Multimodal transport cross-border logistics path optimization method and system based on demand and tax rate uncertainty, medium and processor

Abstract

The invention discloses a multimodal transport cross-border logistics path optimization method and system based on demand and tax rate uncertainty, a medium and a processor, and relates to the technical field of cross-border logistics, comprising: constructing a cross-border logistics network system, the system comprising a supply point, a domestic transit point, a foreign transit point and a most total user demand point; model assumption and related variables are determined, the related variables comprise parameters and decision variables, the parameters comprise deterministic parameters and uncertainty parameters, a multi-objective optimization model is established, and the purpose is to minimize the total transportation cost and the total transportation time; constructing constraint conditions of the multi-objective optimization model; and solving the multi-objective optimization model by adopting a hybrid algorithm of a genetic algorithm and a hill climbing method to obtain an optimal transportation scheme. According to the method, the uncertainty of the demand and the tax rate is comprehensively considered, the multi-target optimization model is established, the actual situation of cross-border logistics can be comprehensively reflected, and the accuracy and practicability of path optimization are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cross-border logistics, and particularly relates to a multi-modal cross-border logistics path optimization method and system based on demand and tax rate uncertainty, a medium and a processor. BACKGROUND

[0002] Firstly, globalization exposes the logistics network to diverse risks from different countries and regions, such as politics, economy, and nature, which threatens the security and stability of the logistics network. Secondly, information asymmetry, technological progress, and high uncertainty of market demand also bring more uncertainties to the efficient operation of the logistics network. In particular, the failure of a single transportation mode or node can lead to the interruption of the entire logistics path, thereby triggering a chain reaction. Therefore, the optimization of cross-border logistics paths based on multi-modal transportation is of great significance for improving the logistics efficiency between Southeast Asia and China, reducing operating costs, and effectively responding to potential risks. By reasonably designing multi-modal transportation paths, the resilience of the logistics network can be enhanced, and the overall competitiveness of the regional economy can be improved.

[0003] Previous researches mostly focused on road transportation, with less research on water transportation and multi-modal transportation. However, in the cross-border logistics network, due to the limitations of single transportation mode in terms of timeliness, cost, risk control, etc., multi-modal transportation integrates various transportation modes such as road, rail, waterway, and air, realizes flexible switching of routes and collaborative allocation of resources, and effectively responds to unexpected events such as transportation interruption and node failure. As an efficient mode of cargo transportation, the optimization of multi-modal transportation routes significantly improves the transportation efficiency and the robustness and flexibility of the logistics network, showing stronger adaptability.

[0004] The single shortest path or the lowest cost path cannot meet the actual needs of multimodal transport. Logistics path optimization must take into account multiple objectives such as transportation cost, timeliness, service reliability and policy risk. When studying cross-border logistics network optimization decisions, most scholars consider the interaction and influence between different factors. For example, Chen X et al. considered goods with different attributes such as value and time, and established a multimodal transport low-carbon route decision optimization model with the goal of minimizing the sum of transportation carbon emission cost and time cost; Temizceri F T et al. proposed a double-objective multimodal transport problem focusing on carbon emission reduction, emphasizing the importance of sustainable multimodal transport in the context of global climate goals and government initiatives. Li S et al. took international trade between China, Myanmar and Vietnam as an example, and considered the characteristics of geographical space factors, compared the influence of geographical factors on cross-border logistics route selection by land and sea. In addition, the increasing demand of customers for perishable agricultural products also makes cold chain logistics very important, and efficient completion of cold chain logistics distribution within a time window, while involving temperature, humidity, truck capacity and other factors, ultimately ensures the quality of agricultural products and food safety. Chen J et al. determined the cold chain logistics path scheme of vehicles from the distribution center to the front warehouse by focusing on fixed cost, refrigeration cost, fuel consumption cost, carbon emission cost, time window violation cost and damage cost. Bei H et al. considered the use of refrigerated containers and customer satisfaction, and studied a multimodal transport route optimization method for perishable goods.

[0005] Cross-border logistics paths are not only subject to the structure of the transportation network, but also influenced by regional policies, tariff systems, infrastructure coordination levels and other institutional factors. Few existing path optimization models comprehensively consider demand and uncertain factors for path optimization.

[0006] Therefore, there is a need for a multimodal transport cross-border logistics path optimization method, system, medium and processor based on demand and tax rate uncertainty. SUMMARY

[0007] In view of the problems in the prior art, the present application provides a multimodal transport cross-border logistics path optimization method, system, medium and processor based on demand and tax rate uncertainty, which can comprehensively consider the uncertainty of demand and tax rate, establish a multi-objective optimization model, and fully reflect the actual situation of cross-border logistics, improving the accuracy and practicality of path optimization. The specific technical solutions are as follows: A multimodal transport cross-border logistics path optimization method based on demand and tax rate uncertainty, comprising: A network system of cross-border logistics is constructed, which includes a supply point, a domestic transfer point, a foreign transfer point and a total user demand point; Determine the model assumptions and relevant variables, including parameters and decision variables, including deterministic parameters and uncertain parameters, and establish a multi-objective optimization model with the goal of minimizing the total transportation cost and the total transportation time; Construct the constraints for the multi-objective optimization model; The multi-objective optimization model is solved using a hybrid algorithm combining genetic algorithm and hill climbing method to obtain the optimal transportation scheme.

[0008] Preferably, the model assumptions include: Consider cargo transportation from a single origin, with multiple modes of transport coexisting; there is no transportation between domestic transit points and overseas transit points; one domestic transit point can transport goods to multiple overseas transit points, and one overseas transit point can receive goods from multiple domestic transit points; one overseas transit point can transport goods to multiple demand points, and each demand point is served by only one overseas transit point; transshipment only occurs at transit nodes, and each order undergoes at most one change of transport mode at each node; the demand volume of goods at each demand point is uncertain; only the transportation and delivery of the same type of goods are considered.

[0009] Preferably, the deterministic parameters include the operating costs of domestic and overseas transit points, transshipment costs and time at the nodes, customs clearance costs and time, transportation distance, transportation costs, transportation speed, cargo handling capacity, penalty costs, cargo loss rate, and cargo value. The uncertain parameters include tariff rates, cargo transportation volume, and fuzzy demand for goods at the demand point.

[0010] Preferably, the decision variable is a 0-1 variable, used to represent the selection of transportation mode and the conversion of transportation mode.

[0011] Preferably, the total transportation cost includes the operating costs of domestic and international transit nodes, the transshipment costs of goods at the transit nodes, the transportation costs between nodes, penalty costs, customs duties, and customs clearance fees, as detailed below: The formula for calculating the operating costs of domestic and international transit nodes is as follows: in, , These represent the sets of domestic transit nodes and overseas transit nodes, respectively. , These represent domestic transit points and overseas transit points, respectively. , , , These are the operating costs of domestic and international transit nodes, respectively. The formula for calculating the transshipment cost of the goods at the transit node is as follows: in, For goods at the node By mode of transportation Change to Unit operating cost, For 0-1 decision variables, if the goods are... Nodes are determined by transportation methods Convert to ,but Otherwise , For goods at the node By mode of transportation Change to Unit operating cost, For 0-1 decision variables, if the goods are... Nodes are determined by transportation methods Convert to ,but Otherwise ; The formula for calculating the transportation cost between the nodes is as follows: in , , Transportation Goods in , , The transportation distance in each stage of transportation. As a supplier, This represents an index and set of overseas customer demand points. , , Transportation Goods in , , Unit transportation distance cost in each stage of transportation; The formula for calculating the penalty cost is as follows: in, This indicates the penalty cost incurred when a quality problem occurs. For goods by mode of transport exist The quantity of goods transported during the transportation phase. This refers to the loss rate of goods throughout the entire transportation process. The formula for calculating the tariff is as follows: in, For 0-1 decision variables, if the goods are transported by a certain method... From domestic transit points Delivery to overseas transit points , Otherwise , For the trade environment In the case of the first Domestic transit points transport to the first Tariff rates at overseas transit points Goods by mode of transport exist The quantity of goods transported during the transportation phase. The value of the goods; The formula for calculating the customs clearance fee is as follows: in, , For goods at the node ,node Customs clearance costs; In summary, the total transportation cost is expressed as follows: ; The total transportation time includes the time spent transporting the product from the supply point to the domestic transit point, then from the domestic transit point to the overseas transit point, and finally delivering it to the user's demand point, as well as the transshipment time of the goods at domestic and overseas transit nodes. The specific total transportation time is expressed as follows: in Representatives adopted transportation methods Goods from the supply point The time it takes to transport goods to the domestic transit point. Representatives adopted transportation methods Goods from domestic transit points Transport to overseas transit points Time, Representatives adopted transportation methods Transiting goods from overseas points The time required to transport the goods to the customer's desired location q.

[0012] Preferably, the objective function of the multi-objective optimization model is obtained by setting the transportation cost weight value as ω, which takes the value [0, 1], and normalizing the cost and time to the same dimension: This objective function can be expressed as: in Historical average cost or ideal cost value can be used. The historical average time or the shortest possible time can be used.

[0013] Preferably, the step of using a hybrid algorithm combining genetic algorithm and hill climbing method to solve the multi-objective optimization model to obtain the optimal transportation scheme specifically includes the following steps: Encoding Design: A hybrid encoding strategy is adopted, combining binary encoding and real number encoding. The first part of the chromosome is the binary part, representing the transportation mode, and the second part is the real number part, representing the route nodes. Population initialization: A set of chromosomes is randomly generated, and each gene is randomly selected from its corresponding range; Fitness function calculation: The fitness of each chromosome is evaluated based on the objective function, which is a comprehensive score of transportation cost and transportation time; Selection operation: A roulette wheel selection strategy is used to determine the individuals participating in crossbreeding; Crossover operation: Crossover operation is performed using a partial matching crossover method; Mutation operation: Randomly modify some genes on some chromosomes to increase the diversity of solutions; The hill-climbing method is used to perform a local search for new individuals: a neighbor solution is generated by a small perturbation. If the neighbor solution has higher fitness, it replaces the current solution. This operation is repeated until the maximum number of iterations is reached. Termination condition: The algorithm terminates when the number of iterations reaches the set number or the fitness change is less than a certain threshold.

[0014] A multimodal transport cross-border logistics route optimization system based on uncertain demand and tax rates, applied to the method described, includes: The network construction module is used to build a cross-border logistics network system, which includes supply points, domestic transit points, overseas transit points, and end-user demand points. The variable determination module is used to determine model assumptions and relevant variables, including parameters and decision variables, with parameters covering deterministic parameters and uncertain parameters; The model building module is used to build a multi-objective optimization model with the goal of minimizing transportation costs and transportation time. The solution module is used to solve the multi-objective optimization model using a hybrid algorithm of genetic algorithm and hill climbing method to obtain the optimal transportation scheme.

[0015] A computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method.

[0016] A processor for running a program, wherein the program executes the method during runtime.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a four-layer cross-border logistics network system, comprehensively considers the uncertainties of demand and tax rates, and establishes a multi-objective optimization model that can fully reflect the actual situation of cross-border logistics and improve the accuracy and practicality of route optimization.

[0018] 2. This invention uses a hybrid algorithm combining genetic algorithm and hill climbing method to solve the problem. It combines the global search capability of genetic algorithm with the local optimization capability of hill climbing method, which reduces the fluctuation of fitness function and improves convergence speed. It can quickly find the optimal transportation scheme and improve the solution efficiency and accuracy.

[0019] 3. The present invention can provide the optimal transportation solution according to different trade environments and demand preferences, which helps to improve logistics efficiency, reduce operating costs, and enhance the overall competitiveness of the regional economy. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0021] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a diagram of the cross-border trade logistics and distribution network for multimodal transport according to the present invention; Figure 3 This is a flowchart of the genetic-hill-climbing hybrid algorithm of the present invention; Figure 4 This is a schematic diagram of chromosome coding in this invention; Figure 5 This is an example diagram of the cross-operation process of the present invention; Figure 6 This is an example diagram of the variation operation process of the present invention; Figure 7 This is a visualization of the starting point, intermediate point, and demand point of this invention. Figure 8 This is the iterative graph of the genetic algorithm of this invention; Figure 9 This is a comparison chart of the GA and GA-HC iterations of this invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1 First, the relevant variables are explained. The relevant variables mentioned in this embodiment are explained as follows: 1) Parameters g: Represents a supply point described in this article; K1, K2: Represent the sets of domestic transit nodes and overseas transit nodes, respectively; i, j: Represent domestic transit points and overseas transit points, i∈K1, j∈K2; q, Q: Represent the index and set of overseas customer demand points; m, n: Different modes of transportation, (m, n)∈N, with values ​​of 1, 2, 3, m≠n (where 1 represents road transportation, 2 represents rail transportation, and 3 represents water transportation); s, S: Cross-border trade environment index and set, s∈S.

[0024] 2) Deterministic parameters , : Operating cost of the i-th domestic transit point / the j-th overseas transit point; The unit cost of changing the mode of transport from m to n for goods at node i; The cost required for customs clearance of goods at node i / node j; , , : The transportation distance of goods at each stage of transportation, using the transportation method m; , , : The unit transportation distance cost of goods at each stage of transportation using transportation method m; : The unit transshipment time of goods at node i when the mode of transport is changed from m to n ; : The unit transshipment time of goods at node j when the mode of transport is changed from m to n; : The time required for goods to clear customs at node i / node j; , , : Indicates the speed of goods transported by mode m in each corresponding transportation stage; : Represents the quantity of goods transported from supply point g to the i-th domestic transit point using transportation method m; The cargo handling capacity of domestic transit nodes; Cargo handling capacity of overseas transit nodes; The penalty costs incurred when quality problems occur; Loss rate of goods throughout the entire transportation process (referring to the occurrence of quality problems); The value of the goods; A large number.

[0025] 3) Uncertainty parameters : Tariff rates for goods transported from domestic transit point i to overseas transit point j under trade environment s; , , : The quantity of goods transported by mode of transport m at each stage of transport; : Fuzzy demand quantity of goods at demand point q.

[0026] 4) Decision variables : 0-1 decision variable, if the goods are transported by mode m from a domestic transit point Delivery to overseas transit points , Otherwise ; : 0-1 decision variable, if the goods are transported by mode m from Delivery from overseas transit points to customer demand points , ;otherwise ; : 0-1 decision variable, if the goods are changed from transportation mode m to n at node i, then Otherwise ; : 0-1 decision variable, if the goods are changed from transportation mode m to n at node j, then Otherwise .

[0027] Please see Figure 1 This embodiment provides a method for optimizing multimodal cross-border logistics routes based on uncertain demand and tax rates, including the following steps: S1. Construct a cross-border logistics network system, which includes supply points, domestic transit points, overseas transit points, and the ultimate user demand point; specifically as follows: This embodiment considers a four-layer cross-border logistics network system, including supply points, domestic transit points, overseas transit points, and end-user demand points. It addresses the complex scenario of one supplier, multiple domestic transit points, multiple overseas transit points, and multiple customer demand points coexisting, constructing a multi-objective optimization model to minimize transportation costs and time. The research focuses on determining the appropriate transportation scheme for exported products, starting from the supply point and selecting the appropriate mode of transport at each stage to ultimately reach the demand point, based on known objectives. Uncertain factors considered include tariffs and the uncertainty of demand within the network. A specific cross-border logistics network distribution diagram is shown below. Figure 2 As shown.

[0028] S2. Determine the model assumptions and relevant variables, including parameters and decision variables. The parameters include deterministic parameters and uncertain parameters. Establish a multi-objective optimization model with the objective of minimizing the total transportation cost and total transportation time. Specifically: The model makes the following assumptions: it considers cargo transportation from a single origin point, with multiple modes of transport coexisting; there is no transportation between domestic and overseas transit points, i.e., transportation is not considered within nodes of the same level; a domestic transit point can transport goods to multiple overseas transit points, and an overseas transit point can receive goods from multiple domestic transit points; an overseas transit point can transport goods to multiple demand points, and a demand point is served by only one overseas transit point; it is assumed that transshipment only occurs at transit nodes, and each order undergoes at most one change of mode of transport at a node; 40-foot containers are used during transportation; the cargo demand at each demand point is uncertain; and only the transportation and delivery of the same type of goods are considered.

[0029] The deterministic parameters include the operating costs of domestic and overseas transit points, transshipment costs and time at the nodes, customs clearance costs and time, transportation distance, transportation costs, transportation speed, cargo handling capacity, penalty costs, cargo loss rate, and cargo value; the uncertain parameters include tariff rates, cargo transportation volume, and fuzzy demand for goods at the demand points; the decision variables are 0-1 variables used to represent the choice of transportation mode and the conversion of transportation modes.

[0030] The total transportation cost consists of the operating costs of domestic and international transit nodes, the transshipment costs of goods at transit nodes, the transportation costs between nodes, penalty costs, and customs duties, which are described below using... express: This represents the total operating cost, which is the sum of the operating expenses of all domestic and overseas transit nodes. The operating costs of the domestic and overseas transit nodes are respectively... and ; This represents the total cost of transshipment of goods at domestic and international transit points, taking into account the costs incurred by changing the mode of transport at domestic and overseas transit points. This represents transportation costs, which include all transportation expenses from the point of supply to domestic transit points, from domestic transit points to overseas transit points, and from overseas transit points to the point of demand. This refers to the penalty cost incurred due to the loss of goods caused by quality issues throughout the entire transportation process, where the penalty imposed by the cross-border e-commerce platform on each unit of product is... ; This indicates customs duties, which need to be paid when goods are transferred from a domestic transit point to an overseas transit point. This refers to customs clearance fees. During the import and export of goods, certain fees are incurred to complete customs declarations, inspections, and tax collection procedures. In this text, these fees are referred to as customs clearance fees.

[0031] In summary, considering all the above costs, the total transportation cost can be expressed as: The total time includes the time spent transporting the product from the supply point to the domestic transit point, then from the domestic transit point to the overseas transit point, and finally delivering it to the user's demand point, as well as the transshipment time at domestic and overseas transit nodes; the total transportation time can be expressed as: in This represents the time required to transport goods from the supply point g to the domestic transit point using transportation method m. This represents the time required to transport goods from domestic transit point i to overseas transit point j using transportation method m. This represents the time it takes to transport goods from an overseas transit point j to the customer's desired location q using transportation method m. , , The value can be obtained by considering the transportation distance and speed at the corresponding stage of the transportation mode. The specific calculation method is as follows: , , .

[0032] S3. Construct the constraints for the multi-objective optimization model. The specific constraint formulas are as follows: (1) , (2) , (3) , (4) , (5) , (6) (7) Constraint formula (1) indicates that the amount of goods transported from supply point g to domestic transit point i shall not exceed the transit capacity of domestic transit point i; constraint formula (2) indicates that the amount of goods transported from domestic transit point i to overseas transit point j shall not exceed the transit capacity of overseas transit point j; constraint formula (3) indicates that after goods are damaged due to quality issues during the entire transportation process, the amount of goods transported from overseas transit point j to demand point q must meet the demand of that customer point, i.e., without considering the situation of shortage at the demand point; constraint formula (4) indicates that a customer demand point is transported by only one overseas transit point; constraint formula (5) indicates that an overseas transit point can receive goods transported by multiple domestic transit points, and a domestic transit point can transport goods to multiple overseas transit points; constraint formula (6) indicates that the amount of goods transported by transportation mode m in each stage is greater than or equal to the amount of goods transported in the next stage, indicating that there are situations where losses are caused by poor quality of goods during the transportation process. Constraint formula (7) is the range of values ​​for the variable.

[0033] It should be noted that this model has two optimization objectives: the transportation time of goods and their transportation cost. The importance of these two objectives varies in different scenarios, and their units of measurement also differ. Therefore, this paper sets the weight of the transportation cost as ω, with a value of [0, 1]. and transportation time Normalize to the same dimensions.

[0034] Therefore, the objective function can be expressed as: in Historical average cost or ideal cost value can be used. The historical average time or the shortest possible time can be used.

[0035] S4. A hybrid algorithm combining genetic algorithm and hill-climbing method is used to solve the multi-objective optimization model to obtain the optimal transportation scheme. The specific calculation process is as follows: Figure 3 As shown: S41. Encoding Design: A hybrid encoding strategy is adopted, combining binary and real-number encoding. The first part of the chromosome is the binary part representing the transportation mode, and the latter part is the real-number part representing the route nodes. Specifically: Genetic algorithms cannot directly process the data used to solve the problem; they need to be encoded to transform the data into data that can be processed by the genetic algorithm. The encoding strategy has a significant impact on the execution of genetic operations such as crossover and mutation in genetic algorithms, and largely determines the efficiency of the genetic evolution process. Common encoding methods for genetic algorithms include binary encoding, Gray code, symbolic encoding, permutation encoding, and real number encoding. Among these methods, binary encoding and real number encoding are more commonly used. This paper considers the path transportation selection problem involving multiple transportation modes. Therefore, this paper adopts a hybrid encoding strategy, combining binary encoding and real number encoding to handle the selection of transportation modes and nodes.

[0036] In this model, the structure of a chromosome represents the logistics path of goods from the supply point to the end user using appropriate transportation methods. It is encoded by the transportation method and the number of nodes. Each chromosome can be divided into two parts: the first part is a binary part representing the transportation method (road, rail, waterway), encoded as 1, 2, and 3; the second part is a real number part representing the route nodes, categorized as origin point, domestic transit point, overseas transit point, and demand point. Considering the actual transportation situation of the nodes, it is stipulated that each node supports the three transportation methods given in the text. The encoding method is shown in Table 1.

[0037] Table 1 Chromosome coding methods The final chromosome structure is shown in the example. Example: [0001 0010 0011 | 1, 3, 7, 25]. The chromosome coding diagram is as follows: Figure 4 As shown.

[0038] S42. Population Initialization: Randomly generate a set of chromosomes, and randomly select each gene from its corresponding range, specifically as follows: A set of chromosomes is randomly generated. Each gene is randomly selected from its corresponding range. The binary part of the gene value can be 1, 2, or 3, while the real part of the gene value can be set according to the specific number of domestic and overseas transit points. The specific steps are as follows: For each supply point, a domestic transit point is randomly selected from 0 to K1; for each domestic transit point, an overseas transit point is randomly selected from 0 to K2; for each overseas transit point, an end user is randomly selected from 0 to Q.

[0039] S43. Fitness Function Calculation: The fitness of each chromosome is evaluated based on the objective function. The fitness function is a comprehensive score of transportation cost and transportation time, specifically: The fitness of each chromosome x is evaluated based on the objective function. In this paper, the fitness function is a combined score of total transportation cost and total transportation time, which can be set as follows: , In this embodiment, the objective problem is minimized, and the fitness function is calculated by normalizing the total cost and total time.

[0040] S44. Selection Operation: A roulette wheel selection strategy is used to determine the individuals participating in crossbreeding, specifically: The selection operation determines which individuals will act as "parents" and participate in crossbreeding to produce the next generation. This process reflects the principle of natural selection, where individuals with higher fitness have a greater probability of being selected. In this paper, a roulette wheel selection strategy is used, randomly selecting a number between 0 and the total fitness. The fitness value of each individual is iterated and gradually increased until the accumulated fitness value exceeds the randomly selected number; at this point, that individual is selected as a parent. The selected individual is used for reproduction. If the selection value is not reached after iterating through all individuals, the last individual in the population is returned. The probability calculation for this selection is shown in the formula: S45. Mutation Operation: Randomly modify some genes on some chromosomes to increase the diversity of solutions. Specifically: Crossover generates two offspring individuals from two parent individuals. It simulates the biological reproduction process by combining the genes of the parents (transportation methods, selection of transit nodes) to create new individuals. Since different positions in the encoding process represent different functions, using full-fragment crossover is prone to computational errors; therefore, partial matching crossover is used in this paper. Due to the limitation on the number of starting points, the value of this gene position is always 1. Assuming we have two parent chromosomes, the specific crossover operation procedure is as follows: Figure 6 As shown.

[0041] S46. Local search for new individuals using the hill-climbing method: Neighboring solutions are generated through small perturbations. If a neighboring solution has higher fitness, it replaces the current solution. This process is repeated until the maximum number of iterations is reached. Specifically: By defining a local perturbation range or neighborhood generation rule, several "neighbor solutions" are generated by slightly perturbing certain genes (variables) of x. The fitness of the new solution x' is calculated. If the fitness of the new solution is greater than the original fitness value, i.e., the fitness is improved, x' is accepted to replace the current solution x. The neighborhood perturbation is then performed on the new x, and the above steps are repeated until the maximum number of iterations is reached.

[0042] S47. Termination condition: The algorithm terminates when the number of iterations reaches the set number or the fitness change is less than a certain threshold.

[0043] Example 2 A multimodal transport cross-border logistics route optimization system based on uncertain demand and tax rates, applied to the method described, includes: The network construction module is used to build a cross-border logistics network system, which includes supply points, domestic transit points, overseas transit points, and end-user demand points. The variable determination module is used to determine model assumptions and relevant variables, including parameters and decision variables, with parameters covering deterministic parameters and uncertain parameters; The model building module is used to build a multi-objective optimization model with the goal of minimizing transportation costs and transportation time. The solution module is used to solve the multi-objective optimization model using a hybrid algorithm of genetic algorithm and hill climbing method to obtain the optimal transportation scheme.

[0044] Example 3 A computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method.

[0045] Example 4 A processor for running a program, wherein the program executes the method during runtime. Furthermore, to better understand the present invention, specific examples are provided below for further illustration: 1. Case Background The development of modern technology and the internet has driven changes in consumer lifestyles, leading to an increasing reliance on convenience foods. Simultaneously, influenced by livestream e-commerce, snail rice noodles have become a major category of fast food. According to the "2023 China Snail Rice Noodle Industry Development Status and Consumer Behavior Analysis Report" released by FoodTalks Global Food Information Network, the total sales revenue of the Liuzhou snail rice noodle industry chain reached 60.07 billion yuan in 2022, a year-on-year increase of 19.8%, and is expected to reach 88.99 billion yuan by 2025; among which, the sales revenue of Liuzhou packaged snail rice noodles reached 18.18 billion yuan in 2022, a year-on-year increase of 19.6%. Snail rice noodle exports are booming, with Liuzhou snail rice noodle exports reaching 83 million yuan in 2022, a year-on-year increase of 61%.

[0046] 2. Instance Data Taking a food company that sells packaged instant food products to Southeast Asia as an example, the company obtained basic export information for its packaged instant food products from e-commerce websites. The visualization results of the coordinates of domestic and overseas transit points along the route of goods shipped from Nanning to various demand points in Southeast Asia are shown below. Figure 7 As shown: Let the penalty cost e be 10 yuan / piece, and the loss rate of goods due to quality issues during the entire transportation process be... The assumption is 0.9%. Basic information on domestic and overseas transit points was obtained from existing literature and Google Maps. The coordinates of the starting point, Nanning, are (108.366, 22.818). Other node information is shown in Table 2-3 below.

[0047] Table 2 Basic Information on Domestic Transit Points Table 3 Basic Information on Overseas Transit Points Since the demand at the customer point is difficult to determine, this paper adopts a random generation method. In the process of my country's foreign trade, its trade environment has great uncertainty, which is directly reflected in the change of tariffs. Therefore, this paper considers three situations: normal (s=1), good (s=2) and poor (s=3) trade environment, and considers the different demand and tariffs under each scenario. According to the literature

[18] , the domestic and foreign transportation costs and transportation speeds of road, rail and waterway transportation, the costs and time of transshipment and customs clearance, and the transit time and transit costs between different modes of transportation. [18,19] Specific data is shown in Table 4-6 below. In the following text, FEU refers to a 40-foot container. When calculating the target cost, the US dollars in the transportation cost unit will be converted to yuan using the exchange rate of April 2025, i.e., 1 US dollar = 7.2 yuan.

[0048] Table 4 Parameters for different modes of transportation Table 5. Time and Costs for Transit Using Different Modes of Transportation Table 6. Costs and time incurred for transshipment and customs clearance between different modes of transport. The distance between supply points, domestic transit points, overseas transit points, and finally demand points is calculated using the latitude and longitude coordinates of each node. Because the actual transport distance between nodes differs from the distance calculated using latitude and longitude coordinates when different modes of transport are used, the distance calculated using the node's latitude and longitude coordinates is recorded as the road transport distance between nodes. A coefficient is introduced when calculating the railway and waterway transport distances between nodes. and The distance between nodes is denoted by railway transport distance. The distance between nodes is denoted by waterway transportation. ,in Additionally, when no transportation mode m (m=1, 2, 3) exists between two nodes, the transportation distance using transportation mode m is recorded as +∞. The transportation distance from the overseas transit point to the demand point is also calculated using the node coordinates. The transportation time between nodes at each level is obtained through the transportation distance between nodes and the transportation speed of the corresponding transportation mode.

[0049] Since the demand for this fast food product at each customer demand point is difficult to measure, this paper uses a method of randomly generating the demand, setting the random generation range to [6000, 10000]. The coordinates of the customer demand points were also obtained from Google Maps.

[0050] Table 7 shows the tariffs on the company's products exported to other countries from its supply points, as found through the international trade tax website. Table 7. Tariffs on Chinese exports to some Southeast Asian countries when trade scenario s=1 The above shows the demand and tariffs under the trade scenario s=1. Due to the constantly changing international trade environment and the many uncertainties facing global economic and trade development, some factors considered in this paper may change accordingly in this evolving trade environment. Specifically, when the trade environment is favorable (s=2) or unfavorable (s=3), the quantity demanded increases by 10% and decreases by 10% respectively compared to the trade environment s=1; regarding tariffs, they decrease and increase by 50% respectively compared to the trade environment s=1.

[0051] 3. Results Analysis This paper uses Python 3.11 to solve the model established above. The population size is set to 100, the maximum number of iterations to 200, and the mutation probability to 0.1. The time and cost weights in the cross-border trade process are each set to 0.5. Under normal trade conditions (s=1), the generated genetic algorithm iterative graph is as follows. Figure 8 As shown: It can be clearly seen from the iteration graph that the process graph shows an overall downward trend, continuously optimizing until it gradually stabilizes into a straight line. After 65 iterations, it becomes a straight line, indicating that the optimal solution has been found. Finally, the optimized transportation scheme is output. The optimal transportation path node selection of this four-layer network is shown in Table 8.

[0052] Table 8 Service status of each node under normal trade environment conditions The table above shows the routes for goods to be delivered to the demand points. The specific routes and transportation methods for the top ten demand points are shown below: Nanning (water) - 2 Fangchenggang (water) - 8 Yangon (rail) - 1 Bangkok. The total cost of this route is 473,467.58 yuan, and the total time is 81.87 hours.

[0053] The route from Nanning (rail) to Fangchenggang (rail) to Mandalay (road) to Phitsanulok has a total cost of 549,772.22 yuan and a total travel time of 30.86 hours.

[0054] Nanning (rail) - 4 Kunming (water) - 4 Chiang Mai (water) - 3 Pattaya, the total cost of this route is 536,037.52 yuan, and the total time is 78.87 hours.

[0055] The route from Nanning (rail) to Pingxiang (water) to Yangon (road) to Hat Yai has a total cost of 638,043.52 yuan and a total travel time of 79.43 hours.

[0056] Nanning (road) - 4 Kunming (road) - 10 Vientiane (rail) - 5 Chiang Rai. The total cost of this route is 535,474.38 yuan, and the total time is 30.45 hours.

[0057] Nanning (water) - Kunming (water) - Vientiane (public) - Lampang, the total cost of this route is 311,314.24 yuan, and the total time is 65.13 hours.

[0058] The route from Nanning (public) to Fangchenggang (public) to Haiphong (public) to Kongjing has a total cost of 335,031.29 yuan and a total travel time of 14.82 hours.

[0059] The route from Nanning (public) to Pingxiang (water) to Yangon (water) to Ayutthaya has a total cost of 384,155.08 yuan and a total travel time of 85.26 hours.

[0060] The route from Nanning (rail) to Qinzhou (water) to Sihanoukville (rail) to Hua Hin has a total cost of 393,792.45 yuan and a total travel time of 77.52 hours.

[0061] Nanning (road) - 4 Kunming (road) - 2 Da Nang (road) - 10 Luang Prabang. The total cost of this route is 383,212.63 yuan, and the total time is 32.42 hours.

[0062] Through continuous iterative optimization, the fitness value was significantly improved, and the results converged quickly during the optimization process. Based on the final results, under normal trade conditions and with a cost weight w=0.5, the average total cost of all transportation routes for a certain food company was 300,395.78 yuan, and the average total time for all routes was 65.67 hours.

[0063] The above uses Python 3.11 programming to solve the multimodal transport route selection model using a strategy combining genetic methods and hill climbing methods, obtaining the optimal transportation schemes for different demand points, as well as the total cost and total time spent by all optimal transportation schemes. Figure 8 To compare the genetic iteration graphs obtained by solving the model using the traditional genetic algorithm and the GA-HC algorithm in this paper. Depend on Figure 8 The comparison shows that the combined genetic algorithm and hill-climbing method converges in about 65 iterations, while the traditional genetic algorithm converges in about 170 iterations. Furthermore, the GA-HC algorithm achieves a smaller final fitness value than the GA algorithm. Therefore, the solution strategy combining genetic algorithm and hill-climbing method proposed in this paper converges faster and achieves better results, making it more suitable for solving the established model.

[0064] 3. Parameter sensitivity analysis Since the optimization objectives in this model are cost and time, the weights of cost and time in the model are changed. w is set to 0.1, 0.3, 0.7, and 0.9 respectively. Under normal trade conditions, the optimal route selection schemes for each demand point are shown in Table 9. The total cost and total time corresponding to the optimal transportation scheme under different weights are shown in Table 9.

[0065] Table 9. Average total cost and total time of the optimal transportation plan for the demand point under different weight values. The results in the table above show that, keeping other parameters constant, the optimal route selection for each demand point differs depending on the weighting of cost and time—that is, when the demand points' preferences for cost and time during the transportation process change. Simultaneously, the average cost and average time spent on transportation schemes for all demand points in the network also change. As the cost weighting increases, meaning demand point customers place greater emphasis on timeliness during transportation, the cost of goods transportation increases while the time decreases. Conversely, when demand points have limited budgets, they tend to prefer more economical transportation schemes. In this case, the importance of transportation time decreases, and the overall transportation scheme exhibits characteristics of lower cost and higher time.

[0066] The weights of cost and time in the objective function are set to 0.5, while other parameters remain unchanged. Table 10 shows the optimal transportation routes, total costs, and total time incurred at each demand point under different trade environments, with varying tariff rates (increased or decreased).

[0067] Table 10 Impact of Tariff Reductions and Increases on Transportation Options As shown in the table above, when faced with changes in the trade environment, this paper uses tariff changes as a case study, reflecting the favorable or unfavorable trade environment by raising or lowering tariffs. The final transportation plan results reveal that changes in the trade environment have a significant impact on transportation costs. When tariffs increase, the average total cost across all routes increases relative to the value under normal trade conditions, and vice versa. However, the impact on transportation time is not significant. The results show that regardless of whether tariffs increase or decrease, the average time spent in all optimal transportation plans decreases relative to the value under normal trade conditions.

[0068] In summary, this paper explores the optimization problem of multimodal cross-border logistics networks under the combined influence of multiple factors. A multi-objective optimization model is constructed, and a strategy-solving algorithm combining genetic algorithm and hill climbing method is designed to solve the model. This algorithm is applied to a cross-border e-commerce trade scenario of a food technology company targeting Southeast Asian countries. Based on collected data, the case model is solved, ultimately yielding the optimal transportation scheme for different demand points. When the emphasis on cost and time during the cargo transportation process changes for different demand points, i.e., by adjusting the weight ratio of cost and time in the objective function, the optimal transportation scheme corresponding to the demand point is obtained, revealing that the transportation mode changes with the weight. By comparing the solution algorithm presented in this paper with the traditional genetic algorithm, it is found that the algorithm used in this paper outperforms the traditional genetic algorithm in terms of transportation cost and number of iterations, demonstrating the effectiveness of the algorithm in solving this problem.

Claims

1. A method for optimizing a multimodal cross-border logistics path based on demand and tax rate uncertainty, characterized in that, The application relates to a method for optimizing cross-border logistics, and belongs to the field of logistics optimization. The method comprises the following steps: Building a network system for cross-border logistics, which comprises a supply point, a domestic transfer point, a foreign transfer point and a final user demand point; Determining model assumptions and related variables, wherein the related variables comprise parameters and decision variables, the parameters comprise deterministic parameters and uncertain parameters, and a multi-objective optimization model is established to minimize total transportation cost and total transportation time; Building constraint conditions of the multi-objective optimization model; 2.The method of claim 1, wherein, Solving the multi-objective optimization model by using a genetic algorithm and a hill climbing method to obtain an optimal transportation scheme. The model assumptions comprise the following aspects: 3.The method of claim 1, wherein, Considering the case that a starting point of goods transportation coexists with multiple transportation modes, and there is no transportation between each domestic transfer point and each overseas transfer point; one domestic transfer point can transport goods to multiple overseas transfer points, and one overseas transfer point can receive goods transported by multiple domestic transfer points; one overseas transfer point can transport goods to multiple demand points, and one demand point is only served by one overseas transfer point; transfer only occurs at transfer nodes, and each order at the node at most changes one transportation mode; the demand of goods at each demand point is uncertain; and only the transportation and distribution of the same type of goods are considered. The deterministic parameters comprise the operation cost of the domestic transfer point and the overseas transfer point, the rehandling cost and time of goods at the node, the customs clearance cost and time, the transportation distance, the transportation cost, the transportation speed, the goods handling capacity, the penalty cost, the goods loss rate and the goods value. 4.The method of claim 1, wherein, The uncertain parameters comprise the tariff rate, the goods transportation volume and the fuzzy demand of goods at the demand point.

5. The method of claim 3, wherein, The decision variables are 0-1 variables, which are used for representing the selection of a transportation mode and the conversion of a transportation mode. The total transportation cost comprises the operation cost of the domestic and overseas transfer nodes, the rehandling cost of goods at the transfer nodes, the transportation cost between the nodes, the penalty cost, the customs duty and the customs clearance cost, and the total transportation cost is specifically as follows: wherein, , respectively represent a set of domestic transit nodes, overseas transit nodes, , respectively represent a domestic transit node and an overseas transit node, , , , respectively represent the operating costs of domestic and overseas transit nodes; The calculation formula of the operation cost of the domestic and overseas transfer nodes is as follows: in, For goods at the node By mode of transportation Change to Unit operating cost, For 0-1 decision variables, if the goods are... Nodes are determined by transportation methods Convert to ,but Otherwise , For goods at the node By mode of transportation Change to Unit operating cost, For 0-1 decision variables, if the goods are... Nodes are determined by transportation methods Convert to ,but Otherwise ; The calculation formula of the rehandling cost of goods at the transfer nodes is as follows: wherein , , using a transportation mode the goods are in , , the transportation distance of the segment transportation stage, the supply point, representing the index and set of overseas customer demand points, , , using a transportation mode the goods are in , , the unit transportation distance cost of the segment transportation stage; The calculation formula of the transportation cost between the nodes is as follows: wherein, represents a penalty cost generated when a quality problem occurs, is the quantity of goods transported by the transportation mode In the quantity of goods transported during the transportation phase, is the loss rate of goods during the entire transportation phase. The calculation formula of the penalty cost is as follows: wherein is a 0-1 decision variable if the shipment is transported by mode from the domestic transit point to the overseas transit point , ; otherwise , is the trade environment in the case of a shipment transported from the domestic transit point to the overseas transit point, is the customs duty rate for a shipment transported by mode in the transportation phase, is the value of the shipment; The calculation formula of the customs duty is as follows: wherein, , the cost of customs clearance required for the goods at the node , the node . The calculation formula of the customs clearance cost is as follows: ； The total transportation cost is represented as follows: in Representatives adopted transportation methods Goods from the supply point The time it takes to transport goods to the domestic transit point. Representatives adopted transportation methods Goods from domestic transit points Transport to overseas transit points Time, Representatives adopted transportation methods Transiting goods from overseas points The time required to transport the goods to the customer's desired location q.

6. The method of claim 1, wherein, The total transportation time comprises the time consumed by the process that products are transported from a supply point to a domestic transfer point, then transported from the domestic transfer point to an overseas transfer point and finally distributed to a user demand point, and the transfer and rehandling time of goods at the domestic transfer node and the overseas transfer node, and the total transportation time is specifically represented as follows: The objective function of the multi-objective optimization model is obtained by setting the transportation cost weight value as omega, and the value of omega is [0, 1], and the cost and time are normalized to the same dimension: wherein a desirable historical average cost or ideal cost value, a desirable historical average time or minimum possible time.

7. The method of claim 6, wherein, The objective function can be represented as follows: The genetic algorithm and the hill climbing method are used to solve the multi-objective optimization model to obtain an optimal transportation scheme, and the method specifically comprises the following steps: Coding design: a hybrid coding strategy is adopted, combining binary coding and real number coding. The front part of the chromosome is a binary part representing the transportation mode, and the rear part is a real number part representing the route nodes. Population initialization: a set of chromosomes is randomly generated, and each gene is randomly selected within its corresponding range. Fitness function calculation: the fitness of each chromosome is evaluated according to the objective function, which is a comprehensive score of transportation cost and transportation time. Selection operation: the roulette wheel selection strategy is used to determine the individuals participating in cross breeding. Crossover operation: partial matching crossover method is used for crossover operation. Mutation operation: some genes of some chromosomes are randomly modified to increase the diversity of solutions. Local search of new individuals using hill climbing method: neighbor solutions are generated by small perturbations. If the fitness of the neighbor solution is higher, the current solution is replaced. The operation is repeated until the maximum number of iterations is reached. Termination condition: when the number of iterations reaches a certain number or the fitness changes less than a certain threshold, the algorithm terminates.

8. A demand and tax rate uncertainty based intermodal cross-border logistics path optimization system, characterized in that, Applied to the method of any one of claims 1-7, comprising: a network construction module for constructing a cross-border logistics network system, which includes supply points, domestic transfer points, foreign transfer points and end-user demand points; a variable determination module for determining model assumptions and related variables, including parameters and decision variables, and parameters covering deterministic parameters and uncertain parameters; a model establishment module for establishing a multi-objective optimization model to minimize transportation cost and transportation time; a solution module for solving the multi-objective optimization model using a hybrid algorithm of genetic algorithm and hill climbing method to obtain the optimal transportation scheme.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium includes a stored program, wherein when the program runs, the device where the computer readable storage medium is located executes the method of any one of claims 1-7.

10. A processor, comprising: The processor is used to run the program, wherein the program runs to execute the method of any one of claims 1-7.

Images