these ratios results in an updated model that is more realistic. Finally, allocating warehouse locations based on the updated model improves the rationality and effectiveness of warehouse location allocation, thereby optimizing warehouse management.

[0038] In one embodiment of this application, based on the scheme execution compliance rate and performance deviation rate, the weight coefficients of the constraints and / or the comprehensive objective function of the storage location optimization model are adjusted to obtain an updated storage location optimization model, including: If the compliance rate of the implementation plan is less than the first preset threshold, and multidimensional data regression analysis confirms that the deviation is due to inaccurate model constraints, then the constraints are corrected based on the actual storage location data that has not been implemented. If the performance deviation rate is greater than the second preset threshold, and the deviation is confirmed by the business rule engine to be due to the mismatch between the target weight and business requirements, then the weight coefficient of the corresponding target in the comprehensive objective function will be adjusted based on the deviation of each sub-performance indicator and the business priority.

[0039] In this embodiment, the execution compliance rate is the degree of matching between the actual warehouse location data and the warehouse location allocation plan, representing the accuracy of the plan's implementation. A low execution compliance rate indicates a disconnect between the warehouse location optimization model constraints and the actual scenario. The first preset threshold is a critical value for determining whether the execution compliance rate meets the standard, for example, 90%, set by the business scenario. If the execution compliance rate is less than the first preset threshold, it indicates a large deviation in plan execution, requiring investigation of the cause. Multidimensional data regression analysis is a statistical analysis method that identifies the causes of deviations by performing regression calculations on multidimensional variables such as unexecuted warehouse location data, constraint parameters, and actual operating environment data. For example, it may indicate whether unreasonable constraints prevent the plan from being executed. Unexecuted actual warehouse location data refers to records where the warehouse locations specified in the warehouse location allocation plan are not actually used, and goods are moved to other warehouse locations. For example, the plan allocates warehouse location 101 in area A, but warehouse location 202 in area B is actually used. This includes the physical attributes of the unused warehouse location and the reasons for non-execution, such as warehouse location damage or insufficient space.

[0040] In this embodiment, the performance deviation rate is the degree of difference between actual operational performance and predicted performance, representing the rationality of the objective function of the warehouse location optimization model. The second preset threshold is a critical value for judging whether the performance deviation is too large, for example, ±5%. If the performance deviation rate is greater than the second preset threshold, it indicates that the warehouse location optimization model's optimization of the objective does not match the actual needs. The business rule engine evaluation is an automated evaluation tool based on preset business rules, such as prioritizing picking efficiency during promotional periods and space utilization during peak seasons. This is used to determine whether the performance deviation stems from a mismatch between the target weights and business needs, rather than accidental factors. The deviation of sub-performance indicators is the deviation between the actual value and the predicted value of each optimization objective, such as a picking efficiency deviation of -4.17% and a cost deviation of 3.57%, used to locate objectives that need adjustment. Business priority is the ranking of the importance of each objective in business operations, such as picking efficiency being greater than warehousing cost during peak outbound periods, warehousing cost being greater than space utilization, and space utilization being greater than efficiency during inventory backlog periods. This serves as the basis for adjusting weights.

[0041] As can be seen from the above, this embodiment constructs a warehouse location optimization model by acquiring commodity data information and solves it to generate a warehouse location allocation scheme. The scheme is executed by collecting actual data and comparing it with predicted data to obtain the scheme execution compliance rate and performance deviation rate. If the scheme execution compliance rate is less than a first preset threshold and the deviation stems from inaccurate constraints in the warehouse location optimization model, the constraints are corrected based on the actual warehouse location data that was not executed, making the constraints more consistent with the actual situation. If the performance deviation rate is greater than a second preset threshold and the deviation stems from a mismatch between the target weights and business requirements, the weight coefficients of the corresponding targets in the comprehensive objective function are adjusted based on the deviations of each sub-performance indicator and business priorities, making the target weights more consistent with business requirements. Ultimately, a better updated warehouse location optimization model is obtained for warehouse location allocation, improving warehouse space utilization, picking efficiency, and reducing warehousing costs.

[0042] In one embodiment of this application, modifying constraints based on actual storage location data that has not been executed includes: Identify target storage locations that are allocated in the storage location allocation scheme but are not actually used; Based on the data from the storage location sensors and historical usage records, analyze the physical attributes, equipment status, and usage frequency of the target storage location; The constraint correction factor is calculated based on the probability of unavailability of storage locations and the efficiency index of usage. The constraint conditions are adjusted using the constraint correction factor.

[0043] In this embodiment, the unused actual storage location data refers to the storage locations assigned to goods in the storage location allocation scheme that were not used in actual execution. For example, goods may have been transferred to other storage locations. This includes the unused storage location number, corresponding goods information, the time of unuse, and the reason for unuse, such as storage location damage or insufficient space. The target storage location is the storage location planned for use in the storage location allocation scheme but not actually used; that is, the unexecuted storage location. It is the object of constraint deviation analysis.

[0044] In this embodiment, the storage location sensor data is data collected by sensors installed in the storage location, such as load-bearing sensors, temperature and humidity sensors, and infrared sensors. This data represents the physical state of the storage location, including actual load capacity, current temperature and humidity, whether it is occupied, and whether equipment is malfunctioning. Historical usage records are data on the past usage of the storage location, including the number of times it was allocated, the number of times it was actually used, the frequency of times it was not used for reasons such as insufficient load capacity or insufficient space, and the average usage efficiency, such as the space utilization rate per use. Physical attributes are the inherent physical characteristics of the storage location itself, such as size, maximum load capacity, material (e.g., whether it is earthquake-resistant), location (e.g., cold storage / normal temperature zone), and distance to the exit. Equipment status is the operating status of the equipment related to the storage location, such as whether the shelves are damaged, whether the sensors are malfunctioning, and whether the corresponding AGV (Automated Guided Vehicle) passage is clear, affecting the actual availability of the storage location. Usage frequency is the ratio of the number of times the storage location was actually used to the number of times it was allocated within a historical period, representing the actual availability of the storage location. For example, if a storage location is allocated 10 times but only used 3 times, the usage frequency is 30%.

[0045] In this embodiment, the unavailability probability of a storage location is calculated based on historical data, representing the likelihood that the location cannot be used. The calculation formula is: Unavailability Probability = Historical Unused Times ÷ Historical Allocation Times. A higher value indicates that the storage location is more likely to cause the plan to fail due to its own problems. Usage efficiency indicators represent the actual usage effect of the storage location, such as space utilization rate (actual storage volume ÷ storage location volume) and turnover efficiency, such as the number of times it is used per unit time, indicating whether the storage location can function as expected by the plan.

[0046] In this embodiment, the constraint correction factor is a parameter calculated based on the probability of storage location unavailability and the utilization efficiency index. It is used to adjust the strictness of the constraints. For example, a factor greater than 1 indicates that the constraints need to be relaxed; a factor less than 1 indicates that the constraints need to be tightened, making the constraints more closely match the actual state of the storage location. Constraint adjustment involves modifying the constraint rules in the original storage location optimization model according to the constraint correction factor. For example, adjusting the upper limit of load-bearing capacity or relaxing space restrictions, so that the modified constraints can adapt to the actual situation of the target storage location.

[0047] As can be seen from the above, this embodiment can identify unused target storage locations in the storage location allocation scheme, analyze their physical attributes, equipment status and usage frequency, calculate constraint correction factors, adjust the constraints of the storage location optimization model, make the constraints more in line with the actual situation, improve the execution compliance rate of the storage location allocation scheme and the accuracy of the storage location optimization model.

[0048] In one embodiment of this application, adjusting the weight coefficients of the corresponding objectives in the comprehensive objective function based on the deviations of each sub-performance index includes: Calculate the deviation rate of each performance indicator and allocate weight adjustment amount according to the deviation rate ratio and business priority factor; the deviation rate includes the deviation rate of picking efficiency indicator, the deviation rate of space utilization indicator and the deviation rate of warehousing cost indicator. If the deviation rate of the picking efficiency index is the largest, then increase the weight of maximizing picking efficiency based on the first step length, while decreasing other weights proportionally. If the deviation rate of the space utilization rate index is the largest, then the weight of maximizing space utilization is increased based on the second step size, while other weights are reduced proportionally. If the deviation rate of the warehousing cost indicator is the largest, then the weight of minimizing the warehousing cost is increased based on the third step length, while other weights are reduced proportionally.

[0049] In this embodiment, the deviation rate of each performance indicator is the difference between the actual value and the predicted value of each sub-objective in the comprehensive objective function. Examples include the deviation rate of picking efficiency, the deviation rate of space utilization, and the deviation rate of warehousing costs. This represents the gap between the optimization effect of a single objective and the expected result. The deviation rate ratio is the proportion of each sub-item's deviation rate to the total deviation rate. It is used to determine which objective's deviation has the greatest impact on overall performance. For example, if the absolute value of the picking efficiency deviation rate is 60%, it is the main deviation item. The business priority factor is a coefficient of importance for each objective set according to actual business needs. For example, during a promotional period, the priority factor for picking efficiency is 0.6, cost is 0.3, and space is 0.1, determining the direction of weight adjustment. The weight adjustment amount is the total amount used to correct the weight coefficients of the comprehensive objective function. For example, if the total adjustment amount is 0.1, the weight of a certain objective increases by 0.1, while the weights of other objectives decrease by a total of 0.1, so that the total weight remains 1. The first, second, and third step lengths correspond to the unit adjustment increments of the weights for picking efficiency, space utilization, and warehousing cost, respectively. For example, a first step length of 0.05 means the picking efficiency weight increases by 0.05. The step size is set based on the business's sensitivity to the target; for example, if efficiency has a greater impact on the business, the step size should be larger. Proportional reduction of other weights means that when the weight of a target is increased, the reduction is distributed proportionally to the original weights of other targets. For example, if the original weights are: picking efficiency 0.5, space utilization 0.3, and warehousing cost 0.2, and picking efficiency increases by 0.1, then space utilization decreases by 0.1 × 3 / 5 = 0.06, and warehousing cost decreases by 0.1 × 2 / 5 = 0.04, keeping the total weight unchanged.

[0050] The step size is calculated as follows: Step Size = Basic Step Size × Deviation Coefficient × Business Sensitivity Coefficient. The basic step size (1) is the initial adjustment range set by the business, for example, 0.05, representing the minimum adjustment unit in a normal scenario. The deviation coefficient is the normalized value of the deviation rate of the picking efficiency indicator, calculated as: Deviation Coefficient = min(|Picking Efficiency Deviation Rate| / Maximum Allowable Deviation Rate, 1.5). The maximum allowable deviation rate is the absolute value of the second preset threshold, for example, 5%. If the picking efficiency deviation rate is greater than 1.5 times the maximum allowable deviation rate, the coefficient is set to 1.5 to avoid an excessively large step size. The first business sensitivity coefficient is set according to the priority of picking efficiency based on the business scenario, for example, 1.5 for promotional periods, 1.0 for regular periods, and 0.8 for off-peak periods.

[0051] Specifically, for example: during the promotion period, the first business sensitivity coefficient = 1.5, the picking efficiency deviation rate = -6% (absolute value 6%), the maximum allowable deviation rate = 5%; the deviation coefficient = min(6% / 5%, 1.5) = 1.2; the first step length = 0.05 × 1.2 × 1.5 = 0.09, that is, the weight of maximizing picking efficiency in a single step is 0.09.

[0052] The second step size = second basic step size × deviation coefficient × second business sensitivity coefficient. Where, the basic step size 2 is the initial adjustment range for space utilization, smaller than the first step size, for example, 0.04, because space optimization takes longer to show results; the deviation coefficient is the normalized value of the deviation rate of the space utilization index, calculated as: deviation coefficient = min(|space utilization deviation rate| / maximum allowable deviation rate, 1.5); the second business sensitivity coefficient is the priority of the business towards space utilization, for example, 1.5 for inventory backlog period, 0.9 for regular period, and 0.6 for empty period.

[0053] Specifically, for example: during the inventory backlog period, the second business sensitivity coefficient = 1.5, the space utilization deviation rate = -7% (absolute value 7%), the maximum allowable deviation rate = 5%; the deviation coefficient = min(7% / 5%, 1.5) = 1.4; the second step size = 0.04 × 1.4 × 1.5 = 0.084, that is, the weight of maximizing space utilization is increased by 0.084 in a single step.

[0054] The third step size = third basic step size × deviation coefficient × third business sensitivity coefficient. Where, the basic step size of 3 is the initial magnitude of the warehouse cost adjustment (minimum, e.g., 0.03, as warehouse cost fluctuations have a slow impact on short-term operations); the deviation coefficient is the normalized value of the deviation rate of the warehouse cost indicator, calculated as: deviation coefficient = min(|warehousing cost deviation rate| / maximum allowable deviation rate, 1.5); the third business sensitivity coefficient is the business priority for warehouse costs, for example, 1.5 during the warehouse cost control period, 0.8 during the normal period, and 0.5 during the expansion period.

[0055] Specifically, for example: during the warehousing cost control period, the third business sensitivity coefficient = 1.5, the warehousing cost deviation rate = 8% (absolute value 8%), the maximum allowable deviation rate = 5%; the deviation coefficient = min(8% / 5%, 1.5) = 1.5; the third step size = 0.03 × 1.5 × 1.5 = 0.0675, that is, the minimum warehousing cost weight is increased by 0.0675 each time.

[0056] As can be seen from the above, this embodiment acquires commodity data information from the warehousing system, constructs a warehouse location optimization model including a comprehensive objective function and constraints, generates a warehouse location allocation scheme and predicted performance data, executes the scheme and collects actual data, compares the scheme execution compliance rate and performance deviation rate, and allocates weight adjustment amounts according to the deviation rate of each sub-performance indicator and business priority factor. If the picking efficiency indicator has the largest deviation rate, its weight is increased and other weights are decreased; if the space utilization indicator has the largest deviation rate, its weight is increased and other weights are decreased; if the warehousing cost indicator has the largest deviation rate, its weight is increased and other weights are decreased. This makes the target weight of the warehouse location optimization model more in line with business needs, improves warehouse picking efficiency and space utilization, and reduces warehousing costs.

[0057] In an embodiment of the present application, it further includes: When a promotion event is recognized, based on the promotion product list and the expected sales volume, temporarily increase the weights of maximizing picking efficiency and minimizing warehousing costs; When a season conversion event is recognized, based on the demand forecast of seasonally sensitive products, temporarily increase the weight of maximizing picking efficiency and adjust the storage location constraints of seasonally sensitive products.

[0058] In this embodiment, the promotion event is a product promotion activity related to warehousing, such as 618, Double 11, store anniversary, etc. Along with a sharp increase in the sales volume of specific products and a huge increase in the order volume, there are special requirements for the warehousing picking speed and warehousing cost control. The promotion product list is a list of products participating in the promotion activity, including product codes, product names, promotion intensity, etc. It is the object for temporarily adjusting the warehouse location optimization model. For example, popular products need to ensure picking efficiency first. The expected sales volume is the predicted sales volume of products during the promotion period based on historical promotion data, current inventory, and market popularity, and is used to evaluate the outbound frequency of products. The higher the sales volume, the more frequent the outbound, and the more necessary to optimize the picking path.

[0059] In this embodiment, the season conversion event is the alternation of seasons, such as the transition between spring and summer, autumn and winter, which causes a drastic change in the demand for some products. For example, air conditioners in summer and down jackets in winter, and it is necessary to adjust the warehousing strategy to adapt to the seasonal demand fluctuations. Seasonally sensitive products are products significantly affected by seasons, such as sunscreen and heaters. The demand for these products surges in specific seasons and is sluggish in other seasons, and their storage locations need to be optimized. The demand forecast is the predicted sales volume and outbound frequency of seasonally sensitive products in the target season based on historical sales data, weather forecasts, market trends, etc., and provides a basis for adjusting the storage strategy. Temporarily increasing the weight is to briefly adjust the weight coefficient of the comprehensive objective function during the event period (such as the promotion period, season conversion period), and restore the original weight after the event ends to adapt to short-term business needs. For example, during the promotion period, priority is given to ensuring the shipping speed. Adjusting the storage location constraint is to modify the original warehouse location allocation rule for seasonally sensitive products, such as shortening the distance to the outbound port and expanding the storage area, so that the products can be quickly shipped out during the peak demand period.

[0060] From the above, it can be concluded that in this embodiment, by obtaining the external promotion plan when recognizing the promotion event and calculating the promotion product heat index and the predicted value of inventory fluctuation, when the heat index is greater than the heat threshold, increasing the weight of maximizing picking efficiency can enable the warehouse to complete product picking more quickly during the promotion period and improve the picking efficiency; when the predicted value of inventory fluctuation is greater than the fluctuation threshold, increasing the weight of minimizing warehousing costs can enable reasonable control of warehousing costs during the promotion period, adapt to the business changes brought by the promotion activity, and achieve more efficient utilization of warehousing resources.

[0061] In an embodiment of the present application, when a promotion event is recognized, based on the promotion product list and the expected sales volume, the weights of maximizing picking efficiency and minimizing warehousing costs are temporarily increased, including: Obtain an external promotion plan, and calculate the popularity index and inventory fluctuation prediction value of the promoted products; If the popularity index is greater than the popularity threshold, increase the weight of maximizing picking efficiency based on the fourth step size; If the inventory fluctuation prediction value is greater than the fluctuation threshold, increase the weight of minimizing warehousing costs based on the fifth step size.

[0062] In this embodiment, the external promotion plan is the promotion activity arrangement from outside the enterprise, such as e-commerce platforms, brand owners, including the promotion time, such as June 1st - 20th, the scope of participating products, and the promotion form, such as discounts, full reduction, buy-one-get-one-free, which is the source basis for triggering warehousing adjustment. The popularity index is an indicator of the degree of market attention of the promoted products, comprehensively considering the historical promotion sales volume of the products, the current pre-sale data, the user search volume, the discount intensity, etc. The higher the value, the higher the outbound frequency of the products during the promotion period. For example, the popularity index of best-selling products is greater than that of ordinary products. The inventory fluctuation prediction value is the amplitude of inventory change calculated based on the expected sales volume and the current inventory of the promoted products, indicating the inbound and outbound frequency of the products during the promotion period. For example, a certain product has a daily sales volume of 100 pieces on weekdays, and the expected daily sales volume during the promotion period is 500 pieces, and the inventory fluctuation prediction value is 400 pieces / day.

[0063] In this embodiment, the popularity threshold is the critical value for judging whether the product popularity requires priority to ensure picking efficiency. For example, if the popularity index is greater than or equal to 80 points, it indicates that the product is a highly concerned best-selling product, which is statistically obtained from historical promotion data. The fluctuation threshold is the critical value for judging whether inventory fluctuations require priority to control warehousing costs. For example, if the inventory fluctuation prediction value is greater than or equal to 300 pieces / day, it indicates that the inventory turnover is frequent and the replenishment / handling costs need to be reduced. The fourth step size is the adjustment amplitude for temporarily increasing the weight of maximizing picking efficiency. For example, 0.08, which is positively correlated with the product popularity. The higher the popularity, the larger the step size. The fifth step size is the adjustment amplitude for temporarily increasing the weight of minimizing warehousing costs. For example, 0.05, which is positively correlated with the inventory fluctuation amplitude. The larger the fluctuation, the larger the step size.

[0064] The fourth step size is calculated as follows: Fourth Base Step Size × Popularity Influence Coefficient × Promotion Urgency Coefficient. The fourth base step size is the baseline range for adjusting picking efficiency during promotional scenarios, for example, 0.05, determined by historical promotional data. The base step size is smaller during regular promotional periods and increased during major promotional periods. The popularity influence coefficient represents the amplification effect of the promotional product's popularity on the step size. The calculation formula is: Popularity Influence Coefficient = 1 + (Average Popularity Index of Promotional Product - Popularity Threshold) ÷ Popularity Threshold. If the average popularity index is less than or equal to the popularity threshold: Popularity Influence Coefficient = 1, no additional amplification is needed; if the average popularity index is greater than the popularity threshold: Popularity Influence Coefficient increases linearly with the magnitude of the increase in the average popularity index. For example, if the average popularity index is 50% greater than the popularity threshold, Popularity Influence Coefficient = 1.5, with an upper limit of 2 to avoid excessively large step sizes. The urgency factor for a promotion is set based on the length of the promotion period. The shorter the period, such as a flash sale, the higher the urgency factor and the larger the factor: Short-term promotion (less than or equal to 3 days): 1.5; Medium-term promotion (4-10 days): 1.2; Long-term promotion (more than 10 days): 1.0.

[0065] Specifically, for example: a certain e-commerce warehouse's 618 promotion (mid-term promotion, urgency coefficient 1.2): fourth basic step length = 0.05; average popularity index of promotional products = 95 points, popularity threshold = 70 points; popularity influence coefficient = 1 + (95-70) ÷ 70 ≈ 1 + 0.357 = 1.357 (less than or equal to 2, valid); fourth step length = 0.05 × 1.357 × 1.2 ≈ 0.081, that is, the picking efficiency weight increases by 0.081 per time.

[0066] The fifth step length = fifth base step length × volatility impact coefficient × turnover rate coefficient. Here, the fifth base step length is the benchmark range for adjusting warehousing costs in promotional scenarios, smaller than the fourth step length (e.g., 0.03), because warehousing cost optimization has a lower priority than picking efficiency, but should be strengthened in high-volatility scenarios. The volatility impact coefficient represents the amplification effect of inventory fluctuations on the step length. The calculation formula is: Volatility Impact Coefficient = 1 + (Average Inventory Volatility Forecast for Promotional Items - Volatility Threshold) ÷ Volatility Threshold. If the average volatility forecast is less than or equal to the volatility threshold: Volatility Impact Coefficient = 1; if the average volatility forecast is greater than the volatility threshold: Volatility Impact Coefficient increases linearly with the magnitude of the volatility increase. For example, if the volatility is 80% greater than the volatility threshold, the Volatility Impact Coefficient = 1.8, with an upper limit of 1.5. Adjustments to warehousing cost weights require greater caution. The turnover rate coefficient is set based on the inventory turnover rate (sales / inventory) of promotional products. The higher the turnover rate (the faster the inventory turnover), the larger the turnover rate coefficient: High turnover rate (greater than or equal to 5 times / promotion period): 1.3; Medium turnover rate (3-5 times / promotion period): 1.1; Low turnover rate (less than 3 times / promotion period): 1.0.

[0067] Specifically, for example: in the same 618 promotion scenario: the fifth basic step length = 0.03; the predicted average inventory fluctuation of promotional items = 300 units / day, the fluctuation threshold = 200 units / day; the fluctuation impact coefficient = 1 + (300-200) ÷ 200 = 1.5 (less than or equal to 1.5, valid); the average inventory turnover rate of promotional items = 4 times / promotion period (medium turnover rate, coefficient 1.1); the fifth step length = 0.03 × 1.5 × 1.1 ≈ 0.0495, that is, the warehousing cost weight increases by 0.0495 per time.

[0068] As can be seen from the above, this embodiment obtains external promotional plans and calculates the popularity index and inventory fluctuation forecast of promotional products when identifying promotional events. When the popularity index is greater than the popularity threshold, the weight of maximizing picking efficiency is increased, which enables the warehouse to complete product picking more quickly during the promotion period and improve picking efficiency. When the inventory fluctuation forecast is greater than the fluctuation threshold, the weight of minimizing warehousing costs is increased, which can reasonably control warehousing costs during the promotion period, adapt to business changes brought about by promotional activities, and achieve more efficient utilization of warehousing resources.

[0069] In one embodiment of this application, when a seasonal transition event is identified, based on demand forecasts for seasonally sensitive goods, the weight for maximizing picking efficiency is temporarily increased, and the storage location constraints for seasonally sensitive goods are adjusted, including: Based on weather forecasts and historical sales patterns, identify target seasons and seasonally sensitive products; Calculate the seasonal demand intensity index for seasonally sensitive goods; If the seasonal demand intensity index is greater than the preset seasonal threshold, the weight of maximizing picking efficiency will be increased based on the sixth step length, and the maximum storage distance constraint for seasonally sensitive goods will be tightened to within the range of the first distance threshold.

[0070] In this embodiment, seasonal transition events refer to the periods of seasonal change throughout the year, such as the transition from spring to summer in March-April and from autumn to winter in September-October. During these times, the demand structure for goods changes; for example, summer sandals replace winter shoes, and winter cotton shoes become popular. Warehouse strategies need to be adjusted to accommodate these demand fluctuations. Weather forecasts, data from meteorological departments on temperature, precipitation, and other conditions for a future period (e.g., 1-3 months), are used to determine the timing of seasonal transitions. For example, an average temperature below 15°C for 10 consecutive days indicates the arrival of autumn. Historical sales patterns are the historical data patterns of goods sales in different seasons. For example, down jacket sales account for 40% of the annual total in November, used to identify goods significantly affected by seasonality. The target season is the upcoming season; for example, based on the September weather forecast, the target season is autumn, serving as a reference for warehouse adjustments. Seasonally sensitive goods are those whose demand fluctuates dramatically with the seasons, such as air conditioners in summer and hand warmers in winter. Sales of these goods increase during the target season and decrease in other seasons. Demand forecasting is based on historical sales patterns, weather forecasts, market trends, etc., to predict the sales volume and shipment frequency of seasonally sensitive goods in the target season. For example, it predicts that the monthly sales volume of light autumn jackets will increase by 500%.

[0071] In this embodiment, the seasonal demand intensity index is an indicator of the strength of demand for seasonally sensitive goods in the target season. It takes into account factors such as demand growth rate, outbound frequency, and inventory turnover rate. The higher the value, the stronger the urgency of outbound delivery of the goods in the target season. The preset seasonal threshold is a critical value for determining whether the warehouse location optimization model needs to be adjusted. For example, if the intensity index is greater than or equal to 60 points, when the seasonal demand intensity index of a goods is greater than the preset seasonal threshold, the picking efficiency weight is increased and the storage constraints are adjusted.

[0072] In this embodiment, the sixth step size is an adjustment range used to temporarily increase the weight for maximizing picking efficiency, for example, 0.06, which is positively correlated with the intensity of seasonal demand; the higher the intensity of seasonal demand, the larger the step size. The maximum storage distance constraint is the maximum allowable distance from the product storage location to the outbound gate specified in the original warehouse location optimization model, for example, less than or equal to 30 meters for regular products, so that picking efficiency is within a reasonable range. The first distance threshold is the maximum storage distance tightened for seasonally sensitive products during the seasonal transition period, for example, less than or equal to 15 meters, which is stricter than the regular constraint to shorten the picking path.

[0073] Specifically, the sixth step size = sixth base step size × demand intensity coefficient × seasonal transition urgency coefficient. The sixth base step size is the baseline adjustment range for seasonal transition scenarios, for example, 0.05, determined by historical seasonal transition data. For instance, the base step size is larger in autumn than in spring because demand fluctuations are more pronounced in autumn. The demand intensity coefficient represents the amplification effect of the step size on the strength of demand for seasonally sensitive goods. The calculation formula is: Demand Intensity Coefficient = Average Intensity Index of Seasonally Sensitive Goods ÷ Preset Seasonal Threshold. If the average intensity index is less than or equal to the preset seasonal threshold: Demand Intensity Coefficient = 1, no additional amplification is needed, the base step size is sufficient; if the average intensity index is greater than the preset seasonal threshold: the demand intensity coefficient increases linearly with the magnitude of the increase. For example, if the average intensity index is 1.5 times the preset seasonal threshold, the demand intensity coefficient = 1.5, with an upper limit of 2 to avoid excessively large step sizes. The seasonal transition urgency coefficient is set based on the speed of seasonal transition (e.g., the rate of temperature change). The more rapid the transition, the larger the coefficient: rapid transition (e.g., an average temperature drop of 10°C or more within 10 days): 1.4; normal transition (e.g., an average temperature drop of 5-10°C within 10-20 days): 1.2; slow transition (e.g., an average temperature drop of less than 5°C within more than 20 days): 1.0.

[0074] Specifically, for example: Calculating the demand intensity in a summer-to-autumn transition scenario, given the following conditions: base step size 6 = 0.05 (baseline value for the autumn transition period); average intensity index of seasonally sensitive goods (light jackets, long-sleeved T-shirts) = 87.5 points; preset seasonal threshold = 60 points; weather forecast shows temperature dropping from 28℃ to 18℃ within 10 days (rapid transition, urgency coefficient = 1.4). Demand intensity coefficient = 87.5 ÷ 60 ≈ 1.458 (less than or equal to 2, valid), sixth step size = 0.05 × 1.458 × 1.4 ≈ 0.05 × 2.041 ≈ 0.102. Constraint adjustment: Because the calculated result 0.102 is greater than the upper limit 0.1, the final sixth step size is set to 0.1 to prevent excessive weight adjustment.

[0075] As can be seen from the above, this embodiment can accurately determine the products affected by the season by identifying the target season and seasonally sensitive products based on weather forecasts and historical sales patterns; it can calculate the seasonal demand intensity index of seasonally sensitive products to obtain the demand level of the products in the target season; when the seasonal demand intensity index is greater than the preset seasonal threshold, the weight of maximizing picking efficiency is increased based on the sixth step, which can improve the picking efficiency of seasonally sensitive products; tightening the maximum storage distance constraint of seasonally sensitive products to the range of the first distance threshold can optimize the storage location of seasonally sensitive products, further improve picking efficiency, and better adapt to the demand of seasonal transition events for warehouse space.

[0076] In one embodiment of this application, it further includes: Before using the optimization algorithm to solve the storage location optimization model, the hyperparameters of the optimization model are determined based on the historical convergence speed and the quality of the solution to the storage location optimization model; the optimization algorithm is the ant colony algorithm.

[0077] In this embodiment, the optimization algorithm employs the ant colony optimization (ACO) algorithm. The ACO algorithm is a heuristic optimization algorithm that simulates the foraging behavior of ant colonies. It uses pheromones to transmit path information and searches for the optimal solution in the solution space, such as a warehouse location allocation scheme. In warehouse location optimization, it is used to solve the combined optimization problem of matching goods and warehouse locations, and converges quickly to a relatively optimal solution through group cooperation. Historical convergence speed refers to the number of iterations or time required for the algorithm to iterate from the initial solution to a stable optimal solution (convergence) when solving similar warehouse location optimization models using the ACO algorithm in the past. For example, a historical average of 50 iterations for convergence indicates the algorithm's solution efficiency. Solution quality refers to the degree to which the solution output by the optimization algorithm (warehouse location allocation scheme) optimizes the comprehensive objective function. For example, whether the objective function value is close to the theoretical optimum and whether it satisfies all constraints is measured by the degree of matching between the optimality of the solution and the actual requirements. Hyperparameters are parameters pre-set before the ACO algorithm runs (not parameters automatically learned during algorithm iteration), and directly affect the convergence speed and solution quality of the ACO algorithm. The hyperparameters of the optimization model are determined based on historical convergence speed and solution quality. The hyperparameters of the ant colony algorithm are adjusted so that the optimization algorithm can converge quickly and output high-quality solutions in the current reservoir optimization model.

[0078] As can be seen from the above, this embodiment determines the hyperparameters of the optimization model based on historical convergence speed and solution quality before using the ant colony algorithm to solve the warehouse location optimization model. This allows the ant colony algorithm to better adapt to the warehouse location optimization model, improves solution efficiency and solution quality, and thus obtains a better warehouse location allocation scheme, thereby improving the rationality of warehouse location allocation.

[0079] In one embodiment of this application, the hyperparameters include: pheromone evaporation factor and heuristic factor; If the improvement rate of the fitness function value is less than the improvement threshold for N consecutive generations, then the pheromone evaporation factor is increased based on the first adjustment amount. If the quality variance of the solution is greater than the variance threshold, then a heuristic factor is added based on the second adjustment.

[0080] In this embodiment, the pheromone evaporation factor is a parameter in the ant colony algorithm that controls the decay rate of old pheromones. Its value ranges from 0 to 1. A larger value indicates faster pheromone evaporation, and the algorithm is more inclined to explore new solutions; a smaller value indicates more old pheromones accumulate, and the ant colony algorithm tends to utilize existing solutions. The heuristic factor is a parameter in the ant colony algorithm that represents the influence of heuristic information (such as distance to storage locations or product association) on the ants' path selection. A larger value indicates that the ants rely more on currently known high-quality solutions, such as short paths; a smaller value indicates that they rely more on random pheromone exploration.

[0081] In this embodiment, "N consecutive generations" refers to the number of consecutive generations in the ant colony algorithm iteration process that update the solution, for example, 5 consecutive generations. A generation is a unit of one complete iteration in the ant colony algorithm, and each generation includes the process of all ants completing path selection and pheromone updates. The fitness function value is a score of the quality of the solution (storage allocation scheme) generated by the ant colony algorithm, directly related to the comprehensive objective function of the storage optimization model. For example, a higher objective function value results in a higher fitness function value, indicating a higher degree of solution optimization. The improvement rate is the percentage increase in the fitness function value between two consecutive generations, calculated as: Improvement Rate = (Current Generation Fitness - Previous Generation Fitness) / Previous Generation Fitness × 100%. A smaller value indicates a slower optimization speed. The improvement threshold is a critical value for determining whether fitness improvement has stagnated, for example, 1%. If the improvement over N consecutive generations is less than the improvement threshold, it indicates that the ant colony algorithm is trapped in a local optimum (difficult to find a better solution).

[0082] In this embodiment, the first adjustment is used to increase the magnitude of the pheromone evaporation factor, for example, by 0.05. The purpose is to accelerate the evaporation of old pheromones, prompting the algorithm to explore new solutions to break the stagnation. The quality variance of the solution is the degree of dispersion of the fitness function values ​​of solutions generated through multiple iterations. A larger variance indicates greater fluctuations in solution quality, with some solutions being high-quality and others poor, indicating insufficient algorithm stability. The variance threshold is a critical value for judging whether the quality of solutions is too dispersed, for example, 0.02. If the variance is greater than the variance threshold, it indicates that the ant colony algorithm's ability to focus on high-quality solutions is insufficient. The second adjustment is used to increase the magnitude of the heuristic factor, for example, by 0.5. The purpose is to enhance the ants' preference for high-quality solutions, reduce random exploration, and improve solution stability.

[0083] Specifically, the first adjustment amount = first base adjustment amount × stagnation coefficient × historical correction coefficient. The base adjustment amount of 1 is the baseline range for adjusting the pheromone evaporation factor, for example, 0.05, determined by historical data, ensuring that a single adjustment significantly affects the pheromone evaporation rate without being excessive. The stagnation coefficient represents the severity of stagnation in the ant colony algorithm, calculated as: Stagnation coefficient = (Improvement threshold - Average improvement rate over N generations) ÷ Improvement threshold × N ÷ Baseline number of generations. Here, the average improvement rate over N generations is the average of the improvement rates over the N generations being adjusted; the smaller the value, the more severe the stagnation. The baseline number of generations is the set standard number of generations for monitoring stagnation, for example, 5 generations. The larger N is, the longer the stagnation lasts, and the larger the stagnation coefficient. The upper limit of the stagnation coefficient is 2 to avoid excessive adjustment leading to excessively rapid pheromone evaporation. The historical correction coefficient is set based on the adjustment effects of similar stagnation scenarios in the past. For example, if the improvement rate increases by more than 50% after a historical adjustment, the historical correction coefficient = 1.0; if it only increases by 20%, the historical correction coefficient = 1.2, to increase the adjustment amount, ranging from 0.8 to 1.5.

[0084] For example, in a certain scenario, the improvement threshold is 1%, the baseline consecutive generations is 5, the first basic adjustment amount is 0.05, and the historical correction coefficient is 1.0. When the adjustment is triggered, the average improvement rate of consecutive N=5 generations is 0.2%, which is significantly less than the improvement threshold; the stagnation coefficient is (1%-0.2%)÷1%×5÷5=0.8×1=0.8; the first adjustment amount is 0.05×0.8×1.0=0.04.

[0085] For example, if N=8 consecutive generations (greater than the baseline generation), the average improvement is 0.1%: the stagnation coefficient is (1%-0.1%) ÷ 1% × 8 ÷ 5 = 0.9 × 1.6 = 1.44; the first adjustment is 0.05 × 1.44 × 1.0 = 0.072. Since the coefficient is not greater than 2, it is valid.

[0086] Specifically, the second adjustment amount = second base adjustment amount × volatility coefficient × high-quality solution proportion coefficient. The second base adjustment amount is the benchmark magnitude for adjusting the heuristic factor, for example, 0.5, which is greater than the first adjustment amount because the heuristic factor has a more direct impact on the focus of the solution. The volatility coefficient represents the severity of the quality fluctuation of the solution, calculated as: volatility coefficient = (current variance - variance threshold) ÷ variance threshold. Wherein, if the current variance ≤ variance threshold: volatility coefficient = 0, no adjustment is needed; if the current variance is greater than variance threshold: volatility coefficient increases linearly with the magnitude of the difference, with an upper limit of 1.5, to avoid the heuristic factor becoming too large and causing the algorithm to lose its exploration ability. The high-quality solution proportion coefficient is set based on the proportion of high-quality solutions in the current solution: proportion less than or equal to 20%, few high-quality solutions: high-quality solution proportion coefficient = 1.3; proportion 20%-50%, moderate high-quality solutions: high-quality solution proportion coefficient = 1.0; proportion greater than 50%, many high-quality solutions: high-quality solution proportion coefficient = 0.8.

[0087] For example, in a certain scenario, the variance threshold is 0.02, and the second basic adjustment amount is 0.5. During adjustment, if the current variance is 0.03, which is greater than the threshold of 0.01, the proportion of high-quality solutions is 15% (less than or equal to 20%); the volatility coefficient is (0.03-0.02)÷0.02=0.5; the proportion coefficient of high-quality solutions is 1.3; and the second adjustment amount is 0.5×0.5×1.3=0.325.

[0088] For example, if the current variance = 0.05, the proportion of high-quality solutions = 30%, the volatility coefficient = (0.05-0.02) ÷ 0.02 = 1.5, reaching the upper limit; the proportion coefficient of high-quality solutions = 1.0; the second adjustment amount = 0.5 × 1.5 × 1.0 = 0.75.

[0089] As can be seen from the above, this embodiment constructs a warehouse location optimization model by acquiring product data information, uses the ant colony algorithm to solve the model to generate a warehouse location allocation scheme, executes the scheme and collects actual data, adjusts the model based on the scheme execution compliance rate and performance deviation rate, and determines hyperparameters based on historical convergence speed and solution quality before solving. If the improvement rate of the fitness function value for N consecutive generations is less than the improvement threshold, adding a pheromone evaporation factor can prevent the algorithm from getting trapped in local optima and improve global search capability. If the quality variance of the solution is greater than the variance threshold, adding a heuristic factor can accelerate the convergence speed of the algorithm and improve the solution efficiency, thereby more accurately allocating warehouse locations, improving space utilization, picking efficiency and reducing warehousing costs.

[0090] Corresponding to the intelligent warehouse location optimization method in the above embodiment, Figure 2 This is a structural block diagram of an intelligent warehouse location optimization system provided in one embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2 The intelligent warehouse location optimization system 20 includes: a data acquisition module 21, a model building module 22, a solution generation module 23, a solution execution module 24, a solution analysis module 25, a performance comparison module 26, a model adjustment module 27, and a warehouse location allocation module 28.

[0091] in, Data acquisition module 21 is used to acquire data information of goods in the warehousing system. The data information includes attribute data, business data and external environment data. Model building module 22 is used to build a storage location optimization model based on data information. The storage location optimization model includes a comprehensive objective function aimed at maximizing space utilization, maximizing picking efficiency and minimizing storage costs, as well as constraints based on physical rules and business rules. The scheme generation module 23 is used to solve the storage location optimization model using optimization algorithms to generate storage location allocation schemes and corresponding prediction performance data; The execution plan module 24 is used to execute the storage location allocation plan and collect actual storage location data and actual operational performance data; The scheme analysis module 25 is used to compare the actual storage location data with the storage location allocation scheme to obtain the scheme execution compliance rate; The performance comparison module 26 is used to compare the actual operational performance data with the predicted performance data to obtain the performance deviation rate. The model adjustment module 27 is used to adjust the constraints and / or weight coefficients of the comprehensive objective function of the storage location optimization model based on the scheme execution compliance rate and performance deviation rate, so as to obtain the updated storage location optimization model. The storage location allocation module 28 is used to allocate storage locations based on the updated storage location optimization model.

[0092] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the data acquisition module 21, model building module 22, scheme generation module 23, scheme execution module 24, scheme analysis module 25, performance comparison module 26, model adjustment module 27, and storage location allocation module 28 are shown.

[0093] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0094] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0095] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0096] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in any embodiment of the intelligent warehouse location optimization method provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0097] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0098] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0099] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0100] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0101] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or units, or it may be an electrical, mechanical, or other form of connection.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0103] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0104] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent warehousing bin location optimization method, characterized in that, The method comprises: acquiring data information of goods in a warehouse system, the data information comprising attribute data, business data and external environment data; constructing a storage location optimization model based on the data information, the storage location optimization model comprising a comprehensive objective function with the objectives of maximizing space utilization, maximizing picking efficiency and minimizing warehouse cost, and constraint conditions based on physical rules and business rules; solving the storage location optimization model using an optimization algorithm to generate a storage location allocation scheme and corresponding predicted performance data; executing the storage location allocation scheme and collecting actual storage location data and actual operation performance data; comparing the actual storage location data with the storage location allocation scheme to obtain a scheme execution compliance rate; comparing the actual operation performance data with the predicted performance data to obtain a performance deviation rate; adjusting the constraint conditions of the storage location optimization model and / or the weight coefficients of the comprehensive objective function based on the scheme execution compliance rate and the performance deviation rate to obtain an updated storage location optimization model; allocating storage locations in the warehouse based on the updated storage location optimization model. 2.The intelligent warehouse location optimization method of claim 1, wherein, The adjusting of the constraint conditions of the storage location optimization model and / or the weight coefficients of the comprehensive objective function based on the scheme execution compliance rate and the performance deviation rate to obtain an updated storage location optimization model comprises: if the scheme execution compliance rate is less than a first preset threshold and it is confirmed through multidimensional data regression analysis that the deviation is caused by inaccurate model constraints, then the constraint conditions are corrected based on actual storage location data that have not been executed; if the performance deviation rate is greater than a second preset threshold and it is confirmed through a business rule engine evaluation that the deviation is caused by a mismatch between the target weight and business demand, then the weight coefficients of the corresponding targets in the comprehensive objective function are adjusted based on the deviation of each performance indicator and the business priority. 3.The intelligent warehouse location optimization method of claim 2, wherein, The correction of the constraint conditions based on actual storage location data that have not been executed comprises: identifying target storage locations that are allocated in the storage location allocation scheme but not used in practice; analyzing the physical properties, equipment status and usage frequency of the target storage locations based on storage location sensor data and historical usage records; calculating a constraint correction factor based on the storage location unavailability probability and usage efficiency indicators; and adjusting the constraint conditions using the constraint correction factor. 4.The intelligent warehouse location optimization method of claim 2, wherein, The adjustment of the weight coefficients of the corresponding targets in the comprehensive objective function based on the deviation of each performance indicator comprises: calculating the deviation rates of each performance indicator and allocating weight adjustment amounts according to the deviation rate proportions and business priority factors; the deviation rates comprise the deviation rate of the picking efficiency indicator, the deviation rate of the space utilization indicator and the deviation rate of the warehouse cost indicator; if the deviation rate of the picking efficiency indicator is the largest, then the weight of the maximized picking efficiency is increased based on a first step size, and the weights of the other targets are reduced according to a proportion; if the deviation rate of the space utilization indicator is the largest, then the weight of the maximized space utilization is increased based on a second step size, and the weights of the other targets are reduced according to a proportion; if the deviation rate of the warehouse cost indicator is the largest, then the weight of the minimized warehouse cost is increased based on a third step size, and the weights of the other targets are reduced according to a proportion. 5.The intelligent warehouse location optimization method of claim 1, wherein, The method further comprises: temporarily increasing the weight of the maximizing picking efficiency and the minimizing warehouse cost based on a promotion item list and expected sales volume when a promotion event is identified; temporarily increasing the weight of the maximizing picking efficiency and adjusting the storage location constraint of a seasonally sensitive item based on demand prediction of the seasonally sensitive item when a season transition event is identified.

6. The intelligent warehousing location optimization method according to claim 5, wherein, The temporarily increasing the weight of the maximizing picking efficiency and the minimizing warehouse cost based on a promotion item list and expected sales volume when a promotion event is identified includes: obtaining an external promotion plan, calculating a heat index and an inventory fluctuation prediction value of a promotion item; if the heat index is greater than a heat threshold value, increasing the weight of the maximizing picking efficiency based on a fourth step length; if the inventory fluctuation prediction value is greater than a fluctuation threshold value, increasing the weight of the minimizing warehouse cost based on a fifth step length.

7. The intelligent warehousing location optimization method according to claim 5, wherein, The temporarily increasing the weight of the maximizing picking efficiency and adjusting the storage location constraint of a seasonally sensitive item based on demand prediction of the seasonally sensitive item when a season transition event is identified includes: identifying a target season and a seasonally sensitive item based on weather forecasts and historical sales patterns; calculating a seasonal demand intensity index of the seasonally sensitive item; if the seasonal demand intensity index is greater than a preset season threshold value, increasing the weight of the maximizing picking efficiency based on a sixth step length, and tightening the maximum storage distance constraint of the seasonally sensitive item to a first distance threshold range. 8.The intelligent warehouse location optimization method of claim 1, wherein, Further comprising: determining hyperparameters of the optimization model based on historical convergence speed and quality of solving the storage location optimization model before solving the storage location optimization model using an optimization algorithm; wherein the optimization algorithm is an ant colony algorithm. 9.The intelligent warehouse location optimization method of claim 8, wherein, The hyperparameters include: pheromone evaporation factor and heuristic factor; if the improvement rate of the fitness function value of the continuous N generations is less than an improvement threshold value, increasing the pheromone evaporation factor based on a first adjustment amount; if the quality variance of the solution is greater than a variance threshold value, increasing the heuristic factor based on a second adjustment amount.

10. An intelligent warehousing bin location optimization system, characterized in that, Comprising: a data acquisition module for acquiring data information of items in a warehouse system, the data information including attribute data, business data and external environment data; a model construction module for constructing a storage location optimization model based on the data information, the storage location optimization model including a comprehensive objective function with the objectives of maximizing space utilization, maximizing picking efficiency and minimizing warehouse cost, and constraint conditions based on physical rules and business rules; a scheme generation module for solving the storage location optimization model using an optimization algorithm to generate a storage location allocation scheme and corresponding predicted performance data; an execution scheme module for executing the storage location allocation scheme and collecting actual storage location data and actual operation performance data; a scheme analysis module for comparing the actual storage location data with the storage location allocation scheme to obtain a scheme execution compliance rate; a performance comparison module for comparing the actual operation performance data with the predicted performance data to obtain a performance deviation rate; The model adjustment module is configured to adjust the constraint condition of the storage location optimization model and / or the weight coefficient of the comprehensive objective function based on the scheme execution compliance rate and the performance deviation rate, to obtain an updated storage location optimization model. The storage location allocation module is configured to perform warehouse storage location allocation based on the updated storage location optimization model.

Images