Cross-border transportation intercontinental freight flow prediction analysis method

Abstract

The invention relates to the technical field of cross-border freight, and discloses a cross-border transportation intercontinental freight flow prediction analysis method, which comprises the following steps of: firstly, constructing a multimodal transport Chinese-European freight network structure, particularly a multi-mode multi-cargo-type cargo transportation network, based on a path topological structure diagram; the freight traffic congestion time is analyzed for freight traffic congestion bottleneck facilities in the path topology network, namely ports, Suishi canals and rail switching stations; the transportation time multiplied by the time value is converted into corresponding currency cost, and the transportation cost is calculated; and analyzing and calculating spatial distribution data of the Chinese and European cargo flow. The congestion caused by the bottleneck and the transport capacity upper limit of the transport service are considered at the same time. And an accurate and practical calculation tool is provided for predicting and analyzing the spatial distribution of the Chinese and European cargo flow and predicting the freight flow.

Description

Technical Field

[0001] This invention relates to the field of cross-border freight technology, specifically a method for predicting and analyzing intercontinental freight traffic in cross-border transportation. Background Technology

[0002] Compared to traditional China-Europe liner shipping services, the biggest advantage of the China-Europe Railway Express lies in its shorter transit time. However, because container trains on the China-Europe Railway Express need to traverse railway systems with different gauges, they must perform gauge-changing operations at stations, which leads to significant delays and severely impacts the timeliness of the service. For example, at Malaszewicze station, one of the most important gateways for westbound trains, delays have reportedly reached 4-6 days due to the current infrastructure's inability to handle the surge in China-Europe rail freight volume (ZZIA, 2021). On the other hand, each port has limited container ship service capacity, and when the number of arriving ships approaches or exceeds the port's capacity, delays due to queuing also occur. In addition to port congestion, the Suez Canal, with its large volume of vessels and limited throughput, has become a long-standing bottleneck on the China-Europe liner shipping route. Despite an expansion in 2015, 41.4% of the Suez Canal still only allows one-way traffic (SCA, 2021). Therefore, ships at both ends of the Suez Canal must be arranged to pass separately, and there are limits on the number of ships in each direction to avoid collisions between ships from different directions on a one-way segment. In summary, the bottleneck effect caused by the limited service capacity of the three types of infrastructure mentioned above will generate congestion and delays, which are key factors affecting the transit time of the entire transportation service route. Therefore, it is essential to consider and quantify the congestion caused by these bottlenecks.

[0003] Compared to traditional China-Europe liner shipping services, the biggest advantage of the China-Europe Railway Express lies in its shorter transit time. However, because container trains on the China-Europe Railway Express need to traverse railway systems with different gauges, they must perform gauge-changing operations at stations, which leads to significant delays and severely impacts the timeliness of the service. For example, at Malaszewicze station, one of the most important gateways for westbound trains, delays have reportedly reached 4-6 days due to the inability of current infrastructure to handle the surge in China-Europe rail freight volume (ZZIA, 2021). On the other hand, each port has limited container ship service capacity. When the number of arriving ships approaches or exceeds the port's capacity, delays due to queuing also occur. For instance, an IHS Markit (2021) report stated that factors such as port container backlogs, equipment and labor shortages caused by the COVID-19 pandemic resulted in Qingdao Port in China needing an average of over 50 hours to serve more than 6,000 ships, while Felixstowe Port in the UK needed an average of over 92 hours. Besides port congestion, the Suez Canal, with its large volume of vessels and limited throughput, has long been a bottleneck segment on the China-Europe liner shipping route. Despite an expansion in 2015, 41.4% of the Suez Canal still only allows one-way traffic (SCA, 2021). Therefore, vessels at both ends of the Suez Canal must be arranged to pass separately, and there are limits on the number of vessels in each direction to avoid collisions between vessels from different directions on one-way segments. In summary, the bottleneck effect caused by the limited service capacity of these three types of infrastructure will generate congestion and delays, which are key factors affecting the transit time of the entire transport service route. Therefore, it is essential to consider and quantify the congestion caused by these bottlenecks. This invention uses a series of queuing theory-based models to estimate the transit time of seaports, the Suez Canal, and rail-changing stations, and to evaluate the navigation performance of these bottleneck nodes.

[0004] Therefore, to address the above issues, a cross-border transportation intercontinental freight flow prediction and analysis method is needed to analyze and predict freight flow based on congestion and cargo transfer information at various transportation nodes, and to address the problem of low analysis efficiency. Summary of the Invention

[0005] The purpose of this invention is to provide a method for predicting and analyzing intercontinental freight traffic in cross-border transportation. This invention simultaneously considers congestion caused by bottlenecks and the capacity limits of transportation services. It provides an accurate and practical computational tool for predicting and analyzing the spatial distribution of freight flows between China and Europe and for forecasting freight traffic.

[0006] This invention is implemented as follows:

[0007] This invention provides a method for predicting and analyzing intercontinental freight traffic in cross-border transportation, specifically implemented according to the following steps:

[0008] S1: First, construct a multimodal freight network between China and Europe based on the route topology diagram, specifically a multi-mode, multi-cargo type freight transport network;

[0009] In step S1, in the path topology diagram, the starting point and the ending point are respectively represented by OD, which are domestic cities producing goods and foreign cities consuming goods.

[0010] The path topology network is graph G(N,A), where N represents the set of nodes and A represents the set of road segments. The set of nodes includes four types of nodes: O is the set of origin cities for goods; D is the set of destination cities for goods; Φ is the set of inbound and outbound systems for ports, canals (i.e., the Suez Canal), rail transfer stations, and railway stations; the set of road segments consists of highway segments, shipping segments, railway segments, port segments, canal segments, and rail transfer station segments.

[0011] Use L s This represents a collection of liner shipping routes, where each service is provided by a liner shipping company and consists of at least one sea segment; L r This represents a collection of railway transport service routes, each provided by the China-Europe Railway Express Company and consisting of at least one railway transport segment. Ω s Indicates a shipping segment, representing a port pair segment consisting of consecutive shipping segments within each liner service route; Ω r It indicates a railway transportation section.

[0012] S2: Analyze freight traffic congestion time for bottleneck facilities (i.e., ports, Suez Canal, and rail transfer stations) in the route topology network; use the M / M / c queuing model to characterize port congestion; specifically, follow these steps:

[0013] S2.1: Assuming that the arrival time of customers follows a Poisson distribution, there are c service counters in the system, and the service time follows an exponential distribution. For the port system, the container ship is the customer in the queuing theory model, and the container ship berth is the service counter. Then the average stay time of the container ship in the port is as shown in equation (1).

[0014]

[0015] Among them, t p λ represents the average time a container ship spends in a port. p It is the average arrival rate of container ships, μ p c represents the average service rate of container ship berths. pIt is the number of berths in the port, ρ p =λ p / c p μ p This indicates the average utilization rate of the port.

[0016] S2.2: The port's designed annual throughput is defined as the port's maximum handling capacity. Therefore, the port's utilization rate is the sum of China-Europe container freight volume and non-China-Europe container freight volume divided by the port's maximum handling capacity. By defining different utilization rates, the average service delay of the port was evaluated according to equation (1), as shown in equation (2).

[0017]

[0018] Where, x p It is the China-Europe container flow rate at port p. It is a pre-determined non-China-Europe container flow rate, u p Let p be the maximum throughput capacity of port p. β p , These are the parameters to be calibrated.

[0019] Furthermore, an M / M / 1 queuing theory model is constructed to assess congestion delays of the fleet when passing through the Suez Canal, where the ships are considered customers and the port of Tufic is considered a service counter; as shown in equations (3)-(4);

[0020]

[0021] Among them, t w Indicates the transit time of a ship sailing north through the Suez Canal, μ w It is the draw service rate of Port Tufiq, λ w It is the arrival rate of ships sailing north, τ w This refers to the transit time through the Suez Canal without queuing;

[0022]

[0023] Where x w To improve the flow rate of container cargo between China and Europe via the Suez Canal. For a predetermined non-Central Europe northbound container cargo flow rate, γ w To convert container cargo volume into a range factor for the number of container ships. It is a predetermined number of non-container ships, u w For the northbound ship throughput capacity of the Suez Canal, β w , These are the parameters to be calibrated.

[0024] Furthermore, the M / M / c queuing theory model is adopted to capture the delay caused by container transfer at the track-changing station; accordingly, the average delay time of container trains at the track-changing station is as shown in equation (5).

[0025]

[0026] Among them, t b λ is the average delay time of container trains at the track-changing station. b It is the arrival rate of container trains, μ b It is the average service rate of the track-changing station, c b It is the number of container transshipment terminal equipment, ρ b It is the average utilization rate of the track-changing station, ρ b =λ b / c b μ b .

[0027] The average service time of the station is evaluated by defining different utilization rates, as shown in Equation (6);

[0028]

[0029] Where, x b It is the China-Europe container flow rate at the rail transfer station. It is a pre-determined non-China-Europe container flow rate, u b This is the maximum throughput capacity of the track-changing station. β b , These are the parameters to be calibrated.

[0030] S3: Multiply the transportation time by the time value to convert it into the corresponding monetary cost and calculate the transportation cost;

[0031] Furthermore, in step S3, the transportation time is multiplied by the time value and converted into the corresponding monetary cost to calculate the generalized transportation cost. The generalized cost of transporting a unit of TEU of m-type goods through route k is calculated as shown in equations (7)-(8).

[0032]

[0033] Equation (8) guarantees that, under the optimal freight flow allocation model, the delay time caused by the upper limit of transport capacity is minimized. It will only appear on routes that are already at full capacity.

[0034] Furthermore, with The perceived generalized transportation cost of goods of class m on path k connecting od is expressed as: in To determine the generalized transportation cost of the actual path under a given traffic allocation pattern, and This represents the stochastic term in perceived cost, indicating the perceived error in a shipper's route selection behavior due to numerous invisible influencing factors. The logit model characterizes the shipper's pattern-route selection behavior, and the stochastic term in the perceived generalized cost of the route represents the cost. If the route follows a Gumbel distribution, then the probability that the shipper chooses this route is... It is obtained by calculation using equation (9);

[0035]

[0036] Among them, positive numbers It is the range factor related to the variance of the perceived generalized transportation cost of the shipper of m-type goods. The model of the multi-mode multi-cargo category network equilibrium problem is as shown in Equation (10).

[0037]

[0038] in,

[0039]

[0040]

[0041] S4: Analyze and calculate the spatial distribution data of China-Europe freight flows.

[0042] Furthermore, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a cross-border transportation intercontinental freight flow prediction and analysis method as described in any one of the above descriptions.

[0043] Furthermore, the present invention provides a computer-readable storage medium, the computer-readable storage medium including an embedded processing system and a stored program, wherein the embedded system controls the execution of any one of the above-described methods for predicting and analyzing intercontinental freight traffic in cross-border transportation during the operation of the embedded system control program.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] 1. This invention simultaneously considers congestion caused by bottlenecks and the capacity limits of transportation services. It provides an accurate and practical calculation tool for predicting and analyzing the spatial distribution of freight flows between China and Europe and forecasting freight volume.

[0046] 2. The proposed network equilibrium model considering capacity constraints, solved using the subgradient algorithm within the Lagrange relaxation framework, is entirely feasible. Within this framework, the DSD algorithm demonstrates high efficiency and performance in solving large-scale network equilibrium problems using the relaxed Lagrange problem. During the algorithm's iteration process, the number of iterations required to reach equilibrium for each relaxed Lagrange problem gradually decreases, significantly shortening the overall solution time for the entire spatial information. For planning or strategic decision-making objectives, the accuracy, solution efficiency, and effectiveness of this invention greatly improve work efficiency. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart of the method of the present invention;

[0049] Figure 2 This is a schematic diagram of the node-path topology of the intermodal freight network between China and Europe according to the present invention;

[0050] Figure 3 This invention relates to the functional relationship between the average dwell time of containers in Shanghai Port and their container throughput rate.

[0051] Figure 4 This invention relates to the Suez Canal layout;

[0052] Figure 5 This invention relates to the functional relationship between the average transit time of the Suez Canal and the throughput of northbound vessels.

[0053] Figure 6 This invention relates to the functional relationship between container flow rate and average dwell time at the Mavarsheviche station.

[0054] Figure 7 This is an example of the convergence curve of the subgradient method of the present invention for solving the Lagrange multiplier problem, showing a portion of the month.

[0055] Figure 8 This is an example of the convergence curve of the DSD algorithm of this invention for solving the relaxed Lagrange problem, specifically for the month portion.

[0056] Figure 9 This is a schematic diagram of the convergence curve of the complete algorithm of this invention in December 2019. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 a part of the embodiments of the present invention, not all of them. 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. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to describe selected 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.

[0058] Please see Figures 1-9 This invention provides a method for predicting and analyzing intercontinental freight traffic in cross-border transportation, specifically implemented according to the following steps:

[0059] S1: First, construct a multimodal freight network between China and Europe based on the route topology diagram, specifically a multi-mode, multi-cargo type freight transport network;

[0060] In step S1, in the path topology diagram, the starting point and the ending point are respectively represented by OD, which are domestic cities producing goods and foreign cities consuming goods.

[0061] The path topology network is graph G(N,A), where N represents the set of nodes and A represents the set of road segments. The set of nodes includes four types of nodes: O is the set of origin cities for goods; D is the set of destination cities for goods; Φ is the set of inbound and outbound systems for ports, canals (i.e., the Suez Canal), rail transfer stations, and railway stations; the set of road segments consists of highway segments, shipping segments, railway segments, port segments, canal segments, and rail transfer station segments.

[0062] Use L s This represents a collection of liner shipping routes, where each service is provided by a liner shipping company and consists of at least one sea segment; L r This represents a collection of railway transport service routes, each provided by the China-Europe Railway Express Company and consisting of at least one railway transport segment. Ω s Indicates a shipping segment, representing a port pair segment consisting of consecutive shipping segments within each liner service route; Ω rThis represents a railway transport segment. The OD demand is exogenous and includes M types of goods, each denoted by m, (m∈M), representing transport from origin o∈O to destination d∈D via a chosen path. For each specific domestic goods-producing city and foreign goods-consuming city in the OD, as shown... Figure 2 ,by Represents all sets based on maritime routes, where the routes are defined by Ω. s In the shipping segment, A p Port section, A w Section A of the canal h The road sections in the middle are composed of; This represents the set of all routes based on rail transport; similarly, the routes within it are represented by Ω. r The railway transport section, A b In the section of the track switching station, A φ In the railway station section, A h The road is composed of sections of highway, with K as the main road. od express and The set that is formed, that is

[0063] In this embodiment, the variable symbols of the present invention are shown in Table 1;

[0064] Table 1 Variable Symbol Table

[0065]

[0066]

[0067]

[0068]

[0069] This example illustrates a multimodal freight network between China and Europe, with Chongqing, China, and Hamburg, Germany, serving as the origin and destination cities, respectively, to showcase the network structure and components. The example includes two liner shipping routes and two rail freight routes. Figure 2 Only one westbound liner service route is shown in the table. The ports or railway stations along the westbound routes of the four routes mentioned above are listed in Table 2:

[0070] Table 2 lists the ports or railway stations along the four westbound routes.

[0071]

[0072]

[0073] For example, a shipper wishing to transport goods from Chongqing to Hamburg has four options in the example network shown: two based on sea transport and two based on rail transport. On one hand, they can choose liner service 1 or 2. If choosing liner service 1, the goods must first be transported by truck from Chongqing to Shanghai port, then the Shanghai-Hamburg segment of liner service 1 is used. On this route, the goods will be loaded at Shanghai port, pass through the Suez Canal, and unloaded at Hamburg port. On the other hand, they can choose one of the two rail transport services. If choosing rail transport service 2, the goods must first be transported by truck from Chongqing to Chengdu station, then the Chengdu-Nuremberg rail segment of rail transport service 2 is used, and finally, the goods are transported again by road from Nuremberg to their destination, Hamburg. In this multimodal transport process, the goods will undergo a gauge change at Zamyn-Uud and Małaszewicze stations.

[0074] S2: Analyze freight traffic congestion time for bottleneck facilities (i.e., ports, Suez Canal, and rail transfer stations) in the route topology network; use the M / M / c queuing model to characterize port congestion; specifically, follow these steps:

[0075] S2.1: Assuming that the arrival time of customers follows a Poisson distribution, there are c service counters in the system, and the service time follows an exponential distribution. For the port system, the container ship is the customer in the queuing theory model, and the container ship berth is the service counter. Then the average stay time of the container ship in the port is as shown in equation (1).

[0076]

[0077] Among them, t p λ represents the average time a container ship spends in a port. p It is the average arrival rate of container ships, μ p c represents the average service rate of container ship berths. p It is the number of berths in the port, ρ p =λ p / c p μ p This indicates the average utilization rate of the port.

[0078] S2.2: The port's designed annual throughput is defined as the port's maximum handling capacity. Therefore, the port's utilization rate is the sum of China-Europe container freight volume and non-China-Europe container freight volume divided by the port's maximum handling capacity. By defining different utilization rates, the average service delay of the port was evaluated according to equation (1), as shown in equation (2).

[0079]

[0080] like Figure 4 , where x p It is the China-Europe container flow rate at port p. It is a pre-determined non-China-Europe container flow rate, u p Let p be the maximum throughput capacity of port p. β p , These are the parameters to be calibrated.

[0081] Furthermore, such as Figures 5-6 M / M / 1 queuing theory model is constructed to evaluate the congestion delay of the fleet when passing through the Suez Canal, where the ships are regarded as customers and the port of Tufic is regarded as a service desk; as in equations (3)-(4);

[0082]

[0083] Among them, t w Indicates the transit time of a ship sailing north through the Suez Canal, μ w It is the draw service rate of Port Tufiq, λ w It is the arrival rate of ships sailing north, τ w This refers to the transit time through the Suez Canal without queuing;

[0084]

[0085] Where x w To improve the flow rate of container cargo between China and Europe via the Suez Canal. For a predetermined non-Central Europe northbound container cargo flow rate, γ w To convert container cargo volume into a range factor for the number of container ships. It is a predetermined number of non-container ships, u w For the northbound ship throughput capacity of the Suez Canal, β w , These are the parameters to be calibrated.

[0086] Furthermore, the M / M / c queuing theory model is adopted to capture the delay caused by container transfer at the track-changing station; accordingly, the average delay time of container trains at the track-changing station is as shown in equation (5).

[0087]

[0088] Among them, t b λ is the average delay time of container trains at the track-changing station. b It is the arrival rate of container trains, μ b It is the average service rate of the track-changing station, c bIt is the number of container transshipment terminal equipment, ρ b It is the average utilization rate of the track-changing station, ρ b =λ b / c b μ b .

[0089] like Figure 7 The average service time of the station is evaluated by defining different utilization rates, as shown in equation (6);

[0090]

[0091] Where, x b It is the China-Europe container flow rate at the rail transfer station. It is a pre-determined non-China-Europe container flow rate, u b This is the maximum throughput capacity of the track-changing station. β b , These are the parameters to be calibrated.

[0092] S3: Multiply the transportation time by the time value to convert it into the corresponding monetary cost and calculate the transportation cost;

[0093] Furthermore, in step S3, the transportation time is multiplied by the time value and converted into the corresponding monetary cost to calculate the generalized transportation cost. The generalized cost of transporting a unit of TEU of m-type goods through route k is calculated as shown in equations (7)-(8).

[0094]

[0095] Equation (8) guarantees that, under the optimal freight flow allocation model, the delay time caused by the upper limit of transport capacity is minimized. It will only appear on routes that are already at full capacity.

[0096] In this embodiment, The perceived generalized transportation cost of goods of class m on path k connecting od is expressed as: in To determine the generalized transportation cost of the actual path under a given traffic allocation pattern, and This represents the stochastic term in perceived cost, indicating the perceived error in a shipper's route selection behavior due to numerous invisible influencing factors. The logit model characterizes the shipper's pattern-route selection behavior, and the stochastic term in the perceived generalized cost of the route represents the cost. If the route follows a Gumbel distribution, then the probability that the shipper chooses this route is... It is obtained by calculation using equation (9);

[0097]

[0098] Among them, positive numbers It is the range factor related to the variance of the perceived generalized transportation cost of the shipper of m-type goods. The model of the multi-mode multi-cargo category network equilibrium problem is as shown in Equation (10).

[0099]

[0100] in,

[0101]

[0102]

[0103] In this embodiment, the above formula (10) is calculated using the following steps, as shown in formula (11);

[0104]

[0105] Flow of m types of goods on path k in L Taking the partial derivatives of the variables yields equation (12);

[0106]

[0107] The first derivative only occurs when It will only exist when the time is right. Let equation (11) equal to 0 to get equation (13);

[0108]

[0109] in It is a type m cargo on path k. The generalized transportation cost under the traffic volume level; as shown in equation (14);

[0110]

[0111] for Equation (22) is used to sum all path results between each pair of ODs to obtain equation (15);

[0112]

[0113] The path selection probability obtained by combining equations (13) and (14) is equivalent to the logit model function, as shown in equation (16);

[0114]

[0115] In this embodiment, S4: Analyze and calculate the spatial distribution data of China-Europe freight flows. This is done using a subgradient algorithm embedded with the DSD algorithm.

[0116] First, initialize the Lagrange multipliers μ. 0 Let the parameters s1>0, s2≥0 be related to the Lagrange multipliers, the iteration number n=0, the lower bound LB=-∞, the upper bound UB=+∞, and the parameters ∈1,∈2,∈3,k,α,β>0.

[0117] Step 1.1: Calculate the path cost according to equation (7).

[0118] Step 1.2: Calculate the path flow according to equation (9)

[0119] Step 1.3: Update the path cost according to equation (7)

[0120] Step 1.4: Calculate the auxiliary flow according to equation (9) f i

[0121] Step 1.5: Calculate segment flow and auxiliary road section flow x i ,like Then end and let Calculate segment flow x n+1 Then skip to Step 2; otherwise proceed to Step 1.6.

[0122] Step 1.6: Use the Armijo method to find the optimal solution λ for the following univariate convex programming problem. i :

[0123]

[0124] Step 1.7: Update the path flow

[0125] Step 1.8: Let i = i + 1 and jump to Step 1.3.

[0126] Step 2: Feasibility Test

[0127] if If the condition is met, proceed to Step 3; otherwise, proceed to Step 4.

[0128] Step 3: Convergence Test

[0129] Calculate B = L(f) n+1 ),LB=Z(f n+1 If (UB-LB) / LB<∈2 holds, and there is If true, then output f. n+1 Stop the algorithm; otherwise, skip to Step 4.

[0130] Step 4: Update the Lagrange multipliers and step size using the subgradient method.

[0131] If n≤k:ω n =s1 / (n+s2)

[0132]

[0133] Let n = n+1 and jump to Step 1;

[0134] If n>k:ω n =s1 / (n+s2)

[0135]

[0136] Let n = n+1 and jump to Step 1.

[0137] In this embodiment, 50 Chinese cities were selected as originating points and 51 European cities as destinations. Raw Sino-European freight trade data was collected from Chinese customs, including basic trade information on Sino-European freight transport: transport date, commodity category and code (HS code), origin and destination locations, and value. More specifically, in the customs data, the export origin was a Chinese provincial-level administrative unit, and the trade destination was a European country. Therefore, to obtain the trade value of each commodity from Chinese cities to European cities, the raw demand data was further subdivided according to the population size of each European city and the export trade value of each Chinese city to Europe. Furthermore, commodities were divided into nine categories based on 2-digit and 4-digit HS codes, as detailed in Table 2. After determining the value per container for each commodity category, the value was converted into the number of containers. These nine commodity categories can be further divided into three groups according to commodity characteristics: highly time-sensitive commodities, moderately time-sensitive commodities, and low-time-sensitive commodities, as shown in Table 3. Each group of commodities has a corresponding depreciation rate due to spoilage or damage.

[0138] Table 3 Product Categories

[0139] Serial Number Product Categories <![CDATA[Time sensitivity level 1 > Depreciation rate (daily) <![CDATA[Value ($ / TEU) 2 > 1 food high 2.500％ 10,391 2 Clothing middle 0.125％ 21,739 3 Electronic and electrical products middle 0.125％ 277,778 4 Household goods middle 0.125％ 48,475 5 Toys / Sports Goods middle 0.125％ 37,314 6 textile Low 0.025％ 5,053 7 steel products Low 0.025％ 30,968 8 mechanical products Low 0.025％ 147,853 9 plastic Low 0.025％ 14,853

[0140] The complete China-Europe multimodal transport network being built, such as Figure 9As shown, this includes 50 Chinese cities (originating points), 51 European cities (destination points), 39 seaports, 40 railway stations, 5 transshipment stations, 27 China-Europe liner shipping routes, and 55 China-Europe freight train routes. Furthermore, domestic freight transport in China and cross-border freight transport in Europe connect the origin and destination to ports or railway stations via their respective road networks, providing convenient "door-to-door" freight services, which is the primary mode of transport chosen by most shippers. Therefore, our China-Europe multimodal transport network encompasses the necessary major domestic Chinese road networks and cross-border European road networks. Based on this, we have pre-generated optional complete freight service routes for shippers in each pair of China-Europe cities. Each route includes all stages from the origin of the goods, via road transport to the port or railway station, to rail transport or liner shipping to Europe, and then back to road transport to the destination. Meanwhile, to maintain a reasonable number of potentially usable freight routes, the number of ports and railway stations reachable from each origin and destination is limited to 3 to 6, and the longest distance from the origin and destination to a port or railway station cannot exceed twice the shortest distance. For example, in the fourth quarter of 2019, our China-Europe multimodal freight network comprised 180 nodes, 1,082 transport segments, and 2,550 origin-destination pairs, encompassing a total of 68,426 freight service routes, including 60,060 routes to Europe via liner shipping and 8,366 routes to Europe via the China-Europe Railway Express.

[0141] The model was calibrated based on statistical data on the number of freight containers from China to Europe and the throughput of the Suez Canal. The parameters θ = [θ] in the model were calibrated using the least squares method. m This parameter reflects the level of uncertainty a shipper perceives regarding the transportation costs of different categories of goods. Assume... This represents the freight volume from China to Europe predicted by the model for port p in month i. These represent the predicted freight volumes for track-changing station b and canal w, respectively. This represents the actual freight volume. Additionally, |A p |,|A b |,|A w | represents the basic freight volume and container count for the port, rail transfer station, and canal, respectively. The least squares function W(θ) can then be expressed as:

[0142]

[0143] We first assume θ m =θ holds for any m, and then the above problem was solved through trial and error, yielding θ. *=0.0255. Tables 4-6 show the relative differences between the predicted and true values ​​of container throughput for all bottleneck facilities. This calibration result indicates that the proposed model, after calibration, achieves an acceptable level of accuracy for planning or strategic decision-making.

[0144] Table 4. Calibration Differences in Bottleneck Facility Throughput Model

[0145]

[0146]

[0147] Table 5 Track Changing Stations

[0148] Track changing station <![CDATA[Percentage of Average Relative Error 4 > Dostoker 15.09％ Malazevic 11.47％

[0149] Table 6 Canals

[0150] canal <![CDATA[Percentage of Average Relative Error 5 > Suez Canal 0.00％

[0151] In this embodiment, Figure 8 The convergence curve for solving the Lagrange multiplier problem using the subgradient method is shown. Figure 9 The convergence curves of solving the relaxed Lagrange problem (i.e., the stochastic user equilibrium problem without capacity constraints) using the DSD algorithm are shown and compared. Figure 9 Convergence curves for the subgradient method and the DSD algorithm are presented, and the relationship between the two algorithms is illustrated. Overall, from... Figure 9 As can be seen, the algorithm proposed in this invention exhibits excellent convergence, converging to a point where the difference between the upper bound (i.e., the dashed line) and the lower bound (i.e., the solid line) is less than 0.001%. Taking December 2019 as an example, the algorithm terminated after 714 iterations and 6,365 seconds of CPU time, when the objective function value reached 290,205,793.

[0152] from Figure 9 It can be seen that the method of the present invention exhibits excellent performance in large-scale problems. As the overall algorithm iterates, the computation time required for each relaxed Lagrange problem to reach convergence is significantly reduced.

[0153] In this embodiment, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a cross-border transportation intercontinental freight flow prediction and analysis method as described in any one of the above embodiments.

[0154] In this embodiment, the present invention provides a computer-readable storage medium, which includes an embedded processing system and a stored program. When the embedded system control program is running, it controls the execution of any one of the above-described methods for predicting and analyzing intercontinental freight traffic in cross-border transportation.

[0155] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations will be apparent to those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting and analyzing intercontinental freight traffic in cross-border transportation, characterized in that: Follow these steps: S1: First, construct a multimodal freight network between China and Europe based on the route topology diagram, specifically a multi-mode, multi-cargo type freight transport network; S2: Analyze the freight traffic congestion time of ports, Suez Canal and rail transfer stations in the route topology network. S3: Multiply the transportation time by the time value to convert it into the corresponding monetary cost and calculate the transportation cost; S4: Analyze and calculate the spatial distribution data of China-Europe freight flows.

2. The method for predicting and analyzing intercontinental freight traffic in cross-border transportation according to claim 1, characterized in that: In step S1, in the path topology diagram, the starting point and the ending point are respectively represented by OD, which are domestic cities producing goods and foreign cities consuming goods. The path topology network is graph G(N,A), where N represents the set of nodes and A represents the set of road segments. The set of nodes includes four types of nodes: O is the set of origin cities for goods; D is the set of destination cities for goods; Φ is the set of inbound and outbound systems for ports, canals, rail transfer stations, and railway stations; the set of road segments consists of highway segments, shipping segments, railway segments, port segments, canal segments, and rail transfer station segments. Use L s This represents a collection of liner shipping routes, where each service is provided by a liner shipping company and consists of at least one sea segment; L r This represents a collection of railway transport service routes, each provided by the China-Europe Railway Express Company and consisting of at least one railway transport segment. Ω s Indicates a shipping segment, representing a port pair segment consisting of consecutive shipping segments within each liner service route; Ω r It indicates a railway transportation section.

3. The method for predicting and analyzing intercontinental freight traffic in cross-border transportation according to claim 1, characterized in that: In step S2, the M / M / c queuing model is used to characterize port congestion; specifically, the following steps are performed: S2.1: Assuming that the arrival time of customers follows a Poisson distribution, there are c service counters in the system, and the service time follows an exponential distribution. For the port system, the container ship is the customer in the queuing theory model, and the container ship berth is the service counter. Then the average stay time of the container ship in the port is as shown in equation (1). Among them, t p λ represents the average time a container ship spends in a port. p It is the average arrival rate of container ships, μ p c represents the average service rate of container ship berths. p It is the number of berths in the port, ρ p =λ p / c p μ p This indicates the average utilization rate of the port. S2.2: The port’s design annual throughput is defined as the port’s maximum handling capacity. By defining different utilization rates, the port’s average service delay is evaluated according to Equation (1), as shown in Equation (2). Where, x p It is the China-Europe container flow rate at port p. It is a pre-determined non-China-Europe container flow rate, u p Let p be the maximum throughput capacity of port p. β p , These are the parameters to be calibrated.

4. The method for predicting and analyzing intercontinental freight traffic in cross-border transportation according to claim 1, characterized in that: In step S2, an M / M / 1 queuing theory model is constructed to evaluate the congestion delays of the fleet when passing through the Suez Canal, where the ships are regarded as customers and the port of Tufic is regarded as a service desk; as in equations (3)-(4); Among them, t w Indicates the transit time of a ship sailing north through the Suez Canal, μ w It is the draw service rate of Port Tufiq, λ w It is the arrival rate of ships sailing north, τ w This refers to the transit time through the Suez Canal without queuing; Where x w To improve the flow rate of container cargo between China and Europe via the Suez Canal. For a predetermined non-Central Europe northbound container cargo flow rate, γ w To convert container cargo volume into a range factor for the number of container ships. It is a predetermined number of non-container ships, u w For the northbound ship throughput capacity of the Suez Canal, β w , These are the parameters to be calibrated.

5. According to the method for predicting and analyzing cross-border intercontinental freight traffic as described in claim 1, in step S2, the M / M / c queuing theory model is used to capture the delay caused by container transshipment at the track-changing station; accordingly, the average delay time of container trains at the track-changing station is as shown in formula (5). in, t b λ is the average delay time of container trains at the track-changing station. b It is the arrival rate of container trains, μ b It is the average service rate of the track-changing station, c b It is the number of container transshipment terminal equipment, ρ b It is the average utilization rate of the track-changing station, ρ b =λ b / c b μ b . The average service time of the station is evaluated by defining different utilization rates, as shown in Equation (6); Where, x b It is the China-Europe container flow rate at the rail transfer station. It is a pre-determined non-China-Europe container flow rate, u b This is the maximum throughput capacity of the track-changing station. β b , These are the parameters to be calibrated.

6. According to the method for predicting and analyzing cross-border intercontinental freight traffic as described in claim 1, in step S3, the transportation time is multiplied by the time value and converted into the corresponding monetary cost, and the generalized transportation cost is calculated. The generalized cost of transporting m-type goods per unit TEU through route k is calculated as shown in equations (7)-(8).

7. The method for predicting and analyzing intercontinental freight traffic in cross-border transportation according to claim 1, in step S3, using... The perceived generalized transportation cost of goods of class m on path k connecting od is expressed as: in To determine the generalized transportation cost of the actual path under a given traffic allocation pattern, and This represents the stochastic term in perceived cost, indicating the perceived error in a shipper's route selection behavior due to numerous invisible influencing factors. The logit model characterizes the shipper's pattern-route selection behavior, and the stochastic term in the perceived generalized cost of the route represents the cost. If the route follows a Gumbel distribution, then the probability that the shipper chooses this route is... It is obtained by calculation using equation (9); Among them, positive numbers It is the range factor related to the variance of the perceived generalized transportation cost of the shipper of m-type goods. The model of the multi-mode multi-cargo category network equilibrium problem is as shown in Equation (10). in, 8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the cross-border transportation intercontinental freight flow prediction and analysis method according to any one of claims 1-7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes an embedded processing system and a stored program, which executes the cross-border transportation intercontinental freight flow prediction and analysis method according to any one of claims 1-7 when the embedded system control program is running.

Images