A Step-by-Step Optimization Method and System for Warehouse Location and Distribution Relationships

Abstract

This application relates to the field of supply chain optimization technology, and in particular to a method and system for warehouse location selection and distribution relationship optimization based on step-by-step optimization. The method includes: dividing the problem into two models; in the first model, acquiring first modeling data to calculate production and demand calculations, the first modeling data including first decision variables, a first optimization objective, and setting first constraints; in the second model, inheriting the production and demand calculations, and acquiring second modeling data to calculate the actual warehouse location and actual distribution location, the second modeling data including second decision variables, a second optimization objective, and setting second constraints. This application employs a two-step modeling and solution scheme of production-distribution optimization and warehouse location and distribution relationship optimization, simultaneously solving the SKU production line problem and selecting warehouse locations and determining the distribution factories / warehouses for demand points, thereby reducing the solution difficulty.

Description

Technical Field

[0001] This application relates to the technical field of supply chain optimization, and in particular to a method and system for optimizing warehouse location and distribution relationships based on step-by-step optimization. Background Technology

[0002] Because large-scale warehouse location selection and distribution relationship optimization processes involve numerous demand points and candidate warehouse addresses, and the calculation process includes routes from the factory to the warehouse and then from the warehouse to the demand point, the supply network is extremely large, making modeling and solving extremely difficult. In practice, such as... Figure 1 As shown in Figure a, to reduce the complexity of the problem, it is usually necessary to pre-screen the supply routes and candidate warehouse addresses before modeling. This reduces the number of variables entering the solution process and lowers the difficulty of the solution. Pre-screening is generally based on a pre-defined warehouse coverage area, for example, referring to... Figure 1 As shown in Figure b, within 500 kilometers; or refer to... Figure 1 In Figure c, the centroid method is used to calculate the straight-line distance weighted by demand. Based on... Figure 1 Chinese Figure b and Figure 1 Figure c shows the selection of a set of candidate warehouse addresses that meet the criteria. The advantage of this method is that it is simple, fast, and easy to understand. For business operators without a technical background, both the process and the results are easy to manage.

[0003] However, this type of method also has a major drawback: it fails to consider the cost aspect. Supply network optimization and warehouse location selection are essentially optimization problems aimed at minimizing total supply costs while meeting demand. If this objective is not consistently integrated into the decision-making logic, the quality of the decision outcome cannot be guaranteed. (See reference...) Figure 2 A manufacturing company has multiple factories across the country, each with multiple production lines. While there is some overlap in the products each production line can produce, they are not entirely identical. The company has multiple demand points nationwide, and there are two supply paths to meet those demands: one is direct factory-to-demand, i.e., a factory-to-demand point supply path; the other is establishing a network of forward warehouses across the country, where products are first shipped from the factory to the forward warehouses, and then distributed from the forward warehouses to the demand points, i.e., a factory-forward warehouse-to-demand point supply path. The challenge is to select suitable warehouse locations from among multiple candidate warehouses and assign a distribution factory / warehouse to each demand point, so as to achieve the lowest possible total supply cost while maximizing demand fulfillment. Supply costs include production costs, logistics costs, warehousing costs, and inventory costs.

[0004] This is a mixed integer programming problem that requires simultaneous decisions on three issues: the production line of the SKU, the location of the forward warehouse, and the distribution factory / warehouse at the demand point. Capacity constraints, minimum shipment constraints, and maximum throughput constraints closely link the five dimensions of factory, production line, warehouse, demand point, and SKU, resulting in high problem complexity and difficulty in solving. Summary of the Invention

[0005] To ensure the quality of the solution while reducing the difficulty of solving it, this application provides a method and system for optimizing warehouse location and distribution relationships based on step-by-step optimization.

[0006] Firstly, this application provides a method for optimizing warehouse location and distribution relationships based on step-by-step optimization, employing the following technical solution:

[0007] A step-by-step optimization method for warehouse location selection and distribution relationship optimization includes the following steps:

[0008] The problem is divided into two models: the first model is represented as an optimization model of factory-demand point, and the second model is represented as an optimization model of factory-warehouse-demand point.

[0009] In the first problem model, first modeling data is obtained to calculate the production calculation volume and the demand calculation volume. The first modeling data includes a first decision variable, a first optimization objective, and a first constraint condition.

[0010] In the second problem model, the production calculation volume and demand calculation volume are inherited, and the second modeling data is obtained to calculate the actual warehouse construction point and the actual delivery location. The second modeling data includes the second decision variable, the second optimization objective, and the second constraint condition.

[0011] By adopting the above technical solution, two problem models are defined. The first model only needs to consider the production line set and the demand point set, ignoring the intermediate warehouse distribution link. Based on the production capacity of the production line set to be optimized and the factory demand fulfillment rate of the demand point set, the optimal production and demand calculations are initially found. Then, the second problem module performs inherited optimization calculations on the production and demand calculations to obtain the actual warehouse construction points and the actual distribution warehouses corresponding to the demand points. A two-step modeling and solution scheme of production-distribution optimization and warehouse location and distribution relationship optimization is adopted, thereby reducing the solution difficulty.

[0012] In some embodiments, before acquiring the first modeling data to calculate the production computation and the demand computation, the following steps are included:

[0013] Based on the factory to be optimized, obtain the corresponding set of production line locations and the set of demand points, and generate the unit supply cost based on the set of production line locations and the set of demand points. The unit supply cost represents the lowest cost among all paths between the production line location and the demand point location.

[0014] By adopting the above technical solution, the corresponding production line location set and demand point location set are obtained based on the factory to be optimized, and the unit supply cost is generated based on the production line location set and demand point location set. The unit supply cost from each factory to each demand point is obtained. Under the premise of meeting the demand, the supply cost is minimized. Through a two-step modeling and solution scheme, the production line problem is solved first, and then the minimization of transportation cost and warehousing cost is considered to find the supply path with the lowest cost. At the same time, the warehouse location is selected and the distribution factory / warehouse of the demand point is determined, thereby reducing the solution difficulty.

[0015] In some embodiments, the first optimization objective includes a first total supply cost, and the first total supply cost is obtained by the following steps:

[0016] Based on the set of production lines, obtain the corresponding production capacity and production cost of the production lines, and based on the set of demand points, obtain the corresponding factory demand fulfillment quantity.

[0017] The first total supply cost is determined based on the production line capacity, the factory demand fulfillment, the production line cost, and the unit supply cost.

[0018] By adopting the above technical solution, in the first problem model, the production and distribution optimization objective is determined based on the principle of minimizing the first total supply cost. This ensures that the production capacity of each factory's production line and the demand fulfillment of each factory are based on the lowest cost path among the two possible paths. This solves the problem of the factory's production line. Without considering intermediate links, it focuses only on the production end and the demand end to find the best matching relationship between the factory and the demand point. This facilitates the inheritance and optimization of the subsequent preset location and distribution optimization model, and reduces the solution complexity of simultaneously deciding on the factory's production line, the location of the forward warehouse, and the distribution factory / warehouse of the demand point.

[0019] In some embodiments, the second constraint includes a delivery flow conservation constraint, which is generated by the following steps:

[0020] The distribution flow conservation constraint is determined based on the production calculation, demand calculation, and the second decision variable.

[0021] In some embodiments, the second optimization objective includes a second total supply cost, and the second total supply cost is obtained by the following steps:

[0022] Obtain the corresponding set of unit costs based on the second decision variable, and determine the second total supply cost based on the second decision variable and the corresponding set of unit costs.

[0023] In some embodiments, after obtaining the actual delivery locations corresponding to the actual warehouse construction points and demand points, the following steps are also included:

[0024] Two cyclical problem models are set up. The first cyclical problem model represents the cyclical optimization model of factory-demand point, and the second cyclical problem model represents the cyclical optimization model of factory-warehouse-demand point.

[0025] In the first cycle problem model, the first cycle modeling data is obtained to calculate the cycle production and cycle demand.

[0026] In the second cycle problem model, the cycle production volume, cycle demand volume, and actual warehouse construction point are inherited to obtain the second cycle modeling data to calculate the cycle delivery location. The second cycle modeling data includes decision cycle variables, optimization cycle objectives, and setting constraint cycle conditions.

[0027] By adopting the above technical solution, the production capacity of the production line and the demand fulfillment of the factory are optimized and calculated based on the first cycle problem model to obtain the cycle production volume and cycle demand volume. Then, based on the second cycle problem model, the cycle production volume, the cycle demand volume, and the actual warehouse construction point are inherited and calculated to obtain the corresponding cycle delivery location. This solves the problem of how to meet demand at low cost within a few months. The multi-time cycle problem is broken down into three sub-problems: production planning, warehouse location selection, and delivery optimization. These sub-problems are solved step by step and then summarized to obtain the final decision, reducing the difficulty of solving the problem.

[0028] In some embodiments, the periodic constraint includes a periodic flow conservation constraint, and the periodic flow conservation constraint is set by the following steps:

[0029] Based on the set of production line locations and the actual warehouse construction points, the corresponding end-of-cycle inventory is obtained, and the cycle retention constraint is determined based on the cycle production volume, cycle demand volume, decision cycle variables, and end-of-cycle inventory.

[0030] In some embodiments, the optimization cycle objective includes a third total supply cost, which is obtained by means of the following steps:

[0031] The corresponding unit inventory cost is obtained based on the actual warehouse establishment point, and the third total supply cost is determined based on the unit inventory cost, the decision cycle variable, and the unit cost set.

[0032] In some embodiments, the following steps are included before inheriting the cycle production volume, cycle demand volume, and actual warehouse construction point:

[0033] Determine whether a warehouse exists based on the actual warehouse construction point mentioned above;

[0034] If it exists, a periodic optimization signal is generated, and warehouse capacity constraints and warehouse throughput constraints are determined based on the periodic optimization signal. The constraint periodic conditions are then updated based on the warehouse capacity constraints and warehouse throughput constraints.

[0035] By adopting the above technical solution, it is determined whether a warehouse exists based on the actual warehouse construction point. If it exists, a periodic optimization signal needs to be generated, and the warehouse capacity constraint and warehouse throughput constraint are determined based on the periodic optimization signal, thereby updating the constraint periodic conditions and reducing the solution complexity caused by adding a time period.

[0036] Secondly, this application provides a warehouse location selection and distribution relationship optimization system based on step-by-step optimization, which adopts the following technical solution:

[0037] A warehouse location and distribution relationship optimization system based on step-by-step optimization, executing the warehouse location and distribution relationship optimization method based on step-by-step optimization described in the first aspect, includes:

[0038] The model setting module is used to divide the problem into two problem models, wherein the first problem model is represented as a factory-demand point optimization model and the second problem model is represented as a factory-warehouse-demand point optimization model.

[0039] In the first problem model, the delivery optimization module is used to acquire first modeling data to calculate the production calculation volume and the demand calculation volume. The first modeling data includes a first decision variable, a first optimization objective, and sets a first constraint condition.

[0040] In the second problem model, the location optimization module inherits the production calculation volume and demand calculation volume, and obtains the second modeling data to calculate the actual warehouse construction point and the actual delivery location. The second modeling data includes the second decision variable, the second optimization objective, and sets the second constraint condition.

[0041] In summary, this application includes at least one of the following beneficial technical effects:

[0042] 1. The problem is divided into two models. The first model only needs to consider the production line set and the demand point set, ignoring the intermediate warehouse distribution link. Based on the production capacity of the production line set to be optimized and the factory demand fulfillment rate of the demand point set, the optimal production and demand calculations are initially found. Then, the second problem module performs inherited optimization calculations on the production and demand calculations to obtain the actual warehouse construction points and the actual distribution warehouses corresponding to the demand points. A two-step modeling and solution scheme of production-distribution optimization and warehouse location and distribution relationship optimization is adopted, which can reduce the solution difficulty.

[0043] 2. For large-scale warehouse location selection and supply network optimization problems, the solution difficulty is reduced by decomposing the model into multiple steps. Compared with existing technologies, this application takes meeting demand and reducing costs as the optimization orientation, which can better guarantee the quality of the solution. The two methods are not mutually exclusive and can be used in combination. For example, the conditions can be relaxed first to exclude impossible candidate points or delivery routes based on the warehouse's coverage area, thus pre-pruning the problem. Then, this solution can be implemented to further accelerate the solution.

[0044] 3. The original problem is split into two sub-problems: "production-distribution optimization" and "warehouse location and distribution relationship optimization". When solving the problem, the preset location and distribution optimization model inherits the production calculation amount corresponding to the production line set and the demand calculation amount corresponding to the demand point set obtained from the preset production and distribution optimization model.

[0045] 4. In “Production-Distribution Optimization”, the total unit cost of the path with the lowest cost among all possible paths between the factory and the demand point is used as the unit supply cost of the factory-demand point, so as to generate a set of unit supply costs corresponding to the production factory element and the demand point element.

[0046] 5. For multi-period problems, the "warehouse location and delivery relationship optimization" problem is further broken down into problems without time variables and problems with time variables, which is equivalent to completing the warehouse location and delivery relationship optimization in two steps. Attached Figure Description

[0047] Figure 1 This is a conventional processing method for the warehouse location selection and distribution relationship optimization process provided in the embodiments of this application;

[0048] Figure 2 This is a schematic diagram of the factory direct delivery and forward warehouse transit distribution method provided in the embodiments of this application;

[0049] Figure 3 This is a block diagram of a warehouse location and distribution relationship optimization method based on step-by-step optimization provided in an embodiment of this application;

[0050] Figure 4 This is a block diagram illustrating the method for obtaining the first total supply cost provided in an embodiment of this application;

[0051] Figure 5 This is a block diagram illustrating a method for obtaining periodic delivery locations provided in an embodiment of this application.

[0052] Figure 6 This is a block diagram of the time cycle processing method of the warehouse location selection and distribution relationship optimization system based on step-by-step optimization provided in the embodiments of this application. Detailed Implementation

[0053] To better understand the purpose, technical solutions, and advantages of this application, it has been described and illustrated below with reference to the accompanying drawings and embodiments. However, those skilled in the art should understand that this application can be implemented without these details. In some cases, to avoid obscuring various aspects of this application due to unnecessary description, well-known methods, processes, systems, components, and / or circuits already described at a higher level will not be elaborated upon. It will be apparent to those skilled in the art that various modifications can be made to the embodiments disclosed in this application, and the general principles defined in this application can be applied to other embodiments and application scenarios without departing from the principles and scope of this application. Therefore, this application is not limited to the illustrated embodiments, but conforms to the broadest scope consistent with the scope of protection claimed in this application.

[0054] To reduce the difficulty of solving the problem, this application discloses a step-by-step optimization method for warehouse location and distribution relationship optimization, which is applied to a step-by-step optimization system for warehouse location and distribution relationship optimization. It is mainly applicable to enterprise-level software such as logistics management systems (WMS) and supply chain planning platforms (SCP) to solve the problem of multiple warehouse location and distribution relationship between warehouses and demand points, especially addressing the computational bottleneck in large-scale scenarios.

[0055] like Figure 3 As shown, a method for optimizing warehouse location and distribution relationships based on step-by-step optimization includes the following steps:

[0056] S100, divide into two problem models.

[0057] The first problem model is represented as a factory-demand point optimization model, and the second problem model is represented as a factory-warehouse-demand point optimization model. The first problem model only considers the relationship between any factory and any demand point, ignoring the intermediate warehouse distribution link. It seeks the optimal combination based on the production capacity of the factory-production line, the demand quantity of the demand point, and the supply cost from the factory to the demand point. The second problem model considers minimizing transportation and warehousing costs, finding the lowest-cost supply path, and simultaneously selecting warehouse locations and determining the distribution factory / warehouse for the demand point.

[0058] S200, in the first problem model, obtain the first modeling data to calculate the production calculation amount and the demand calculation amount.

[0059] The first modeling data includes the first decision variables, the first optimization objective, and the first constraints. The first decision variables include production line capacity and the amount of demand fulfilled by the factory. The first optimization objective includes minimizing the amount of unfulfilled demand and minimizing the first total supply cost.

[0060] In one embodiment, before acquiring the first modeling data to calculate the production and demand quantities, the minimum supply cost from each factory to each demand point is calculated. Based on the aforementioned basic idea, if a factory supplies to a demand point, it will choose the path with the lowest cost among all possible paths between them. The unit total logistics cost of this path is the unit supply cost from that factory to that demand point. Therefore, before acquiring the first modeling data to calculate the production and demand quantities, the following steps are also included:

[0061] S110: Obtain the corresponding set of production line locations and the set of demand points based on the factory to be optimized, and generate the unit supply cost based on the set of production line locations and the set of demand points. The unit supply cost represents the lowest cost among all paths between the production line location and the demand point location.

[0062] The formula for calculating unit supply cost is as follows:

[0063] ;

[0064] Where F represents the set of factories, C represents the set of candidate warehouses, and D represents the set of demand points. Characterizes the unit transportation cost from the factory to the candidate warehouse. The unit transportation cost from the warehouse candidate point to the demand point. It represents the unit transportation cost from the factory to the demand point.

[0065] Combination Figure 4 In one embodiment, the first optimization objective includes a first total supply cost, and the method for obtaining the first total supply cost includes the following steps:

[0066] S210: Obtain the corresponding production line capacity and production cost based on the production line set, and obtain the corresponding factory demand fulfillment quantity based on the demand point set.

[0067] S220 determines the first total supply cost based on production line capacity, factory demand fulfillment, production line cost, and unit supply cost.

[0068] The formula for minimizing the first total supply cost is as follows: .

[0069] Where F is the set of factories, C is the set of candidate warehouses, and D is the set of demand points. It is the set of production lines of factory i. This represents the set of SKUs that the factory-production line can produce, where U represents the complete set of SKUs. Production capacity of the production line. To meet the factory's demand. Production line cost for factory-production line-SKU.

[0070] Furthermore, the calculation method for minimizing unfulfilled demand is as follows: .

[0071] in, It is the demand for SKUs at the demand points. It refers to the demand fulfillment rate of each SKU.

[0072] In addition, the first set of constraints includes factory production line constraints, factory demand delivery constraints, and factory demand flow conservation constraints. The formula for calculating the factory production line constraints is as follows: The formula for calculating the factory demand flow conservation constraint is: or The formula for calculating the factory demand delivery constraint is: .

[0073] in, This represents the maximum capacity (production hours) of the factory / production line. This represents the minimum shipment / delivery quantity. By solving the first problem model in step S200, the production calculation quantity for each item can be obtained. The formula for calculating the production calculation quantity is: The formula for calculating the required quantity corresponding to the set of required points is as follows: .

[0074] It should be noted that the specific solution optimization process uses existing technology and the first constraint condition to gradually optimize the production line capacity and the factory demand to meet the first optimization objective, and finally obtains the corresponding production calculation and demand calculation.

[0075] S300, in the second problem model, inherits the production calculation quantity and demand calculation quantity, and obtains the second modeling data to calculate the actual warehouse construction point and the actual delivery location.

[0076] The second modeling data includes a second decision variable, a second optimization objective, and second constraints. The second decision variable includes parameters such as whether a candidate warehouse point is selected, direct demand parameters, warehouse candidate parameters, and warehouse distribution parameters. Direct demand parameters refer to the total quantity of products the factory directly sends to the demand point. Warehouse candidate parameters refer to the total quantity of products the factory sends to the candidate warehouse point. Warehouse distribution parameters refer to the total quantity of products sent from the candidate warehouse point to the demand point.

[0077] It should be noted here that the parameter Z determines whether a warehouse candidate point is selected. p ∈{0,1}, p∈C.

[0078] Requires direct delivery parameters QS i,q,k ∈R+, i∈F, q∈D, k∈S i,j Warehouse candidate parameters QT i,p,k ∈R+, i∈F, p∈C, k∈S i,j Warehouse delivery parameters QD p,q,k ∈R+, p∈C, q∈D, k∈U.

[0079] In another embodiment, the second optimization objective includes a second total supply cost, and the second total supply cost is obtained by the following steps:

[0080] Obtain the corresponding set of unit costs based on the second decision variable, and determine the second total supply cost based on the second decision variable and the corresponding set of unit costs.

[0081] In the second problem model, the unit cost set includes the unit transportation cost CT from the factory to the candidate warehouse. i,p The unit transportation cost (CD) from the candidate warehouse point to the demand point p,q Unit transportation cost CS from the factory to the demand point i,q And the fixed construction cost CB of the warehouse candidate sites p The second optimization objective is based on the principle of minimizing the second total supply cost. Therefore, the formula for calculating the second optimization objective is: min(Σ i,q,k CS i,q *QS i,q,k + Σ i,p,k CT i,p *QT i,p,k + Σ p,q,k' CD p,q *QD p,q,k' + Σ p CB p *Z p ), i∈F, p∈C, q∈D, k∈S i,j , k'∈U.

[0082] In one embodiment, the second constraint includes a delivery flow conservation constraint, which is generated by the following steps:

[0083] S310 determines the distribution flow conservation constraints based on production calculations, demand calculations, and the second decision variable.

[0084] The second constraint includes demand point receiving constraints, warehouse throughput constraints, and warehouse construction conditions and demand delivery conditions determined based on warehouse candidate parameters and warehouse delivery parameters.

[0085] Demand point acceptance constraints include minimum shipment quantity constraints and minimum delivery quantity constraints. Therefore, the formula for calculating demand point acceptance constraints is: QS i,q,k ≥ Q min Or QS i,q,k = 0, QT i,p,k ≥ Q min Or QT i,p,k = 0, QD p,q,k' ≥Q min Or QD p,q,k' = 0, i∈F, p∈C, q∈D, k∈S i,j , k'∈U; warehouse throughput constraint is Σ i,k QT i,p,k +Σ q,k' QD p,q,k' ≤ TP max i∈F, p∈C, q∈D, k∈S i,j , k'∈U. The warehouse construction condition is only if Z p When = 1, the demand delivery condition is QT. i,p,k QD p,q,k' ≥ 0. The inherited flow conservation constraint is QP'. i,k = Σ q QS i,q,k + Σ p QT i,p,k , Σ i QT i,p,k = Σ q QD p,q,k' , Σ p QD p,q,k' = Q' q,k' i∈F, p∈C, q∈D, k∈S i,j , k'∈U. Position size constraint, Σ p Z p ≤ w, p∈C.

[0086] Where w is the maximum position size, TP max This represents the maximum throughput of the warehouse. The formula for calculating production volume is: The formula for calculating the required quantity corresponding to the set of required points is: Only when both the warehouse construction conditions and the demand delivery conditions are met will the candidate warehouse location be the actual warehouse construction location (QS). i,q,k Positive factory and QD p,q,k' A positive value indicates the actual delivery location of the demand point.

[0087] The above models and solution processes are all for a single time period, such as a month, with the input being the monthly capacity of each factory / production line and the monthly demand of each demand point. However, in real-world applications, warehouse location selection is a low-frequency decision, at least on a quarterly basis. The challenge is how to meet demand at low cost over a period of several months. The biggest change after incorporating the time factor is the inventory transfer relationship between adjacent periods. For example, if a month's demand cannot be met due to capacity constraints, production can be advanced, and the demand for subsequent months can be met by maintaining inventory in the factory / warehouse. This change needs to be reflected in the flow conservation constraint, which in the original problem becomes I. i,k,t-1 + Σ j QP i,j,k,t = Σ q QS i,q,k,t + Σ p QT i,p,k,t + I i,k,t I p,k',t-1 + Σ i QT i,p,k,t = Σ q QD p,q,k',t +I p,k',t , Σ p QD pqk' = Q qk' , i∈F, j∈L i p∈C, q∈D, k∈S ij , k'∈U, t represents the time period, and I represents the inventory at the end of the period.

[0088] It's important to note that once inventory is introduced, inventory costs must be considered, meaning a cost item needs to be added to the total supply cost. The total supply cost in the original problem then becomes Σ. i,j,k,t CP i,j,k *QP i,j,k,t + Σ i,q,k,t CS i,q *QS i,q,k,t +Σ i,p,k,t CT i,p *QT i,p,k,t + Σ p,q,k',t CD p,q *QD p,q,k',t + Σp CB p *Z p + Σ p,k',t CW p *I p,k',t , i∈F, j∈L i p∈C, q∈D, k∈S i,j , k'∈U, CW p This represents the unit inventory cost of a warehouse. Inventory costs are complex, including rental costs, personnel operating costs, capital tied up in inventory, and so on. Another inventory-related factor is warehouse capacity constraints, i.e., Σ. k' I p,k',t ≤ SC p SC p Indicates warehouse capacity.

[0089] The addition of these factors will further increase the difficulty of solving the problem, making step-by-step solutions or problem simplification almost a must. The aforementioned solution is still effective. In the first step, "production-distribution optimization", time factors need to be considered to generate production plans for each time period. In the second step, "warehouse location and distribution relationship optimization", warehouse location is a long-term decision and does not need to consider time variables, but distribution relationships may change with demand, so it is necessary to solve for distribution relationships for each time period.

[0090] It's important to note that directly solving the "warehouse location and delivery relationship optimization" problem with time periods is very challenging. First, disregarding the time variable, we obtain the actual warehouse location by following steps S100-S300. Next, building upon the results of the previous step, we introduce the time variable and solve for the delivery relationships for each time period. The specific steps are as follows.

[0091] Reference Figure 5 In one embodiment, after obtaining the actual delivery locations corresponding to the actual warehouse construction points and demand points, the following steps are also included:

[0092] S400 sets up two cyclical problem models.

[0093] The first-cycle problem model represents a cyclical optimization model between the factory and the demand point, while the second-cycle problem model represents a cyclical optimization model between the factory, warehouse, and demand point. The first-cycle problem model considers the optimization problem between the factory and the demand point within a time period, for example, a monthly period. The second-cycle problem model considers the optimization of warehouse location and distribution relationships within a time period, given inventory transfer relationships.

[0094] S500, in the first cycle problem model, obtains the first cycle modeling data to calculate the cycle production and cycle demand.

[0095] It's important to note that the generation method for the first-cycle problem model is the same as the calculation method for the first-cycle problem model. The difference lies in the fact that the first-cycle modeling data for the first-cycle problem model consists entirely of parameters within a specific time period. The first-cycle modeling data includes input parameters, computational objectives, and computational conditions.

[0096] For example, using a monthly time period, the input parameters of the first-cycle problem model are the monthly production capacity of the production line and the monthly demand corresponding to the demand point. The calculation objectives represent minimizing the monthly non-fulfillment volume and the monthly total supply cost. The calculation conditions are the same as or similar to the first constraint conditions, so they will not be elaborated further here. Therefore, after calculation based on the first-cycle problem model, we obtain the cycle production volume and cycle demand volume.

[0097] In the second-cycle problem model, S600 inherits the cycle production volume, cycle demand volume, and actual warehouse construction point to obtain second-cycle modeling data to calculate the cycle delivery location.

[0098] The second-cycle modeling data includes decision-making cycle variables, optimization cycle objectives, and set cycle constraints. Decision-making cycle variables include cycle direct delivery volume, cycle shipment volume, and cycle distribution volume. Cycle direct delivery volume refers to the volume of goods shipped from the factory to the demand point within a given time period. Cycle shipment volume refers to the volume of goods shipped from the factory to the actual warehouse establishment point within a given time period. Cycle distribution volume refers to the volume of goods distributed from the actual warehouse establishment point to the demand point within a given time period. Specifically, cycle direct delivery volume (QS) i,q,k,t ∈R+, i∈F, q∈D', k∈S i,j,t Weekly shipment volume (QT) i,p,k,t ∈R+, i∈F, p∈C, k∈S i,j,t Periodic delivery volume (QD) p,q,k,t ∈R+, p∈C, q∈D', k∈U.

[0099] Reference Figure 6In summary, for multi-time-cycle problems, we essentially break down the original problem into three sub-problems: production planning, warehouse location selection, and distribution optimization. We solve these sub-problems step by step and then summarize the results to arrive at the final decision. First, we use the first-cycle problem model to solve the "production-distribution optimization" problem with time variables, obtaining the factory-production line-SKU-cycle production volume and the demand point-SKU-cycle demand volume. Next, we use the second-cycle problem model to solve the "warehouse location selection and distribution relationship optimization" problem without time variables, inheriting the factory-production line-SKU production volume and demand point-SKU demand volume from steps S100-S200 to obtain the actual warehouse location. Finally, we use the second-cycle problem model to solve the "warehouse location selection and distribution relationship optimization" problem with time variables, inheriting the factory-production line-SKU-cycle production volume, demand point-SKU-cycle fulfillment volume, and actual warehouse location from the first two steps to obtain the corresponding cycle distribution location for the factory / warehouse-demand point.

[0100] In one embodiment, the periodic constraint includes a periodic flow conservation constraint, and the periodic flow conservation constraint is set by the following steps:

[0101] S610 obtains the corresponding end-of-cycle inventory based on the set of production line locations and the actual warehouse construction points, and determines the cycle flow conservation constraints based on cycle production volume, cycle demand volume, decision cycle variables, and end-of-cycle inventory.

[0102] Specifically, the formula for calculating the conservation constraint of periodic flow is as follows:

[0103] I i,k,t-1 + QP' i,k,t = Σ q QS i,q,k,t + Σ p QT i,p,k,t + I i,k,t I p,k',t-1 + Σ i QT i,p,k,t =Σ q QD p,q,k',t + I p,k',t , Σ p QD p,q,k',t = Q' q,k',t , i∈F, j∈L i p∈C, q∈D', k∈S i,j,t , k'∈U;

[0104] Among them, QP' i,k,t This refers to the cycle production volume. Q' q,k',t This represents the periodic demand. D' is the set of actual warehouse construction points obtained from steps S100-S300.

[0105] The constraint period conditions also include demand point receiving constraints and warehouse throughput constraints, and the setting methods for demand point receiving constraints and warehouse throughput constraints are the same as or similar to those in the second problem model, so we will not go into too much detail here.

[0106] In one embodiment, the optimization cycle objective includes a third total supply cost, which is obtained by the following steps:

[0107] S620 obtains the corresponding unit inventory cost based on the actual warehouse construction point, and determines the third total supply cost based on the unit inventory cost, decision cycle variables, and unit cost set.

[0108] Specifically, the optimization cycle objective is based on minimizing the third total supply cost, and the optimization cycle objective is calculated as follows:

[0109] min(Σ i,q,k,t CS i,q,t *QS i,q,k,t + Σ i,p,k,t CT i,p,t *QT i,p,k,t + Σ p,q,k',t CD p,q,t *QD p,q,k',t + Σ p,k',t CW p *I p,k',t ), i∈F, p∈C, q∈D', k∈S i,j,t , k'∈U.

[0110] Among them, CW p Characterizes the unit inventory cost of a warehouse.

[0111] In one embodiment, the following steps are included before inheriting the cycle production volume, cycle demand volume, and actual warehouse construction point:

[0112] S510 determines whether a warehouse exists based on the actual warehouse construction point.

[0113] S520, if it exists, generates a periodic optimization signal, determines the warehouse capacity constraint and warehouse throughput constraint based on the periodic optimization signal, and updates the constraint periodic condition based on the warehouse capacity constraint and warehouse throughput constraint.

[0114] Specifically, the process involves checking if a corresponding warehouse location exists within the set of actual warehouse construction points. If a warehouse location exists within the set, it is determined that a warehouse needs to be built. The processor then generates a periodic optimization signal and determines the warehouse capacity constraint and warehouse throughput constraint based on this signal to update the constraint periodic conditions. The warehouse capacity constraint is calculated as follows: Σ k' I p,k',t ,≤ SCp The warehouse throughput constraint is calculated as follows: Σ i,k QT i,p,k + Σ q,k' QD p,q,k' ≤ TP max i∈F, p∈C, q∈D', k∈S i,j,t , k'∈U, SC p For warehouse storage capacity. TP max This represents the warehouse's maximum throughput.

[0115] If there is no warehouse location within the set, it is determined that there is no warehouse to be built. In this case, the constraint period condition is the same as or similar to the second constraint condition.

[0116] Specifically, based on the inheritance calculation of the second-cycle problem model, QS is obtained. i,q,k,t Positive factory and QD p,q,k',t A positive warehouse is the periodic delivery location for demand points within each time period.

[0117] This application also discloses a step-by-step warehouse location and delivery relationship optimization system, which executes a step-by-step optimization method for warehouse location and delivery relationship optimization.

[0118] The step-by-step warehouse location and delivery relationship optimization system includes a model setting module, a delivery optimization module, and a location optimization module. The model setting module divides the system into two problem models: a first problem model representing a factory-demand point optimization model, and a second problem model representing a factory-warehouse-demand point optimization model. In the first problem model, the delivery optimization module acquires first modeling data to calculate production and demand calculations. This first modeling data includes first decision variables, a first optimization objective, and sets first constraints. In the second problem model, the location optimization module inherits the production and demand calculations and acquires second modeling data to calculate the actual warehouse construction point and the actual delivery location. This second modeling data includes second decision variables, a second optimization objective, and sets second constraints.

[0119] The other functions performed in the above-mentioned model setting module, delivery optimization module, and site selection optimization module, as well as the technical details of each function, are the same or similar to the corresponding features in the warehouse site selection and delivery relationship optimization method based on step-by-step optimization described above, so they will not be repeated here.

[0120] The implementation principle is as follows:

[0121] First, the processor determines the solution model by judging the current solution process. If it is solving a single time period, it enters the first problem model and the second problem model for solution. Based on the execution steps S100-S300, it determines the minimization of the location optimization objective based on the preset location selection and delivery optimization model to obtain the corresponding inheritance calculation results. When the inheritance calculation results meet the warehouse construction conditions, the corresponding warehouse candidate point is taken as the actual warehouse construction point. When the inheritance calculation results meet the demand delivery conditions, the corresponding factory and warehouse are taken as the actual delivery location of the demand point.

[0122] If solving a problem with multiple time periods, the process proceeds to the first problem model, the second problem model, the first cycle problem model, and the second cycle problem model. Based on steps S100-S300, actual warehouse construction points are generated, and the monthly production capacity of the production line and the monthly demand corresponding to the demand points are optimized and calculated based on the first cycle problem model to obtain the cycle production volume and cycle demand volume.

[0123] Finally, based on the second cycle problem model, the cycle production volume, cycle demand volume, and actual warehouse construction point are inherited and calculated to obtain the corresponding cycle delivery location.

[0124] It should be understood that although the steps in the flowcharts in the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order in which these steps are performed, and they may be performed in other orders.

[0125] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A warehouse location and distribution relationship optimization method based on step-by-step optimization, characterized by, The method comprises the following steps: dividing two problem models, wherein a first problem model is represented as a factory-demand point optimization model, and a second problem model is represented as a factory-warehouse-demand point optimization model; in the first problem model, first modeling data is obtained to calculate production calculation and demand calculation, the first modeling data comprises first decision variables, a first optimization target and sets first constraint conditions, the first decision variables comprise line production capacity and factory demand satisfaction, the first optimization target comprises demand non-performance quantity minimization and first total supply cost minimization, and the first constraint conditions comprise factory line constraints, factory demand delivery constraints and factory demand flow conservation constraints; in the second problem model, the production calculation and the demand calculation are inherited, and second modeling data is obtained to calculate actual warehouse building points and actual distribution locations, the second modeling data comprises second decision variables, a second optimization target and sets second constraint conditions, the second decision variables comprise warehouse candidate point selection parameters, demand direct delivery parameters, warehouse candidate parameters and warehouse distribution parameters; wherein, before the first modeling data is obtained to calculate the production calculation and the demand calculation, the following steps are further included: according to a to-be-optimized factory, a corresponding line location set and a demand point location set are obtained, and a unit supply cost is generated according to the line location set and the demand point location set, the unit supply cost representing the lowest cost in all paths between the line location and the demand point location; the first total supply cost is obtained in the following manner: according to the line location set, corresponding line production capacity and line production cost are obtained, and according to the demand point location set, corresponding factory demand satisfaction is obtained; the first total supply cost is determined based on the line production capacity, the factory demand satisfaction, the line production cost and the unit supply cost; the second constraint conditions comprise distribution flow conservation constraints, and the distribution flow conservation constraints are generated in the following manner: the distribution flow conservation constraints are determined based on the production calculation, the demand calculation and the second decision variables; the second optimization target comprises second total supply cost minimization, and the second total supply cost is obtained in the following manner: a corresponding unit cost set is obtained according to the second decision variables, and the second total supply cost is determined based on the second decision variables and the corresponding unit cost set.

2. The warehouse siting and delivery relationship optimization method based on stepwise optimization according to claim 1, characterized in that, after the actual warehouse building points and the actual distribution locations corresponding to the demand points are obtained, the following steps are further included: two periodic problem models are set, wherein a first periodic problem model is represented as a factory-demand point periodic optimization model, and a second periodic problem model is represented as a factory-warehouse-demand point periodic optimization model; in the first periodic problem model, first periodic modeling data is obtained to calculate periodic production and periodic demand; in the second periodic problem model, the periodic production and the periodic demand and the actual warehouse building points are inherited to obtain second periodic modeling data to calculate periodic distribution locations, the second periodic modeling data comprises decision periodic variables, optimization periodic targets and sets constraint periodic conditions.

3. The warehouse siting and delivery relationship optimization method based on stepwise optimization according to claim 2, characterized in that, The constraint cycle condition comprises a cycle flow conservation constraint, and the cycle flow conservation constraint is set in the following manner: The cycle flow conservation constraint is determined based on the set of line positions and the actual warehouse building points.

4. The warehouse siting and delivery relationship optimization method based on stepwise optimization of claim 2, wherein, The optimization cycle target comprises a third total supply cost, and the third total supply cost is obtained in the following manner: The third total supply cost is determined based on the unit inventory cost, the decision cycle variable, and the unit cost set.

5. The warehouse siting and delivery relationship optimization method based on stepwise optimization of claim 3, wherein, Before inheriting the cycle production quantity, the cycle demand quantity, and the actual warehouse building points, the following step is further included: Based on the actual warehouse building points, it is determined whether there is a building warehouse; If there is, a cycle optimization signal is generated, and the warehouse capacity constraint and the warehouse throughput constraint are determined based on the cycle optimization signal, and the constraint cycle condition is updated based on the warehouse capacity constraint and the warehouse throughput constraint.

6. A step-based warehouse siting and distribution relationship optimization system, characterized by, The warehouse site selection and distribution relationship optimization method based on step-by-step optimization comprises the following steps: A model setting module is used to divide two problem models, wherein a first problem model represents a factory-demand point optimization model, and a second problem model represents a factory-warehouse-demand point optimization model; In the first problem model, a distribution optimization module is used to obtain first modeling data to calculate a production calculation quantity and a demand calculation quantity, wherein the first modeling data comprises a first decision variable, a first optimization target, and a first constraint condition; In the second problem model, a site selection optimization module is used to inherit the production calculation quantity and the demand calculation quantity, and obtain second modeling data to calculate actual warehouse building points and actual distribution positions, wherein the second modeling data comprises a second decision variable, a second optimization target, and a second constraint condition.

Images