A two-phase cross-depot vehicle scheduling algorithm for supply disruptions

Abstract

This application provides a two-stage cross-warehouse vehicle scheduling algorithm for supply disruptions. The method includes: acquiring product supply shortage data for inbound vehicles; constructing a demand decomposition algorithm by combining the product supply shortage data with the initial demand matrix of outbound vehicles and the unit shortage cost of each outbound vehicle for different product types, and calculating a demand reduction matrix; calculating the adjusted outbound vehicle demand matrix based on the demand reduction matrix and calculating the shortage cost; constructing a mathematical programming model based on the shortage cost, transit time cost, and outbound vehicle scheduling cost; establishing cross-warehouse center transit models under single and multiple loading modes, and designing a mode conversion algorithm; designing a two-stage percentage threshold heuristic algorithm to solve the cross-warehouse center transit model, and outputting the updated outbound vehicle demand matrix, scheduling scheme, and all cost information; minimizing product time cost loss and shortage loss through dynamic demand adjustment and flexible operation modes, thereby improving the supply chain's resilience.

Description

Technical Field

[0001] This application relates to the field of supply chain cross-delivery scheduling technology, and more specifically, to a two-stage cross-delivery vehicle scheduling algorithm for supply disruptions. Background Technology

[0002] Adopting cross-docking scheduling can effectively reduce transit time and is a key means to improve product logistics efficiency. Traditional cross-docking research focuses on time optimization under supply and demand balance and adopts a single unloading / loading model, meaning that vehicles must complete all unloading / loading tasks before leaving the station. However, when supply disruptions lead to actual arrivals being less than planned demand, the rigid scheduling strategy of the traditional model is relatively inefficient: on the one hand, it cannot intelligently prioritize demand to the most urgent areas (where stockout costs are highest); on the other hand, outbound vehicles may occupy the station for a long time while waiting for a specific product that has not yet arrived, hindering other vehicles that can be loaded immediately and significantly increasing time costs.

[0003] Therefore, there is an urgent need for a flexible scheduling method for cross-depot vehicles in the face of supply disruptions. This method should possess two core functions: (1) dynamically adjusting the allocation of each outbound vehicle (representing different regions) based on the urgency of product demand during supply disruptions; and (2) adaptively selecting whether to adopt a multiple loading (ML) mode (allowing outbound vehicles to enter and exit the platform multiple times) to achieve an optimal balance between increased scheduling costs and reduced time costs. Currently, no research has effectively integrated these two aspects for joint optimization. Summary of the Invention

[0004] The purpose of this application is to provide a two-stage cross-warehouse vehicle scheduling algorithm for supply disruptions. By constructing cross-warehouse center transfer models under single and multiple loading modes, and enabling mode switching, the algorithm minimizes product time cost losses and stockout losses through dynamic demand adjustment and flexible operation modes, thereby improving the supply chain's resilience to disruptions.

[0005] This application also provides a two-stage cross-depot vehicle scheduling algorithm for supply disruptions, including:

[0006] Obtain product supply shortage data due to supply disruptions in inbound vehicles, construct a demand decomposition algorithm by combining the product supply shortage data with the initial demand matrix of outbound vehicles and the unit shortage cost of each outbound vehicle for different product types, and calculate the demand reduction matrix.

[0007] Calculate the adjusted outbound vehicle demand matrix based on the demand reduction matrix, and calculate the stockout cost;

[0008] Based on stockout costs, transit time costs, and outbound vehicle scheduling costs, a cross-warehouse center transit model is constructed with the goal of minimizing total costs.

[0009] Establish cross-warehouse transfer models under single and multiple loading modes, and design a mode conversion algorithm;

[0010] Based on the objective function and constraints, a three-layer coding scheme consisting of vehicle serial number, product type, and loading quantity is constructed. A two-stage percentage threshold heuristic algorithm is designed to solve the cross-warehouse transfer model, and the updated outbound vehicle demand matrix, scheduling scheme, and all cost information are output.

[0011] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruptions described in the embodiments of this application, a decomposition algorithm for depot vehicle demand is designed based on the data on the shortage of inbound vehicles, specifically including:

[0012] Let the total number of cases be the first. Out-of-stock quantity of products that cannot be delivered on time Product category collection Outbound vehicle For the Initial demand for this product , Generate outbound vehicles based on demand decomposition algorithm Demand Reduction Matrix ;

[0013] The requirement decomposition algorithm must at least satisfy the following constraints:

[0014] Constraint 1: The total decrease in demand must match the amount of supply shortage. ;

[0015] Constraint 2: The reduction in demand shall not exceed the initial demand and shall be non-negative. .

[0016] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruptions described in this application embodiment, the calculation of stockout costs based on the demand reduction matrix specifically includes:

[0017] Set up each outbound vehicle For the Unit stockout cost of this product ;

[0018] Calculate the demand reduction matrix based on the demand decomposition algorithm. ;

[0019] Initial demand matrix based on outbound vehicles Calculate the adjusted outbound vehicle demand matrix. The formula is as follows:

[0020] ;

[0021] Calculate the corresponding out-of-stock cost .

[0022] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruptions described in the embodiments of this application, a method is established to minimize the total cost. For the cross-warehouse central transshipment model with the objective of [objective], the total cost of the objective function is expressed as:

[0023]

[0024] in, This represents the total completion time consumed during the unloading / loading process of all products at the cross-docking center, multiplied by... Convert time into cost; the second item is the total cost of stockouts. This represents the unit dispatch cost incurred by a vehicle entering and leaving the loading platform once. This indicates the total number of times a vehicle enters or exits the loading platform.

[0025] The constraints include supply and demand balance for stockouts, balance of inbound vehicle transfers, balance of outbound vehicle demand, transfer-related constraints, unloading / loading time constraints for inbound / outbound vehicles, sorting constraints, and unloading / loading mode constraints.

[0026] Optionally, in the two-stage cross-warehouse vehicle scheduling algorithm for supply disruptions described in the embodiments of this application, cross-warehouse center transfer models are established under single and multiple loading modes, and a mode transformation algorithm is designed:

[0027] A cross-warehouse transfer model is constructed based on the sorting of vehicle serial numbers under a single and continuous loading scheduling mode.

[0028] Introducing batch loading quantity parameters This transforms a single, uninterrupted loading operation of a single outbound vehicle into a multi-stage loading operation that can be interrupted.

[0029] Assume that the outbound vehicles are adjusted due to supply disruption. For the Demand for this product The number of times it was loaded Passing the exam Quantity of each shipment Calculate using the following formula:

[0030]

[0031] in, ;

[0032]

[0033] in, For the floor operation, the original floor is rounded down. The process of loading goods onto a single, continuous vehicle is transformed into an interruptible process. Section loading operation, the quantity loaded in each operation is ;

[0034] Section loading operation introduction Each virtual vehicle corresponds to one loading quantity. ;

[0035] Each loading operation consists of three layers of information: vehicle serial number, product type, and loading quantity. This means the process begins from the vehicle leaving the warehouse. Loading the first The quantity of the products is ;

[0036] Each loading operation is randomly ordered. If adjacent loading operations come from the same outbound vehicle, the loading quantities are combined to obtain the combined outbound vehicle demand matrix. ,in This belongs to the set of vehicles leaving the warehouse after the introduction of virtual vehicles.

[0037] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruption described in this application embodiment, the two-stage percentage threshold heuristic algorithm solution model includes a first stage and a second stage. The first stage includes the following steps:

[0038] Initialize algorithm parameters, including the amount of stockouts due to supply disruptions. Unit out-of-stock cost Population size Probability of choice Maximum number of iterations Batch loading quantity With threshold ;

[0039] A demand reduction matrix is ​​generated based on a demand decomposition algorithm. Calculate the adjusted outbound vehicle demand matrix. ;

[0040] The scheduling mode adopts a single and continuous loading (SL) method, and the vehicle sequence number is randomly generated based on the coding method. The initial population of chromosomes;

[0041] Calculate the fitness value for each chromosome based on the total cost of the objective function. Including time cost, stockout cost, and scheduling cost, the fitness value function is: ;

[0042] Perform the selection operation, using a roulette wheel strategy to select the first... The probability of each chromosome , total selection One parent generation;

[0043] The crossover operation is performed using a partially mapped crossover (PMX) method, and the mutation operation is performed using a random swap mutation method.

[0044] Population update involves recording the chromosomes with the highest and lowest fitness in each generation, replacing the chromosome with the chromosome with the highest fitness, and iterating until the maximum number of iterations is reached. Output the total cost in SL mode. and the scheduling costs incurred. ;

[0045] judge With threshold The size, if The algorithm uses SL mode and outputs the results. The algorithm employs a multi-loading ML model to execute the second phase.

[0046] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruptions described in this application embodiment, the second stage includes the following steps:

[0047] Based on batch loading quantity Each vehicle leaving the warehouse The single and continuous loading process is transformed into Section loading operations;

[0048] right The loading and unloading operations are randomly sorted, and a mode transformation algorithm is applied to generate a randomly sorted outbound vehicle demand matrix. ,in This belongs to the set of vehicles leaving the warehouse after the introduction of virtual vehicles;

[0049] Based on the new outbound vehicle demand matrix It adopts an encoding method based on virtual vehicle serial numbers to perform roulette wheel selection, crossover, and mutation operations;

[0050] Determine if the termination condition is met; if so, terminate the process and output the total cost under the multiple loading ML model. If the conditions are not met, the roulette wheel selection, crossover, and mutation operations will be executed again.

[0051] Optionally, in the two-stage cross-depot vehicle scheduling algorithm for supply disruptions described in this application embodiment, the partial mapping cross-PMX includes the following steps:

[0052] After grouping the parent population into pairs, select the two parent chromosomes that need to be crossovered and generate two distinct random integers.

[0053] Cross over the random integer portion between the two parent genes; after the crossover, non-repeating numbers are retained, and duplicate numbers appearing outside the crossover portion are eliminated by partial mapping, using the correspondence of the crossover portion for mapping;

[0054] If the fitness value of the chromosome after crossover is higher than that before crossover, the crossover operation is accepted; if it is lower than that before crossover, the original chromosome is retained.

[0055] The remaining two parent chromosomes, after being grouped into pairs, are subjected to repeated crossover processing.

[0056] Due to the application of the above technical solution, the present invention has the following advantages compared with the prior art:

[0057] Integrated Innovation: For the first time, dynamic adjustment of demand under supply disruption and adaptive selection of flexible operating modes (SL / ML) are incorporated into a unified decision-making framework for joint optimization.

[0058] The algorithm is highly efficient: the proposed two-stage percentage threshold heuristic algorithm can intelligently determine which mode is more efficient, avoiding blind calculations under both modes and improving solution efficiency. Meanwhile, by introducing batch loading quantity parameters and using a three-layer coding scheme based on vehicle serial number, product type, and loading quantity, the algorithm effectively solves the modeling and solving challenges in flexible and complex ML modes.

[0059] Enhanced resilience: This method can significantly reduce total costs under supply disruptions and improve the ability of cross-warehouse centers to cope with uncertainty. Attached Figure Description

[0060] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 This is a flowchart illustrating the two-stage cross-depot vehicle scheduling method provided in an embodiment of the present invention;

[0062] Figure 2 This is a flowchart of the two-stage percentage threshold heuristic algorithm provided in this embodiment of the invention;

[0063] Figure 3 This is a flowchart of the demand decomposition algorithm for supply disruptions provided in an embodiment of the present invention;

[0064] Figure 4This is a pseudocode diagram of the demand decomposition algorithm provided in an embodiment of the present invention;

[0065] Figure 5 This is a flowchart of cross-warehouse scheduling under a single loading mode based on a demand decomposition algorithm provided in an embodiment of the present invention.

[0066] Figure 6 This is a schematic diagram of loading operations under the multiple loading mode provided in the embodiments of the present invention;

[0067] Figure 7 This is a schematic diagram of converting vehicle sorting in SL mode into loading operation sorting in ML mode, provided by an embodiment of the present invention.

[0068] Figure 8 This is a schematic diagram of an algorithm provided by an embodiment of the present invention for converting ML mode into SL mode by introducing virtual vehicles. Detailed Implementation

[0069] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0070] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0071] Please refer to Figure 1 , Figure 1 This is a flowchart of a two-stage cross-depot vehicle scheduling algorithm for supply disruptions, as described in some embodiments of this application. This two-stage cross-depot vehicle scheduling algorithm for supply disruptions is used in a terminal device and includes the following steps:

[0072] Obtain product supply shortage data due to supply disruptions in inbound vehicles, construct a demand decomposition algorithm by combining the product supply shortage data with the initial demand matrix of outbound vehicles and the unit shortage cost of each outbound vehicle for different product types, and calculate the demand reduction matrix.

[0073] Calculate the adjusted outbound vehicle demand matrix based on the demand reduction matrix, and calculate the stockout cost;

[0074] Based on stockout costs, transit time costs, and outbound vehicle scheduling costs, a cross-warehouse center transit model is constructed with the goal of minimizing total costs.

[0075] Establish cross-warehouse transfer models under single and multiple loading modes, and design a mode conversion algorithm;

[0076] Based on the objective function and constraints, a three-layer coding scheme consisting of vehicle serial number, product type, and loading quantity is constructed. A two-stage percentage threshold heuristic algorithm is designed to solve the cross-warehouse transfer model, and the updated outbound vehicle demand matrix, scheduling scheme, and all cost information are output.

[0077] According to an embodiment of the present invention, a warehouse vehicle demand decomposition algorithm is designed based on warehouse vehicle supply shortage data, specifically including:

[0078] Let the total number of cases be the first. Out-of-stock quantity of products that cannot be delivered on time Product category collection Outbound vehicle For the Initial demand for this product , Generate outbound vehicles based on demand decomposition algorithm Demand Reduction Matrix ;

[0079] The requirement decomposition algorithm must at least satisfy the following constraints:

[0080] Constraint 1: The total decrease in demand must match the amount of supply shortage. ;

[0081] Constraint 2: The reduction in demand shall not exceed the initial demand and shall be non-negative. .

[0082] It should be noted that, according to embodiments of the present invention, a demand decomposition algorithm (such as...) is designed based on shortage data of various types of products caused by supply disruptions. Figure 3 (As shown) The dynamic demand adjustment data for Kuqa is calculated, specifically including:

[0083] When supply disruptions occur, demand needs to be adjusted based on the "out-of-stock cost priority" of fresh produce, with the core being the generation of a demand reduction matrix. ;

[0084] The adjusted outbound demand matrix is ​​calculated based on the demand decomposition algorithm to ensure that the total reduction in demand matches the supply interruption and shortage. Constraint 3 is also included, as follows:

[0085] The adjusted outbound vehicle demand matrix is ​​calculated, and subsequent vehicle scheduling processes are all based on it. The value.

[0086] There are multiple outbound vehicle demand reduction matrices that satisfy the above constraints. The specific choice needs to be made by comprehensively considering factors such as unit stockout cost. In the model solution section, the algorithm design that satisfies the above constraints is as follows: Figure 4 As shown.

[0087] According to an embodiment of the present invention, the stockout cost is calculated based on the demand reduction matrix, specifically including:

[0088] Set up each outbound vehicle For the Unit stockout cost of this product ;

[0089] Calculate the demand reduction matrix based on the demand decomposition algorithm. ;

[0090] Initial demand matrix based on outbound vehicles Calculate the adjusted outbound vehicle demand matrix. The formula is as follows:

[0091] ;

[0092] Calculate the corresponding out-of-stock cost .

[0093] According to embodiments of the present invention, a method is established to minimize total cost. For the cross-warehouse central transshipment model with the objective of [objective], the total cost of the objective function is expressed as:

[0094]

[0095] in, This represents the total completion time consumed during the unloading / loading process of all products at the cross-docking center, multiplied by... Convert time into cost; the second item is the total cost of stockouts. This represents the unit dispatch cost incurred by a vehicle entering and leaving the loading platform once. This indicates the total number of times a vehicle enters or exits the loading platform.

[0096] The constraints include supply and demand balance for stockouts, balance of inbound vehicle transfers, balance of outbound vehicle demand, transfer-related constraints, unloading / loading time constraints for inbound / outbound vehicles, sorting constraints, and unloading / loading mode constraints.

[0097] It should be noted that, based on constraints such as inbound / outbound vehicle transfer association and unloading / loading modes, a mixed-integer cross-warehouse transfer model is constructed using a single-loading (SL) loading mode, with the goal of minimizing time cost, stockout cost, and scheduling cost.

[0098] Objective function:

[0099]

[0100] Constraints:

[0101]

[0102]

[0103] The formula parameters are shown in the table below:

[0104]

[0105] Constraints (2-5) describe the quantity constraints that must be met when a supply disruption occurs. Constraints (6) and (7) ensure that the transfer quantity matches both the loading capacity of the inbound vehicles and the demand of the outbound vehicles. Constraint (8) specifies the variables. and The relationship between the two constraints is as follows: Constraints (9-11) provide a reasonable order rule for the entry and exit times of inbound vehicles. Constraint (12) ensures that no inbound vehicle is ahead of itself in the inbound vehicle sequence. Constraints (13-16) apply to outbound vehicles in a similar manner (i.e., specifying the entry and exit time order and reflexive constraints for outbound vehicles). Constraint (17) then links the entry and exit times of inbound and outbound vehicles.

[0106] like Figure 5 As shown, a specific example illustrates the solution scheme for the hybrid integer cross-warehouse transfer model proposed in this embodiment of the invention under the single-load (SL) mode:

[0107] This embodiment includes 2 inbound vehicles and 3 outbound vehicles, transporting 3 types of products. Inbound vehicle 1 experienced a supply disruption, resulting in the loss of products of types 2 and 3. Furthermore, unloading / loading 1 unit of product requires 1 unit of time, and the transfer time from the inbound platform to the outbound platform, as well as the transit time from the departure of one vehicle to the arrival of the next, are both 2 units.

[0108] Based on the above information, a decision is made among the three outbound vehicles to determine which vehicles' demand should be adjusted to achieve a balance between supply and demand while minimizing cost losses.

[0109] according to Figure 5Loading information for outbound trucks: The initial demand for 3 outbound trucks is... (The first row of the matrix indicates that Kuche 1 is loaded with one piece each of the first and third types of products, and so on).

[0110] based on Figure 4 The proposed requirement decomposition algorithm can generate requirements such as Figure 5 The two scheduling schemes shown are as follows:

[0111] Option ① involves reducing the demand for the third type of product in outbound vehicle 1 and the second type of product in outbound vehicle 3. At this time, the demand for outbound vehicles 1 (OT1), OT2, and OT3 are [1 0 0; 0 1 0; 0 0 1], respectively. It requires 3 vehicles to complete the cross-depot scheduling together, which requires 10 units of time cost, and also incurs the scheduling cost of 3 outbound vehicles.

[0112] Option ② involves reducing the demand for the third type of product in delivery vehicle 1 and the second type of product in delivery vehicle 2. At this point, only two outbound vehicles are needed to complete the cross-warehouse dispatch, requiring 8 units of time cost, while also incurring the dispatch cost of two outbound vehicles. Clearly, the total cost of this solution is less than that of solution ①.

[0113] According to an embodiment of the present invention, based on the traditional cross-warehouse transfer model under the single loading (SL) mode, in order to accelerate the transfer process and reduce costs, a cross-warehouse transfer model under the multiple loading (ML) mode is constructed (see product transfer process). Figure 6 ).

[0114] Figure 6 This demonstrates how, in order to quickly meet the needs of outbound vehicle 1 (requiring one piece each of product types 1 and 2), after loading product type 1, the vehicle leaves the platform, incurring vehicle downtime. The vehicle drives into warehouse 2, then loads one piece each of product type 1 and type 2 as required, and finally, there is an interruption. The outbound vehicle 1 returned to the platform to transfer product type 2, completing the scheduling. During this period, there were no delays in loading due to products not arriving at the loading platform on time, as is the case in the SL loading mode. The ML mode achieved continuity in the loading process and reduced transshipment time costs, but it also increased scheduling costs.

[0115] To establish a cross-warehouse transfer model under the ML (Multi-Level Marketing) model, this invention proposes a batch loading quantity parameter. This means that each time a truck leaves the warehouse to perform a loading operation, it must load at least [amount missing]. The quantity of products allows for the division of the loading process of outbound trucks into multiple segments, each called a loading operation. Each operation consists of three layers of information: from the first... The first truck loaded with goods leaving the warehouse The quantity of the products is .

[0116] Specifically, if based on the batch loading quantity Uninstall For products of a certain quantity, the total number of loading trips required. and the quantity loaded each time They are respectively:

[0117]

[0118] in, This is the floor operator. This represents the sequence number of the loading segment. According to formula (18-19), it can be... The loading process for a large quantity of products is divided into Each segment contains the following quantity of cargo: .

[0119] According to a specific embodiment of the present invention, the SL mode is converted into the ML mode, such as... Figure 7-8 As shown in the diagram. This embodiment includes two delivery vehicles and two types of products. Delivery vehicle 1 needs to load 25 units of product type 1 and 5 units of product type 2; delivery vehicle 2 needs to load 15 units of product type 2. The batch unloading quantities are set. .

[0120] From equations (18) and (19), it can be seen that the loading number for outbound vehicle 1 is 3, loading 10 and 15 units of the first type of product, and 5 units of the second type of product respectively; the loading number for outbound vehicle 2 is 1, loading 15 units of the second type of product. Therefore, the loading process requires 4 loading operations, each loading operation consisting of the outbound vehicle number, product type, and other relevant information. and loading quantity It consists of three elements.

[0121] Performing different sorting operations on the four loading operations will result in different outcomes, corresponding to different loading orders, as shown in Table 1 and... Figure 8 As shown:

[0122] Table 1

[0123]

[0124] As shown in Table 1, based on the outbound vehicle sequence number, the first possible loading order is 1-2-1-1. Since the latter two loading operations are for the same outbound vehicle, they can be combined (after combination, it can be regarded as introducing virtual vehicle 3). Summarizing the four loading operations, the outbound vehicle loading matrix under the introduction of virtual vehicles is obtained as follows: Similarly, the loading matrix corresponding to the second possible loading order is: ,See Figure 8 .

[0125] It is evident that although the loading operation is divided into the same four segments, arranging these four segments into different unloading sequences will result in different loading matrices, and thus different objective function values ​​(such as time cost, scheduling cost, etc.).

[0126] Therefore, for the ML model, the key to its mathematical model and solution algorithm lies in sorting the loading operations consisting of three layers of information, thereby transforming it into a SL model that incorporates virtual vehicles. The model transformation algorithm is summarized as follows:

[0127] Based on batch loading quantity Demand matrix for outbound vehicles The loading operation is divided into For the decomposition method, see equations (18) and (19).

[0128] Will The loading operations are arranged randomly, as shown in Table 1;

[0129] Based on this sorting, if two adjacent loading operations originate from the same outbound vehicle, they are merged; otherwise, a new virtual outbound vehicle is introduced, thus transforming the process into an outbound vehicle demand matrix expressed in matrix form. .

[0130] in, The matrix has more than This is equivalent to changing the behavior of outbound vehicles repeatedly entering and exiting the loading platform into the addition of virtual outbound vehicles performing loading operations.

[0131] Based on the demand matrix of inbound vehicles and the demand matrix for introducing virtual outbound vehicles The cost values ​​of each product during the cross-warehouse scheduling process can be calculated. The specific model is the same as that in equations (1)-(17), the difference being that... Become Thus, the cross-depot vehicle scheduling model under the ML mode has been fully constructed.

[0132] According to an embodiment of the present invention, cross-warehouse transfer models under single and multiple loading modes are established, and a mode transformation algorithm is designed:

[0133] A cross-warehouse transfer model is constructed based on the sorting of vehicle serial numbers under a single and continuous loading scheduling mode.

[0134] Introducing batch loading quantity parameters This transforms a single, uninterrupted loading operation of a single outbound vehicle into a multi-stage loading operation that can be interrupted.

[0135] Assume that the outbound vehicles are adjusted due to supply disruption. For the Demand for this product The number of times it was loaded Passing the exam Quantity of each shipment Calculate using the following formula:

[0136]

[0137] in, ;

[0138]

[0139] in, For the floor operation, the original floor is rounded down. The process of loading goods onto a single, continuous vehicle is transformed into an interruptible process. Section loading operation, the quantity loaded in each operation is ;

[0140] Section loading operation introduction Each virtual vehicle corresponds to one loading quantity. ;

[0141] Each loading operation consists of three layers of information: vehicle serial number, product type, and loading quantity. This means the process begins from the vehicle leaving the warehouse. Loading the first The quantity of the products is ;

[0142] Each loading operation is randomly ordered. If adjacent loading operations come from the same outbound vehicle, the loading quantities are combined to obtain the combined outbound vehicle demand matrix. ,in This belongs to the set of vehicles leaving the warehouse after the introduction of virtual vehicles.

[0143] Based on this, the sorting problem of loading operations in the complex ML model is transformed into the sorting problem of the outbound vehicle sequence number, that is, it is transformed into a single loading (SL) model after introducing virtual vehicles.

[0144] This paper proposes a cross-warehouse transfer model using a flexible multiple loading (ML) mode. It also proposes a model transformation algorithm to convert the model into a more easily solvable single loading (SL) mode cross-warehouse transfer model, thus achieving both feasibility and computational efficiency.

[0145] like Figure 2As shown in the embodiment of the present invention, the two-stage percentage threshold heuristic algorithm solution model includes a first stage and a second stage. The first stage includes the following steps:

[0146] Initialize algorithm parameters, including the amount of stockouts due to supply disruptions. Unit out-of-stock cost Population size Probability of choice Maximum number of iterations Batch loading quantity With threshold ;

[0147] A demand reduction matrix is ​​generated based on a demand decomposition algorithm. Calculate the adjusted outbound vehicle demand matrix. ;

[0148] The scheduling mode adopts a single and continuous loading (SL) method, and the vehicle sequence number is randomly generated based on the coding method. The initial population of chromosomes;

[0149] Calculate the fitness value for each chromosome based on the total cost of the objective function. Including time cost, stockout cost, and scheduling cost, the fitness value function is: ;

[0150] Perform the selection operation, using a roulette wheel strategy to select the first... The probability of each chromosome , total selection One parent generation;

[0151] The crossover operation is performed using a partially mapped crossover (PMX) method, and the mutation operation is performed using a random swap mutation method.

[0152] Population update involves recording the chromosomes with the highest and lowest fitness in each generation, replacing the chromosome with the chromosome with the highest fitness, and iterating until the maximum number of iterations is reached. Output the total cost in SL mode. and the scheduling costs incurred. ;

[0153] judge With threshold The size, if The algorithm uses SL mode and outputs the results. The algorithm employs a multi-loading ML model to execute the second phase.

[0154] It should be noted that the first stage includes steps S1-S8, as detailed below:

[0155] S1, obtains data on stockouts of various product categories caused by supply disruptions, such as... Figure 1 Supply disruptions occurred for product category 2.

[0156] S2. Correspondingly, the demand of all outbound vehicles must also be adjusted. Based on the unit stockout cost of each outbound vehicle for different product types (combined with data such as differences in cargo location, urgency, and freshness), this invention proposes a demand decomposition algorithm to calculate the dynamic demand adjustment data of outbound vehicles.

[0157] S3, based on the demand reduction matrix, calculates the adjusted outbound vehicle demand matrix to ensure that the total demand reduction matches the supply interruption and shortage, and calculates the corresponding cumulative stockout cost;

[0158] S4. Adjustments in outbound vehicle demand significantly impact the supply-demand matching of inbound and outbound vehicles and the cross-warehouse transfer process. Combining constraints such as inbound / outbound vehicle transfer association and unloading / loading modes, a mixed-integer cross-warehouse transfer model is constructed with the goal of minimizing time cost, stockout cost, and scheduling cost.

[0159] S5. Furthermore, based on the traditional single loading (SL) cross-warehouse transfer model, in order to speed up the transfer process and reduce costs, a multiple loading (ML) cross-warehouse transfer model is constructed.

[0160] S6. To improve the efficiency of model calculation, a batch loading quantity parameter is introduced. At the same time, a pattern transformation algorithm was designed to transform the ML pattern into an SL pattern that is easier to solve;

[0161] S7 proposes a two-stage percentage threshold heuristic algorithm to intelligently determine which mode (SL, ML) has higher efficiency and lower cost, avoiding blind calculations in both modes. Meanwhile, the coding scheme adopts a three-layer architecture based on vehicle serial number, product type, and loading quantity, effectively solving the modeling and solving challenges in the ML mode.

[0162] S8 outputs the optimal scheduling scheme based on the algorithm, obtaining the optimal outbound vehicle demand adjustment matrix, each inbound / outbound unloading / loading scheduling scheme, and various cost information.

[0163] According to an embodiment of the present invention, the second stage includes the following steps:

[0164] Based on batch loading quantity Each vehicle leaving the warehouse The single and continuous loading process is transformed into Section loading operations;

[0165] right The loading and unloading operations are randomly sorted, and a mode transformation algorithm is applied to generate a randomly sorted outbound vehicle demand matrix. ,in This belongs to the set of vehicles leaving the warehouse after the introduction of virtual vehicles;

[0166] Based on the new outbound vehicle demand matrix It adopts an encoding method based on virtual vehicle serial numbers to perform roulette wheel selection, crossover, and mutation operations;

[0167] Determine if the termination condition is met; if so, terminate the process and output the total cost under the multiple loading ML model. If the conditions are not met, the roulette wheel selection, crossover, and mutation operations will be executed again.

[0168] According to an embodiment of the present invention, partial mapping cross-PMX includes the following steps:

[0169] After grouping the parent population into pairs, select the two parent chromosomes that need to be crossovered and generate two distinct random integers.

[0170] Cross over the random integer portion between the two parent genes; after the crossover, non-repeating numbers are retained, and duplicate numbers appearing outside the crossover portion are eliminated by partial mapping, using the correspondence of the crossover portion for mapping;

[0171] If the fitness value of the chromosome after crossover is higher than that before crossover, the crossover operation is accepted; if it is lower than that before crossover, the original chromosome is retained.

[0172] The remaining two parent chromosomes, after being grouped into pairs, are subjected to repeated crossover processing.

[0173] A second aspect of the present invention provides a cross-database scheduling system, the system comprising:

[0174] The data acquisition module is used to obtain information such as supply interruption, product type parameters, and cross-database scheduling environment.

[0175] The optimization module is used to execute a two-stage percentage threshold heuristic algorithm and output the optimal unloading / loading sequence, outbound vehicle demand matrix, total cost, etc.

[0176] The control module is used to apply the obtained scheduling scheme to the job scheduling system of the cross-depot center.

[0177] This system can avoid situations where the arrival of goods does not match the current demand of outbound vehicles, resulting in long waiting times, thereby reducing transfer time. However, the multiple entries and exits of outbound vehicles into and out of the station also increase scheduling costs, so a balance needs to be struck between transfer time costs and scheduling costs.

[0178] In summary, this application has the following advantages:

[0179] Integrated Innovation: For the first time, dynamic adjustment of demand under supply disruption and adaptive selection of flexible operating modes (SL / ML) are incorporated into a unified decision-making framework for joint optimization.

[0180] The algorithm is highly efficient: the proposed two-stage percentage threshold heuristic algorithm can intelligently determine which mode is more efficient, avoiding blind calculations under both modes and improving solution efficiency. Meanwhile, by introducing batch loading quantity parameters and using a three-layer coding scheme based on vehicle serial number, product type, and loading quantity, the algorithm effectively solves the modeling and solving challenges in flexible and complex ML modes.

[0181] Enhanced resilience: This method can significantly reduce total costs under supply disruptions and improve the ability of cross-warehouse centers to cope with uncertainty.

[0182] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A two-stage cross-dock vehicle scheduling algorithm oriented to supply disruption, characterized in that, The application comprises the following steps: Obtaining product supply shortage data of inbound vehicle supply interruption, constructing a demand decomposition algorithm for the product supply shortage data and the initial demand matrix of outbound vehicles, the unit shortage cost of different product categories for each outbound vehicle, and calculating a demand reduction matrix; Calculating an adjusted outbound vehicle demand matrix based on the demand reduction matrix and calculating a shortage cost; Constructing a cross-dock center transfer model aiming to minimize the total cost based on the shortage cost, transfer time cost, and outbound vehicle scheduling cost; Establishing a cross-dock center transfer model under single and multiple loading modes and designing a mode conversion algorithm; Based on the objective function and the constraint conditions, a three-layer coding scheme composed of vehicle sequence number, product category, and loading quantity is constructed, a two-stage percentage threshold heuristic algorithm is designed to solve the cross-dock center transfer model, and an updated outbound vehicle demand matrix, scheduling scheme, and all cost information are outputted; Based on the inbound vehicle supply shortage data, an outbound vehicle demand decomposition algorithm is designed, which specifically includes: Let the total number of cases be the first. Out-of-stock quantity of products that cannot be delivered on time Product category collection Outbound vehicle For the Initial demand for this product , Generate outbound vehicles based on demand decomposition algorithm Demand Reduction Matrix ; The demand decomposition algorithm at least needs to meet the following constraint conditions: Constraint 1: The total amount of demand reduction matches the total amount of supply shortfall: ; Constraint 2: Reduction in demand is no more than initial demand and non-negative: ; Based on the demand reduction matrix, the shortage cost is calculated, which specifically includes: Let each outbound vehicle The unit shortage cost of the first kind of product ; Computing a demand reduction matrix based on a demand decomposition algorithm ; Based on the initial demand matrix of outbound vehicles , the demand-adjusted outbound vehicle demand matrix is calculated , as follows: ； Calculating the corresponding stock-out cost ; Establishing a goal of minimizing total cost The goal of the cross-docking center transshipment model, the objective function total cost is expressed as: ； wherein, represents the total completion time consumed by all products during the unloading / loading process at the cross-docking center, c represents the cost per unit time, which is multiplied by the time is converted into the cost; the second term is the total shortage cost; represents the unit scheduling cost consumed by the outbound vehicle to enter and exit the loading platform once, represents the total number of times the outbound vehicle enters and exits the loading platform. The constraint conditions include shortage supply and demand balance, inbound vehicle transfer balance, outbound vehicle demand balance, transfer correlation constraint, inbound / outbound vehicle unloading / loading time constraint, sequencing constraint, and unloading / loading mode constraint; Establishing a cross-dock center transfer model under single and multiple loading modes and designing a mode conversion algorithm: Based on the sequencing of vehicle sequence numbers, a cross-dock center transfer model under a single and continuous loading scheduling mode is constructed; Introducing a batch load quantity parameter Converting a single, uninterrupted continuous load operation of a truck to a multi-segment load operation that can be interrupted; Set the number of trucks to be dispatched due to supply disruption adjustment The demand for the first product , the number of times it is loaded and the quantity of the first load are calculated as follows: ； wherein ; ； wherein, is a floor operator, the originally first vehicle only 1 time and continuous loading process, into an interruptible segment loading operation, each operation loading quantity is ; Segment loading operation introduction A virtual outbound vehicle, and each virtual vehicle corresponds to 1 segment loading quantity ; Each piece of loading operation is composed of three layers of information, i.e. vehicle serial number, product type and loading quantity, which is from the vehicle out of the warehouse to the vehicle into the warehouse. Loading the first product with the quantity of ; Each loading operation is randomly sorted, and if the adjacent loading operations are from the same outbound vehicle, the loading quantities are merged, and finally the merged outbound vehicle demand matrix is obtained wherein belongs to the set of outbound vehicles after the virtual vehicle is introduced The two-stage percentage threshold heuristic algorithm for solving the model includes a first stage and a second stage, and the first stage includes the following steps: initializing algorithm parameters, including backorder quantity due to supply disruption , unit backorder cost , population size , selection probability , maximum number of iterations , batch shipment quantity and threshold value ; Generating a demand reduction matrix based on a demand decomposition algorithm , calculating a demand adjusted outbound truck demand matrix ; The scheduling mode of single and continuous loading SL is adopted, the coding mode based on vehicle serial number is adopted, and the initial population of the number of chromosomes is randomly generated chromosomes According to the target function total cost, the fitness value of each chromosome is calculated, and the total cost The fitness value function includes time cost, out-of-stock cost and scheduling cost ; performing a selection operation, selecting the first chromosome using a roulette wheel strategy with a probability of one chromosome , co-selecting a parent; Partial mapping crossover (PMX) is used to perform crossover operation, and random exchange mutation is used to perform mutation operation; Population update, record the highest and lowest fitness of the chromosomes in each generation of population, replace the lowest fitness chromosome with the highest fitness chromosome, iterate until the maximum iteration number is reached , output the total cost in SL mode , and the consumed scheduling cost ; determining the size of the threshold , if the algorithm adopts the SL mode and outputs the result, if , the algorithm adopts the multiple loading ML mode to perform the second stage.

2. The two-stage interdocking yard vehicle scheduling algorithm for supply disruption, according to claim 1, characterized in that, The second stage includes the following steps: Based on the number of batches loaded Converting each truck 's single and continuous loading process into a series of loading operations; To The loading operation is randomly sorted, and a mode transformation algorithm is applied to generate a randomly sorted outbound vehicle demand matrix Wherein It belongs to the outbound vehicle set after introducing virtual vehicles Based on the new demand matrix of the outbound vehicle , using a virtual vehicle number-based encoding method, performing roulette selection, crossover, and mutation operations; determining whether the termination condition is satisfied, if yes, ending and outputting the total cost under the multiple loading ML mode if not, performing the roulette selection, crossover and mutation operations again.

3. The two-stage interdiction-oriented cross-dock vehicle scheduling algorithm of claim 2, wherein, Partial mapping crossover (PMX) includes the following steps: After grouping the parent population in pairs, two parent chromosomes that need to be crossed are selected, and two different random integers are generated; Cross the parts between the two parent gene random integers; after crossing, the non-repeated numbers are retained, and the repeated numbers appearing outside the crossing part are eliminated by using partial mapping method, and the corresponding relationship of the crossing part is used for mapping; If the fitness value of the crossed chromosome is higher than that before crossing, the crossing operation is accepted, otherwise the original chromosome is retained; The remaining two parent chromosomes grouped in pairs are repeatedly crossed.

Images