Multimodal transport network optimization method considering demand uncertainty and bulk-to-consolidated transport

Abstract

The present application relates to the field of multimodal transport technology, specifically, to a multimodal transport network and optimization method that takes into account demand uncertainty and bulk-to-container conversion, which can, to a certain extent, solve the problem that the existing mathematical model is not practical enough and the existing algorithm cannot obtain the global optimal solution. The network includes: nodes and transportation processes; the nodes include supply ports, transshipment ports and demand nodes, the supply ports are large international ports, the transshipment ports are national coastal ports, the demand nodes are inland factories, and there are bulk-to-container conversion processes at the supply ports and transshipment ports; the transportation process includes international transportation and domestic transportation, the international transportation includes bulk cargo loading and container transportation, the domestic transportation includes transportation between transshipment ports and transportation from transshipment ports to demand nodes, the domestic transportation considers multimodal transportation, and the modes of domestic transportation include inland waterway transportation, railway transportation and road transportation.

Description

Technical Field

[0001] The present application relates to the field of multimodal transport technology, and more specifically, to a multimodal transport network optimization method that takes into account demand uncertainty and bulk-to-bulk conversion. Background Art

[0002] Bulk-to-Containerization can also be called bulk containerization, which means that bulk cargo is concentrated into large standardized containers to simplify loading and unloading work and carry out high-quality transportation. In order to make full use of transportation capacity, implement green and environmentally friendly transportation, and carry out high-quality transportation, my country has trial-run bulk-to-Containerization and multimodal transport in high-throughput ports such as Tianjin Port, Fuzhou Port, and Dalian Port. I believe that more ports will adopt this model. The current bulk cargo transportation model is that dry bulk carriers are unloaded by the unloading equipment of the quay crane after arriving at the port, and then transported to the dry bulk yard by port transport vehicles and then transferred by container trucks. However, there will be damage to bulk cargo during the loading and unloading and transportation process, and the transportation process will also pollute the environment, and bulk cargo is not easy to transfer. At terminals that combine bulk-to-container and multimodal transport, bulk-to-container equipment directly loads bulk carriers into containers after the bulk carriers dock, greatly reducing cargo spillage. Horizontal transport equipment navigation vehicles (internal container trucks) then realize the circulation of containers between the front-end shore logistics subsystem and the yard container area subsystem. In addition, the landside logistics subsystem and the yard container area subsystem are realized through the coordinated operation of external container trucks (External Truck, referred to as external container trucks / ET) and yard cranes.

[0003] As the trade volume of dry bulk cargoes gradually increases, the traditional dry bulk cargo transportation network design may no longer be able to adapt to the future demand for bulk cargo transportation. Most of the traditional bulk cargo transportation is chartered, and the demand for bulk cargoes in each cycle is often uncertain, which poses certain challenges to chartering ships, and there are a certain amount of cargo damage and environmental pollution problems. In the face of the current rise in bulk freight rates, a system combining bulk-to-container and multimodal transport should be developed to fully utilize the transportation capacity. On the one hand, it can cope with the uncertainty of bulk cargo demand, and on the other hand, it can use idle containers to transport bulk cargoes, avoiding large-scale empty container storage, and can also use the low cargo damage of container transportation to improve customer service levels and satisfaction.

[0004] However, there are also three major difficulties in the multimodal transport network optimization problem considering bulk-to-container conversion: how to decide which port to purchase bulk-to-container equipment, how much bulk-to-container conversion is required, and how much each mode of transport transports. These all pose considerable challenges to network design. The main methods currently used to solve the multimodal transport network design problem include: mathematical programming, heuristic algorithms, and exact algorithms.

[0005] However, the existing mathematical models are not realistic enough, and the modeling of multimodal transport optimization problems rarely takes into account uncertainty factors. In addition, there are a large number of heuristic algorithms for multimodal transport problems, but these heuristic algorithms only obtain local optimal solutions rather than global optimal solutions. Summary of the invention

[0006] In order to solve the problem that the existing mathematical models are not practical enough, the modeling of multimodal transport optimization problems rarely considers uncertainty factors; and there are a large number of heuristic algorithms for multimodal transport problems, but these heuristic algorithms only obtain local optimal solutions rather than global optimal solutions. This application provides a multimodal transport network optimization method that considers demand uncertainty and scattered-to-collective transformation.

[0007] The embodiment of the present application is implemented as follows:

[0008] The present application provides a method for optimizing a multimodal transport network taking into account demand uncertainty and bulk-to-bulk conversion, wherein the multimodal transport network includes nodes and transport processes;

[0009] The nodes include supply ports, transit ports and demand nodes. The supply ports are large international ports, the transit ports are national coastal ports, and the demand nodes are inland factories. There is a bulk-to-container process at the supply ports and the transit ports.

[0010] The transportation process includes international transportation and domestic transportation. The international transportation includes bulk cargo loading and container transportation. The domestic transportation includes transportation between transit ports and transportation from transit ports to demand nodes. The international transportation includes ocean transportation. The domestic transportation considers multimodal transportation. The domestic transportation methods include inland waterway transportation, railway transportation and road transportation.

[0011] The costs of the international transportation stage include: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the cost of chartering bulk carriers, the transportation cost of bulk cargoes and containers, the variable cost of bulk-to-container conversion and the environmental cost;

[0012] The costs of the domestic transportation stage include: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the intermodal transportation cost of bulk and container, the penalty cost of shortage, the variable cost of bulk-to-container, the environmental cost and the port inventory cost;

[0013] Optimization methods include:

[0014] Construct a deterministic model, define the set required for constructing the model, define the parameters required for constructing the model, define the decision variables required for constructing the model, define the dependent variables required for constructing the model, and obtain the objective function of the deterministic model based on the set, parameters, decision variables and dependent variables:

[0015] By means of a stochastic optimization method, a problem model is constructed according to the deterministic model, main problem parameters are defined, sub-problem parameters are defined, main problem decision variables are defined, sub-problem decision variables are defined, and dependent variables are defined, and an objective function of the main problem model and an objective function of the sub-problem model are obtained according to the main problem parameters, sub-problem parameters, main problem decision variables, sub-problem decision variables and dependent variables;

[0016] The problem model is solved based on an accurate algorithm to obtain an optimization solution for the multimodal transport network. The first iteration requires initializing the binary decision variables in the main problem model and setting the upper and lower limits of the objective function.

[0017] In each scenario, the feasible solutions of the main problem model are brought into solving the subproblems.

[0018] For the solution of the sub-problem in each scenario, if the sub-problem model has an optimal solution, the optimal cut is generated and added to the constraints of the main problem model; otherwise, a feasible cut is generated and added to the constraints of the main problem model;

[0019] Solve the main problem model after one iteration. If the upper and lower limits are equal, the algorithm terminates and the optimization solution is obtained.

[0020] Otherwise, continue to solve the sub-problems in each scenario. If the sub-problem model has an optimal solution, generate the optimal cut and add it to the constraints of the main problem model. Otherwise, generate a feasible cut and add it to the constraints of the main problem model. This process continues until all sub-problem models have been solved. Then go back to each scenario and bring the feasible solution of the main problem model into solving the sub-problems.

[0021] In a possible implementation, the optimization method also includes a planning period, which includes multiple cycles. Each cycle requires a decision on the purchase of bulk-to-container equipment and a decision on bulk cargo chartering. The supply port and the transit port can only carry out the bulk-to-container process after purchasing the bulk-to-container equipment. Each cycle also requires a decision on the bulk cargo and bulk-to-container freight volume for each mode of transportation in the domestic transportation.

[0022] In a possible implementation, the objective function includes:

[0023] Equipment purchase and maintenance costs, intermodal transport costs of bulk cargo and containers, bulk cargo chartering costs, labor costs for converting bulk cargo to container cargo, warehousing costs and environmental costs.

[0024] In a possible implementation, the objective function further includes:

[0025] Out-of-stock penalty constraints, unmet freight demand constraints, inventory balance constraints, bulk cargo and bulk-to-container cargo constraints and cargo supply limit constraints, port capacity limit constraints, bulk carrier capacity limit constraints, and port bulk-to-container restriction constraints.

[0026] In a possible implementation, solving the problem model based on an accurate algorithm to obtain an optimization solution for the multimodal transport network further includes:

[0027] Put all the sub-problems into the main problem model at the same time to reduce the number of iterations and quickly narrow down the solution space.

[0028] The technical solution provided by this application can at least achieve the following beneficial effects:

[0029] The multimodal transport network optimization method considering demand uncertainty and bulk-to-container conversion provided by the present application adds bulk-to-container conversion to the multimodal transport network model, and optimizes the minimization of comprehensive costs such as facility cost, transportation cost, variable cost of bulk-to-container conversion, inventory cost, and environmental cost. For the ocean transport part of the multimodal transport optimization problem considering bulk-to-container conversion, a transport decision that considers cross-cycle transport is added to the model to achieve a process that is more in line with actual transport. The uncertainty of bulk demand and cross-cycle transport in ocean transport are considered, which are two important aspects that are in line with the actual situation of bulk multimodal transport. The problem is divided into a main problem and a sub-problem for multi-cycle planning, and an accurate algorithm for solving the main and sub-problems is designed.

[0030] Moreover, in the mathematical programming model, we split the model into main problems and sub-problems, use precise algorithms, and add scenario-based methods to precise algorithms for a large number of demand scenarios caused by demand uncertainty, which can speed up model solving and reduce the number of iterations. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0032] Figure 1 is a schematic diagram of a multimodal transport network considering demand uncertainty and bulk-to-bulk transport, as shown in an exemplary embodiment of the present application;

[0033] Figure 2 is a flow chart of a multimodal transport network optimization method considering demand uncertainty and bulk-to-bulk transport, shown in an exemplary embodiment of the present application;

[0034] Figure 3 is a schematic diagram of a process of constructing a deterministic model according to another exemplary embodiment of the present application;

[0035] Figure 4is a flowchart of constructing a problem model according to another exemplary embodiment of the present application;

[0036] Figure 5 is a flow chart of solving a problem model by using an accurate algorithm, shown in another exemplary embodiment of the present application;

[0037] Figure 6 It is a schematic diagram of the plan layout of a dry bulk cargo transshipment terminal in the prior art;

[0038] Figure 7 It is a schematic diagram of the plan layout of a bulk-to-container multimodal transport terminal in the prior art;

[0039] Figure 8 is a flow chart of bulk cargo and containerized cargo in a bulk-to-container operation shown in another exemplary embodiment of the present application;

[0040] Fig. 9 is a flow chart of bulk cargo and containers in a transshipment port shown in another exemplary embodiment of the present application;

[0041] Fig.10 is a schematic diagram of the time points and relationships of the decisions of each cycle within the planning period shown in another exemplary embodiment of the present application;

[0042] Fig.11 It is a schematic diagram of a pseudo code of a Multicut Benders Decomposition algorithm shown in another exemplary embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, implementation mode and advantages of the present application clearer, the exemplary implementation mode of the present application will be clearly and completely described below in conjunction with the drawings in the exemplary embodiments of the present application. Obviously, the described exemplary embodiments are only part of the embodiments of the present application, not all of the embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0044] It should be noted that the brief description of terms in this application is only for the convenience of understanding the embodiments described below, and is not intended to limit the embodiments of this application. Unless otherwise specified, these terms should be understood according to their ordinary and common meanings.

[0045] The terms "first", "second", "third", etc. in the specification and claims of this application and the above drawings are used to distinguish similar or similar objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise noted. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances.

[0046] The terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device comprising a list of components is not necessarily limited to all the components expressly listed but may include other components not expressly listed or inherent to such product or device.

[0047] Before explaining the multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport provided in the embodiment of the present application, the application scenario and implementation environment of the embodiment of the present application are first introduced.

[0048] Figure 6 It is a schematic diagram of the plan layout of the dry bulk cargo transshipment terminal in the prior art. Figure 7 It is a schematic diagram of the plan layout of a bulk-to-container multimodal transport terminal in the prior art.

[0049] like Figure 6 and Figure 7 As shown in the figure, bulk-to-containerization can also be called bulk containerization, which means that bulk cargo is concentrated into large standardized containers to simplify loading and unloading and to carry out high-quality transportation. In order to make full use of transportation capacity, implement green and environmentally friendly transportation, and carry out high-quality transportation, my country has trial-run bulk cargo transportation combining bulk-to-containerization and multimodal transportation in high-throughput ports such as Tianjin Port, Fuzhou Port, and Dalian Port. It is believed that more ports will adopt this model.

[0050] like Figure 2 As shown in the figure, the current bulk cargo transportation mode is that the dry bulk carrier unloads the cargo by the unloading equipment of the quay crane after arriving at the port, and then transports the cargo to the dry bulk cargo yard by the port transport vehicle and then transfers it by the container truck. However, the bulk cargo will be damaged during the loading and unloading and transportation process, and the transportation process will also pollute the environment, and the bulk cargo is not easy to transfer. The terminal that combines bulk-to-container and multimodal transport directly loads the cargo into the container by the bulk-to-container equipment after the bulk carrier docks, which greatly reduces the spillage of the cargo. After that, the horizontal transport equipment navigation vehicle (referred to as the internal container truck) realizes the container circulation of the front shore logistics subsystem and the yard container area subsystem. In addition, the landside logistics subsystem and the yard container area subsystem are realized through the cooperation of the external container truck (External Truck, referred to as the external container truck / ET) and the yard crane.

[0051] As the trade volume of dry bulk cargoes gradually increases, the traditional dry bulk cargo transportation network design may no longer be able to adapt to the future demand for bulk cargo transportation. Most of the traditional bulk cargo transportation is chartered, and the demand for bulk cargoes in each cycle is often uncertain, which poses certain challenges to chartering ships, and there are also a certain amount of cargo damage and environmental pollution problems. In the face of the current rise in bulk freight rates, a system combining bulk-to-container and multimodal transport should be developed to make full use of transportation capacity.

[0052] On the one hand, it can cope with the uncertainty of bulk cargo demand, and on the other hand, it can utilize idle empty containers to transport bulk cargo, avoiding large-scale empty container storage. It can also utilize the low cargo damage in container transportation to improve customer service levels and satisfaction.

[0053] However, there are three major difficulties in optimizing the multimodal transport network considering bulk-to-container conversion: how to decide which port to purchase bulk-to-container conversion equipment, how much bulk-to-container conversion is required, and how much each mode of transport transports. These have brought considerable challenges to network design.

[0054] The main existing methods for multimodal transport network design problems are:

[0055] [1] Mathematical programming

[0056] A mathematical programming model is established, in which the transport volume of various modes of transport and the location of multimodal transport hubs are taken as decision variables, and the minimization of transport cost, comprehensive cost and transport time is taken as the objective function.

[0057] [2] Heuristic Algorithm

[0058] Use common heuristic algorithms such as genetic algorithms, taboo search, greedy algorithms, etc. to solve multimodal transport optimization problems, or add rules or strategies defined based on the characteristics of the problem to the developed algorithms to accelerate the algorithms or improve algorithm performance.

[0059] [3] Exact algorithm

[0060] The use of precise algorithms to solve multimodal transport optimization problems generally involves cutting plane method, integer programming method, branch and bound method, etc. This algorithm often requires the application of software such as LINGO and CPLEX, which can effectively improve the efficiency of solution.

[0061] The main defects are:

[0062] [1] The mathematical model is not realistic enough

[0063] Existing optimization modeling for multimodal transport rarely considers uncertainty factors. Changes in demand are likely to affect the long-term planning of the entire logistics network, so they need to be considered.

[0064] In international transportation, due to its long voyage and long time span, it is very likely that the goods cannot be delivered within one cycle. However, there are very few studies that consider this factor in multimodal transportation, so it is necessary to consider the situation of cross-cycle transportation.

[0065] [2] Algorithm performance is not good enough

[0066] Currently, in business decisions, the global optimal solution is more valuable than the local optimal solution.

[0067] Because business decisions involve many variables and factors, if you only consider the optimal solution in a certain local area, it is easy to fall into limited thinking and miss better opportunities. The global optimal solution can take into account all factors and variables and make the best decision based on long-term interests. Therefore, from a business perspective, precise algorithms should be used.

[0068] However, the exact algorithm will face the problem of long solution time and too large solution space, resulting in low solution efficiency. In real life, it often leads to the solution time exceeding 3600s before the actual scale is reached. Therefore, the exact algorithm needs to be optimized for the solution time.

[0069] Based on this, the present application provides a multimodal transport network optimization method considering demand uncertainty and bulk-to-container conversion. It solves the multimodal transport network optimization problem considering bulk-to-container conversion by improving the design algorithm, takes into account the mixed decision of transport volume and equipment, and takes into account the uncertainty of bulk cargo demand, thereby enhancing the robustness of the model.

[0070] In view of the characteristics of long planning cycle and short transportation cycle, multi-cycle planning is proposed, which takes into account the cross-cycle transportation of goods, making the model more practical and helpful for international ocean shipping operators to allocate resources more reasonably and quickly. This patent mainly establishes a multi-cycle mathematical programming model for bulk multimodal transport considering bulk-to-container under uncertain demand conditions. The goal of the optimization model is the bulk-to-container network design cost and transportation cost.

[0071] This patent divides the multimodal transport network optimization problem of bulk-to-container conversion into a main problem and a sub-problem, where the main problem is only responsible for solving the optimal layout of bulk-to-container equipment and bulk chartering strategy, that is, whether the port purchases bulk-to-container equipment and whether the shipping company leases a bulk carrier; the sub-problem uses the solution of the main problem as a known parameter, aiming to find an optimal multimodal transport solution that transports the bulk cargo demand from the supply port to the demand node in each cycle, that is, the optimal bulk cargo transportation volume and bulk-to-container cargo transportation volume under various transportation modes. Then, using the structure of the main and sub-problems, we developed an improved BendersDecomposition algorithm to apply to this problem. This method of splitting the problem into a main problem and sub-problems for solution can quickly solve the problem to the precise optimal solution under large-scale examples, and is both general and practical.

[0072] Next, the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems will be described in detail through embodiments and in combination with the accompanying drawings. The embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. Obviously, the described embodiments are part of the embodiments of the present application, not all of them.

[0073] Figure 1 It is a schematic diagram of the structure of a multimodal transport network considering demand uncertainty and bulk-to-bulk transport, as shown in an exemplary embodiment of the present application.

[0074] Figure 2 It is a flowchart of a multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport, shown in an exemplary embodiment of the present application.

[0075] In an exemplary embodiment, Figure 1 As shown, a multimodal transport network considering demand uncertainty and bulk-to-bulk conversion is provided, which includes nodes and transportation processes;

[0076] The nodes include supply ports, transit ports and demand nodes. The supply ports are large international ports, the transit ports are national coastal ports, and the demand nodes are inland factories. There is a bulk-to-container process at the supply ports and the transit ports.

[0077] The transportation process includes international transportation and domestic transportation. The international transportation includes bulk cargo shipping and container transportation. The domestic transportation includes transportation between transshipment ports and transportation from transshipment ports to demand nodes. The domestic transportation considers multimodal transportation. The domestic transportation modes include inland waterway transportation, railway transportation and road transportation.

[0078] Among them, the supply port is a large international port (such as the Port of New York and the Port of Los Angeles), the transshipment port is a national coastal port (such as the Port of Shanghai and the Port of Tianjin), and the demand node is an inland factory. International transportation is assumed to be carried out only by sea transportation through bulk chartering or container transportation. Domestic transportation consists of two parts: transportation between transshipment ports and transportation from transshipment ports to demand nodes. Multimodal transport needs to be considered, and the bulk-to-container process only occurs at the supply port and the transshipment port.

[0079] It can be seen that each port can only carry out bulk-to-container conversion after purchasing bulk-to-container equipment, but this will also generate variable costs for bulk-to-container conversion. It is also necessary to decide the bulk and bulk-to-container freight volumes for each mode of transportation in the domestic section. How to use appropriate methods to fully consider the uncertainty of the demand of multiple customers per cycle within a limited time, arrange the purchase of bulk-to-container equipment, bulk chartering strategies, and multimodal transport plans and bulk-to-container plans for each port per cycle to reduce comprehensive costs (equipment purchase costs, bulk chartering costs, bulk-to-container variable costs, unmet demand penalty costs per cycle at demand nodes, environmental costs, transportation costs, and port inventory costs) is crucial for buyers, multimodal transport operators, and ports.

[0080] In a possible implementation, the costs of the international transportation stage include: the purchase cost and per-cycle maintenance cost of bulk-to-container equipment, the cost of chartering a bulk carrier, the transportation cost of bulk cargo and containers, the variable cost of bulk-to-container conversion and the environmental cost.

[0081] Among them, the cost of the ocean transport stage, that is, the first stage of the multimodal transport network we consider, the supply of bulk cargo at each supply port, directly affects the subsequent transportation mode arrangement and the second stage arrangement. The following are several factors closely related to the cost of the ocean transport stage:

[0082] [1] Supply volume at the supply port

[0083] The impact of the supply quantity of the supply port on the cost of the ocean transport stage is obvious. The more the supply quantity of the supply port, the more goods need to be transported, and under the same conditions, the total ocean transport cost will be higher.

[0084] [2] Purchase of bulk-to-container equipment at supply ports

[0085] Since the premise of bulk-to-container transportation is that the port needs to have bulk-to-container equipment in the current cycle, it is necessary to decide in which cycle the port will purchase the equipment. The purchase of equipment will incur fixed costs. Then, when bulk-to-container is carried out in each cycle, the equipment will also have maintenance costs and a certain amount of labor costs per unit of bulk cargo. This part is called the variable cost of bulk-to-container. Faced with the uncertainty of bulk cargo demand in each cycle, some supply ports will not purchase bulk-to-container equipment because they feel that their supply volume is small or the cost of purchasing equipment is high. This will naturally lead to bulk cargo chartering transportation. This will then generate chartering costs.

[0086] [3] Cargo damage and environmental costs

[0087] In the process of bulk cargo transportation, it is normal for bulk cargo to deteriorate due to its lack of airtightness and spillage during transportation. This part of the loss is called bulk cargo damage. After bulk cargo is converted to container cargo, the airtightness of container transportation can greatly reduce cargo damage. However, both container transportation and bulk cargo transportation have carbon emissions, which is an environmental cost.

[0088] After the above analysis, it can be concluded that the cost of ocean shipping includes: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the cost of chartering a bulk carrier, the cost of bulk and container transportation, the variable cost of bulk-to-container conversion and the environmental cost.

[0089] In a possible implementation, the costs of the domestic transportation stage include: bulk-to-container equipment purchase costs and per-cycle maintenance costs, bulk and container multimodal transport costs, out-of-stock penalty costs, bulk-to-container variable costs, environmental costs and port inventory costs.

[0090] Among them, the cost of the domestic transportation stage refers to the process of transshipment between transit ports and transportation from transit ports to demand nodes. Transshipment ports may also have bulk-to-container operations that affect costs. After bulk-to-container operations, there are also costs related to each cycle of transportation. The following are the relevant factors of the domestic transportation stage cost:

[0091] [1] Purchase of equipment for converting bulk to container at transshipment ports

[0092] Like supply ports, transshipment ports also have bulk-to-container conversion. Bulk cargo transported from supply ports or other transshipment ports to the transshipment port will be converted from bulk to container. Since the premise of converting bulk cargo to container is that the port must have bulk-to-container equipment in the current cycle, it is necessary to decide in which cycle the port will purchase the equipment, which will incur fixed costs. Subsequently, when converting bulk cargo to container in each cycle, the equipment will also have maintenance costs and a certain amount of labor costs per unit of bulk cargo.

[0093] [2] Transportation costs and related costs

[0094] Domestic transportation of bulk cargo and containerized cargo both use multimodal transport. Different modes of transport will generate different transportation costs. The domestic transportation process consists of two parts: 1) Transportation between transit ports, which involves the transportation costs of bulk cargo and containerized cargo after bulk conversion, as well as the cost of cargo damage during bulk cargo transportation and the environmental cost corresponding to the transportation process. 2) Transportation from transit ports to demand nodes. Each cycle of transportation consists of the costs of bulk cargo transportation and containerized cargo transportation, and there will also be cargo damage and environmental costs during bulk cargo transportation.

[0095] [3] Out-of-stock penalty costs

[0096] Although multimodal transport is adopted, out-of-stock situations may still occur due to demand uncertainty, bulk cargo damage, supply volume and transportation capacity limitations of each mode. Out-of-stock penalty costs are set to maximize the interests of the ordering party.

[0097] The cost of the domestic transportation stage is affected by the three factors mentioned above, and its components include: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the cost of multimodal transport of bulk and containers, the shortage penalty cost, the variable cost of bulk-to-container conversion, environmental costs and port inventory costs.

[0098] like Figure 2 As shown in the figure, a multimodal transport network optimization method considering demand uncertainty and bulk-to-bulk transport is provided, including:

[0099] Step 100: Build a deterministic model:

[0100] Step 200: constructing a problem model according to the deterministic model by using a stochastic optimization method;

[0101] Step 300: Solve the problem model based on an accurate algorithm to obtain an optimization solution for the multimodal transport network.

[0102] When optimizing the multimodal transport network, we set some basic assumptions as follows:

[0103] 1. Assume that international transportation only involves sea transportation, and domestic transportation involves multimodal transport.

[0104] 2. It is assumed that there is sufficient supply of containers in the port and the empty container transportation process is not considered.

[0105] 3. Assuming that all domestic transportation can be completed within one cycle, international transportation may involve cross-cycle transportation due to the long distance of international transportation.

[0106] 4. Domestic cargo is assumed to be stored only at transit ports.

[0107] There is cargo damage in bulk cargo transportation, but there is no cargo damage in container transportation because the seal is intact and there is no leakage or weathering.

[0108] In a possible implementation, the optimization method also includes a planning period, which includes multiple cycles. Each cycle requires a decision on the purchase of bulk-to-container equipment and a decision on bulk cargo chartering. The supply port and the transit port can only carry out the bulk-to-container process after purchasing the bulk-to-container equipment. Each cycle also requires a decision on the bulk cargo and bulk-to-container freight volume for each mode of transportation in the domestic transportation.

[0109] Figure 3 It is a flowchart of building a deterministic model shown in another exemplary embodiment of the present application.

[0110] In one possible implementation, Figure 3 As shown, the construction of the deterministic model includes:

[0111] Step 110: Define the collection required to build the model.

[0112] Among them, the collection includes:

[0113] S: set of supply ports, s,i∈S;

[0114] K: set of transit ports, k, i∈K;

[0115] P: the set of demand nodes, p,i∈P;

[0116] T: the set of cycle numbers in the entire planning period, cycle t, t'∈T;

[0117] M: the set of domestic transport modes, including inland waterway transport, railway transport, and road transport, m∈M;

[0118] G: The capacity type set of the carrying ship, the type number of the ship is g∈G.

[0119] Step 120: Define the parameters required to build the model.

[0120] The parameters include:

[0121] d pt : The demand of demand node p in period t, p∈P,t∈T,ω∈Ω;

[0122] F it : In period t, the fixed cost of node i to purchase distributed-to-collective equipment, i∈S∪K,t∈T;

[0123] The fixed cost of chartering a bulk carrier with g capacity on the (s, k) route in period t, s∈S, k∈K, t∈T, g∈G;

[0124] R it : The maintenance cost of bulk-to-concentrator equipment at port i in period t, t∈T,i∈S∪K;

[0125] The unit cost of transporting bulk cargo from port s to port k in period t, s∈S, k∈K, t∈T;

[0126] The unit bulk cargo transportation cost from port s to port k in period t, s∈S, k∈K, t∈T;

[0127] The cost of transporting bulk cargo from k to i using m mode of transport in period t, k∈K,i∈K∪P,m∈M,t∈T;

[0128] Unit bulk cargo transportation cost from k to i using transport mode m in period t, k∈K,i∈K∪P,m∈M,t∈T;

[0129] A it : The unit operation cost of converting scattered goods into concentrated goods at port i in period t, i∈S∪K,t∈T;

[0130] In period t, the storage cost of bulk cargo stored in port i is, i∈S∪K,t∈T;

[0131] In period t, the storage cost of converting bulk storage units into bulk storage units at port i is, i∈S∪K,t∈T;

[0132] B pt : The stock-out penalty cost of the p demand node being out of stock in period t, p∈P,t∈T;

[0133] β: cargo damage rate in ocean bulk shipping;

[0134] θ m : The cargo damage rate of bulk cargo transported by domestic transport mode m, m∈M;

[0135] E bul : Unit carbon emission cost of ocean transport of bulk cargo;

[0136] Carbon emission cost of transporting bulk cargo domestically using transport mode m, m∈M;

[0137] E con : Carbon emission costs of converting bulk cargo into container cargo in ocean transport;

[0138] The carbon emission cost of domestically transporting bulk cargoes by transport mode m, m∈M;

[0139] In period t, the capacity of port k, k∈K, t∈T;

[0140] In period t, the maximum supply quantity of the s supply port is k∈K, t∈T;

[0141] Q g : capacity of g-type bulk carrier, g∈G;

[0142] In period t, the ocean transport bulk-to-consolidated capacity from port s to port k is s∈S, k∈K, t∈T;

[0143] The domestic bulk cargo transportation capacity limit for transporting from k to i by mode m in period t is k∈K,i∈K∪P,m∈M,t∈T;

[0144] In domestic period t, the capacity limit for bulk-to-consolidated transportation from k to i using mode m is, k∈K,i∈K∪P,m∈M,t∈T.

[0145] Step 130: Define the decision variables required to build the model.

[0146] The decision variables include:

[0147] z it : 0, 1 variable, whether port i purchases scattered-to-collector equipment in period t, i∈S∪K,t∈T;

[0148] l sktg : 0,1 variable, whether to rent a bulk carrier of g capacity on the (s,k) route in period t, s∈S,k∈K,t∈T,g∈G;

[0149] x skt : continuous variable, the quantity of bulk cargo transported from port s to port k in period t, s∈S,k∈K,t∈T;

[0150] Note: The form of cross-cycle transportation is:

[0151] x kimt : continuous variable, the quantity of bulk cargo transported from k to i by m transportation mode in period t, k∈K,i∈K∪P,m∈M,t∈T;

[0152] y skt : Continuous variable, the number of bulk cargoes converted from port s to port k in ocean transport within period t, s∈S, k∈K, t∈T, Note: Considering the form of cross-period transportation, it is:

[0153] y kimt : Continuous variable, the number of bulk cargo converted to consolidated cargo transported from k to i using m mode of transportation within period t, k∈K,i∈K∪P,m∈M,t∈T.

[0154] Step 140: Define the dependent variables required to build the model.

[0155] Among them, the dependent variables include;

[0156] y kt : continuous variable, the number of k-port scattered groups in t period, k∈K,t∈T;

[0157] Continuous variable, the quantity of bulk cargo stored at port k in period t, k∈K,t∈T;

[0158] Continuous variable, the storage quantity of bulk cargo converted to consolidated cargo at port k in period t, k∈K,t∈T;

[0159] O pt : continuous variable, the number of goods that are not met by the p demand node in the t period, k∈K,t∈T;

[0160] Continuous variable, the number of bulk cargoes arriving at port k in period t, k∈K,t∈T;

[0161] Continuous variable, the number of bulk cargoes leaving port k in period t, k∈K,t∈T;

[0162] Continuous variable, the number of bulk cargoes arriving at port k in period t, k∈K,t∈T;

[0163] Continuous variable: the number of bulk cargoes that leave port k and are converted into consolidated cargoes within period t, k∈K,t∈T.

[0164] Step 150: Obtain the objective function of the deterministic model according to the set, parameters, decision variables and dependent variables.

[0165] Among them, the objective function is:

[0166]

[0167] In a possible implementation, the objective function of the deterministic model includes:

[0168] Equipment purchase and maintenance costs, intermodal transport costs of bulk cargo and containers, bulk cargo chartering costs, labor costs for converting bulk cargo to container cargo, warehousing costs and environmental costs.

[0169] Figure 8 is a flow chart of bulk cargo and containerized cargo in a bulk-to-container operation shown in another exemplary embodiment of the present application. Fig. 9 is a flow chart of bulk cargo and containers in a transshipment port shown in another exemplary embodiment of the present application, Fig.10 It is a schematic diagram of the time points and relationships of each cycle decision within the planning period, shown in another exemplary embodiment of the present application.

[0170] In a possible implementation, the objective function of the deterministic model further includes:

[0171] Out-of-stock penalty constraints, unmet freight demand constraints, inventory balance constraints, bulk cargo and bulk-to-container cargo constraints and cargo supply limit constraints, port capacity limit constraints, bulk carrier capacity limit constraints, and port bulk-to-container restriction constraints.

[0172] The constraint formula is:

[0173]

[0174]

[0175] Constraint (1) indicates that the demand of node p in each period t should be met as much as possible if capacity permits, otherwise there will be a shortage penalty.

[0176] Constraint (2) states that there is no unmet freight demand in the network before the first cycle.

[0177] Constraints (3-4) are inventory balance constraints. Faced with the cargo damage rate and the capacity constraints of domestic transshipment methods (road transport, rail transport, inland waterway transport), it may happen that the current cargo cannot be fully transported. These goods will be left at the port for short-term storage, resulting in inventory costs. Constraints (3-4) also represent the flow of cargo within the port when some bulk cargo is converted to bulk cargo, which will be explained in detail below. Constraints (5-6) indicate that there is no inventory in the first period.

[0178] like Figure 8 and Fig. 9 As shown, taking the transshipment port k as an example, the cargo flow when the port is converted from bulk to container is introduced in detail. represents the bulk cargo volume arriving at the k transit port in period t, It represents the quantity of bulk cargo converted to consolidated cargo arriving at transit port k in period t. There is y kt The goods are converted from bulk to container and then transported through and leave.

[0179] Constraints (7)-(10) represent the bulk cargo and bulk-to-container cargo arriving and departing from port k in period t, respectively. Meaning is The quantity of bulk cargo shipped from the supply port s to the transit port k in the cycle, where It indicates the sailing time from port s to port k. This setting takes into account the sailing time of ocean transportation, which may lead to cross-period transportation if the goods cannot be transported in the current cycle. Constraint (11-12) indicates that there is no cargo transportation before the first cycle.

[0180] Constraint (13) indicates that in period t, the cargo supply at port s cannot exceed the upper limit Constraint (14) represents the capacity limit of port k. Constraint (15) represents that the volume of cargo on the bulk ocean route cannot exceed the capacity limit of the bulk carrier.

[0181] like Fig.10 As shown in Figure 1, the entire planning period includes |T| cycles. Whether the scattered-to-collection conversion can be performed in each cycle depends on whether the scattered-to-collection conversion equipment has been purchased in the previous cycle. Only if the scattered-to-collection conversion equipment has been purchased in the current cycle can the scattered-to-collection conversion be performed in each subsequent cycle.

[0182] In formulas (16)-(18), z st The decision variable is 0 or 1, indicating whether port s will purchase bulk-to-collector equipment in period t. indicates whether the distributed-to-collective equipment was purchased before period t. If it was purchased, the sum is 1. Then the right side of constraint (16) is Indicates that port S can convert bulk into bulk, with the upper limit being the supply limit represents the supply quantity of the supply port s in period t. Constraint (17) similarly indicates whether the transit port k can convert bulk to bulk depends on whether the bulk to bulk equipment was purchased in the previous period. Constraint (18) limits the purchase of equipment to once during the entire planning period.

[0183] Constraints (21-23) are the ranges of x and y under the arc capacity constraints.

[0184] However, considering that the number of demands per cycle in reality is mostly uncertain, in order to be more realistic, we expand the above deterministic model into a model that considers the uncertainty of demand. The random optimization method is used to solve it. Therefore, the optimization problem of the multimodal transport network considering uncertainty is summarized as follows. Its solution method can be extended to all two-stage multi-period mixed integer programming problems considering uncertainty.

[0185] Figure 4 It is a flowchart of constructing a problem model shown in another exemplary embodiment of the present application.

[0186] In one possible implementation, Figure 4 As shown, the problem model is constructed according to the deterministic model by using a random optimization method, including:

[0187] Step 210: Define main problem parameters.

[0188] The main problem parameters include:

[0189] d ptw : In the ω scenario, the demand of the p demand node in period t, p∈P,t∈T,ω∈Ω

[0190] F it : In period t, the fixed cost of node i to purchase distributed-to-collective equipment, i∈S∪K,t∈T

[0191] The fixed cost of chartering a bulk carrier with g capacity on the (s, k) route in period t is: s∈S, k∈K, t∈T, g∈G

[0192] R it : The maintenance cost of bulk-to-concentrator equipment at port i in period t, t∈T,i∈S∪K.

[0193] Step 220: Define sub-problem parameters.

[0194] Among them, the sub-question parameters include:

[0195] Under the ω demand scenario, the unit bulk-to-container transportation cost from port s to port k in period t is s∈S,k∈K,t∈T,ω∈Ω;

[0196] Under the ω demand scenario, the unit bulk cargo transportation cost from port s to port k in period t is s∈S,k∈K,t∈T,ω∈Ω;

[0197] In the ω demand scenario, the unit bulk-to-contained cargo transportation cost from k to i using m transportation mode in period t is k∈K, i∈K∪P, m∈M, t∈T, ω∈Ω;

[0198] Under the ω demand scenario, the unit bulk transportation cost from k to i using m transportation mode in period t, k∈K, i∈K∪P, m∈M, t∈T, ω∈Ω;

[0199] A itω : Under the ω demand scenario, in period t, the unit operation cost of converting scattered data into centralized data at port i, i∈S∪K,t∈T,ω∈Ω;

[0200] In the ω demand scenario, the storage cost of storing unit bulk cargo at port i in period t is, i∈S∪K,t∈T,ω∈Ω;

[0201] In the ω demand scenario, in the t cycle, the storage cost of converting bulk storage units into bulk storage units at port i is, i∈S∪K,t∈T,ω∈Ω;

[0202] B pt : The stock-out penalty cost of the p demand node being out of stock in period t, p∈P,t∈T;

[0203] β: cargo damage rate in ocean bulk shipping;

[0204] θ m : The cargo damage rate of bulk cargo transported by domestic transport mode m, m∈M;

[0205] E bul : Unit carbon emission cost of ocean transport of bulk cargo;

[0206] Carbon emission cost of transporting bulk cargo domestically using transport mode m, m∈M;

[0207] E con : Carbon emission costs of converting bulk cargo into container cargo in ocean transport;

[0208] The carbon emission cost of domestically transporting bulk cargoes by transport mode m, m∈M;

[0209] In period t, the capacity of port k, k∈K, t∈T;

[0210] In period t, the maximum supply quantity of the supply port s, k∈K, t∈T

[0211] Q g : capacity of g-type bulk carrier, g∈G;

[0212] In period t, the ocean transport bulk-to-consolidated capacity from port s to port k is s∈S, k∈K, t∈T;

[0213] The domestic bulk cargo transportation capacity limit for transporting from k to i by mode m in period t is k∈K,i∈K∪P,m∈M,t∈T;

[0214] In domestic period t, the capacity limit for bulk-to-consolidated transportation from k to i using mode m is, k∈K,i∈K∪P,m∈M,t∈T.

[0215] Step 230: Define the main problem decision variables.

[0216] Among them, the main problem decision variables include:

[0217] z it : 0, 1 variable, whether port i purchases distributed-to-collective equipment in period t, i∈S∪K,t∈T

[0218] l sktg : 0,1 variable, whether to charter a bulk carrier of g capacity on the (s,k) route in period t, s∈S, k∈K, t∈T, g∈G.

[0219] Step 240: Define sub-problem decision variables.

[0220] Among them, the sub-problem decision variables include:

[0221] x sktω : Continuous variable, under the ω demand scenario, the quantity of bulk cargo transported from port s to port k in period t, s∈S,k∈K,t∈T,ω∈Ω, Note: Considering the form of cross-period transportation, it is:

[0222] x kimtω : Continuous variable, under the ω demand scenario, the quantity of bulk cargo transported from k to i using m transportation methods within t period, k∈K,i∈K∪P,m∈M,t∈T,ω∈Ω

[0223] y sktω : Continuous variable, under the demand scenario of ω, the number of bulk cargoes converted from port s to port k in ocean transport within period t, s∈S,k∈K,t∈T,ω∈Ω, Note: Considering the form of cross-period transportation, it is:

[0224] y kimtω : Continuous variable. Under the demand scenario ω, within period t, the number of bulk cargoes transported from k to i using transportation mode m, k∈K,i∈K∪P,m∈M,t∈T,ω∈Ω.

[0225] Step 250: Define the dependent variable.

[0226] The dependent variables include:

[0227] y ktω : Continuous variable, under the ω demand scenario, within the t period, the number of k-port scattered changes to sets, k∈K,t∈T,ω∈Ω

[0228] Continuous variable, under the ω demand scenario, the amount of bulk cargo stored at port k in period t, k∈K,t∈T,ω∈Ω Continuous variable, under the ω demand scenario, within the t period, the number of bulk cargo storage at port k, k∈K,t∈T,ω∈Ω

[0229] O ptω : Continuous variable, the number of goods that are not met at p demand node within t period under ω demand scenario, k∈K,t∈T,ω∈Ω

[0230] Continuous variable, under the ω demand scenario, the number of bulk cargoes arriving at port k within period t, k∈K,t∈T,ω∈Ω

[0231] Continuous variable, under the ω demand scenario, the number of bulk cargoes leaving port k in period t, k∈K,t∈T,ω∈Ω

[0232] Continuous variable, under the ω demand scenario, the number of bulk cargoes arriving at port k in period t, k∈K,t∈T,ω∈Ω

[0233] Continuous variable: under the ω demand scenario, the number of bulk cargoes converted to consolidated cargoes at port k within period t, k∈K,t∈T,ω∈Ω.

[0234] Step 260: Obtain the objective function of the main problem model and the objective function of the sub-problem model according to the main problem parameters, sub-problem parameters, main problem decision variables, sub-problem decision variables and dependent variables.

[0235] Among them, the objective function of the main problem model is:

[0236]

[0237] The objective function of the main problem model includes the fixed cost of equipment purchase for the bulk-to-collector network and the cost of chartering bulk cargo ships. Since demand uncertainty mainly affects short-term decisions, the sub-problem part takes demand uncertainty into account, which is also a common way to deal with two-stage problems. We then used a stochastic optimization method to represent demand uncertainty with many deterministic scenarios and solve the sub-problem objective function by taking the average.

[0238] Constraints include:

[0239]

[0240] The objective function of the subproblem model is:

[0241]

[0242] The objective function of the sub-problem model includes the sum of the intermodal transport costs of bulk cargo and containers, bulk cargo chartering costs, labor costs for bulk-to-container conversion, storage costs, and environmental costs under each demand scenario.

[0243] Constraints include:

[0244]

[0245]

[0246] It can be seen that since this problem is a mixed integer programming problem, it contains a large number of 0, 1 decision variables and continuous variables. If there are only continuous variables, linear programming (LP) technology can be used to solve it. With the current commercial solvers such as (Cplex, Lingo, Groubi) The solution capability can be solved in a very short time, but 0, 1 variables are discrete and cannot be expressed by continuous equations. However, there are also a large number of corresponding algorithms for binary variables and integer variables (such as (branch and bound method, cutting plane and branch and cut method, etc.), so we separate the two types of variables into main problems and sub-problems when modeling, which perfectly conforms to the precise algorithm structure of Benders Decomposition.

[0247] Exact algorithms are a type of algorithm that has been less studied before. Compared with the heuristic algorithms (ant colony optimization, genetic algorithms, greedy algorithms, etc.) that have been extensively studied in recent years, exact algorithms ensure the best global solution by ensuring that all possible solutions are systematically evaluated. Heuristic algorithms do not try to evaluate all possible solutions. Instead, they use practical or experience-based methods to quickly find an approximately satisfactory solution to a complex problem.

[0248] The complete structure of the model has been introduced above and will not be repeated here. In this embodiment, we change the model to a compact form to facilitate the description of the algorithm solution process below.

[0249] First, we denote vectors l and z as binary decision variables in the main problem model and vectors as a continuous decision variable.

[0250] Therefore, we have the following simple formula:

[0251]

[0252] st

[0253] Ez≤F(46)

[0254] z,l∈{0,1}(47)

[0255] For each demand scenario ω∈Ω there are the following SPs:

[0256] C(l,z,d(ω))=minGx(ω)+Hy(ω)(48)

[0257] st

[0258] Ix(ω)+Jy(ω)≤K+Sz+Ol(49)

[0259]

[0260] We summarize the standard Benders procedure and describe its improved implementation for the current problem. The Benders algorithm is designed to solve large-scale problems or complex models that cannot be solved as a whole. The overall idea of ​​the algorithm can be summarized as decomposing the model into multiple small-scale problems and solving them iteratively.

[0261] Figure 5 It is a flowchart showing another exemplary embodiment of the present application for solving a problem model through an accurate algorithm.

[0262] In one possible implementation, Figure 5 As shown, the problem model is solved based on an accurate algorithm to obtain an optimization solution for the multimodal transport network, including:

[0263] Step 310: The first iteration requires initializing the binary decision variables in the main problem model and setting the upper and lower limits of the objective function.

[0264] Among them, the binary decisions in the main problem model are z and l, the main problem model has a feasible solution, and the upper and lower limits of the objective function are set.

[0265] Step 320: In each scenario ω, the feasible solution of the main problem model is brought into solving the sub-problem.

[0266] Step 330: DSP solution for each sub-problem in each scenario ω , if the sub-problem model has an optimal solution, the optimal cut is generated and added to the constraints of the main problem model, otherwise a feasible cut is generated and added to the constraints of the main problem model.

[0267] Step 340: Solve the main problem model after one iteration. If the upper and lower limits are equal, the algorithm terminates and an optimization solution is obtained.

[0268] Step 350: Otherwise, return to step 330 to continue solving DSP ω To add constraints, know that all DSP ω All problems are solved once, and then return to step 320 for solution.

[0269] The expression formula is as follows:

[0270]

[0271] Ez≤F(52)

[0272] z,l∈{0,1}(53)

[0273] δ≥0(54)

[0274] The main problem model represents the main problem of relaxation, that is, only some constraints of the original problem are retained for solution. The current problem is the initial relaxation problem. δ represents the objective function value of the sub-problem model under the ω scenario, so we can get the value of the binary variables z', l' and δ' in this iteration. Each scenario ω has a sub-problem model as follows:

[0275] (SP ω (z',l'))minGx(ω)+Hy(ω)(55)

[0276] Ix(ω)+Jy(ω)≤K+Sz'+Ol'(56)

[0277]

[0278] However, we observe the above formula and find that l′ and z′ are still in the constraints. As mentioned before, the presence of 0, 1 variables and continuous variables will seriously drag down the solution time. Therefore, in order to solve efficiently, the dual problem of finding the subproblem model is as follows:

[0279] (DSP ω (z',l'))max(K+Sz'+Ol')u(ω)(58)

[0280] (ω)I≤p(59)

[0281] u(ω)J≤q(60)

[0282] u(ω)≥0(61)

[0283] For DSP ω The solution to will produce different cuts:

[0284] ⑴If DSP ω There exists a solution such that (K+Sz'+Ol')u(ω)>0, then SP ω No feasible solution, add Feasibility cut into the main problem model constraints

[0285] ⑵If DSP ω There is an optimal solution Then add the optimal cut To the main problem model

[0286] Fig.11 It is a schematic diagram of a pseudo code of a Multicut Benders Decomposition algorithm shown in another exemplary embodiment of the present application.

[0287] It can be seen that because the exact algorithm solves the global optimal solution, the calculation efficiency will be relatively slow. For this model, there may be problems such as a large number of iterations and a large solution space. For example, this problem has problems such as a large number of demand scenarios, resulting in a large number of solutions, an excessively large solution space, and repeated constraint additions.

[0288] Then our improvement method focuses on steps 320-350. It can be found that step 320 will bring the solution of the main problem model into the DSP solution under all ω, and then generate the corresponding cuts in step 330. After that, these cuts will be added to the main problem model through continuous iterative calculations for solution. This will result in multiple iterations and seriously increase the calculation time. However, the final result of this process is to put all DSP ω The generated cuts are put into the main problem model to generate solutions, so the multi-cut algorithm is based on this idea and directly puts the cuts of all sub-problems into the main problem model at the same time, reducing the number of iterations and quickly narrowing the solution space.

[0289] The pseudo code of Multicut Benders Decomposition algorithm is as follows Fig.11 shown.

[0290] Here, ε is a very small error value we set to determine whether the optimal solution is found, that is, UB-LB<ε, and ε is generally set to 0.01.

[0291] Some embodiments of the present application mainly solve two problems:

[0292] Firstly, it solves the limitation of the current bulk cargo multimodal transport that lacks consideration of bulk-to-container transport, fills the gap in the research on bulk-to-container transport, and takes into account the multi-cycle and demand uncertainty as well as the cross-cycle transport that may occur in ocean shipping, which is more in line with reality.

[0293] Secondly, in view of the fact that the existing mathematical programming models and heuristic algorithms cannot give the exact optimal solution to the problem under large-scale examples in a short time, after splitting the problem, an improved Benders Decomposition exact algorithm is proposed. That is, a mixed integer programming model for solving the main problem and an exact algorithm for solving the sub-problems. It can give the exact optimal solution to the problem under large-scale examples in a very short time.

[0294] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the instructions, these steps are not necessarily executed in sequence according to the order of the instructions. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the above-mentioned embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0295] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0296] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A multimodal transport network optimization method considering demand uncertainty and bulk-to-bulk transport, characterized in that: The intermodal network includes nodes and transportation processes; The nodes include supply ports, transit ports and demand nodes. The supply ports are large international ports, the transit ports are national coastal ports, and the demand nodes are inland factories. There is a bulk-to-container process at the supply ports and the transit ports. The transportation process includes international transportation and domestic transportation. The international transportation includes bulk cargo loading and container transportation. The domestic transportation includes transportation between transit ports and transportation from transit ports to demand nodes. The international transportation includes ocean transportation. The domestic transportation considers multimodal transportation. The domestic transportation methods include inland waterway transportation, railway transportation and road transportation. The costs of the international transportation stage include: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the cost of chartering bulk carriers, the transportation cost of bulk cargoes and containers, the variable cost of bulk-to-container conversion and the environmental cost; The costs of the domestic transportation stage include: the purchase cost of bulk-to-container equipment and the maintenance cost per cycle, the intermodal transportation cost of bulk and container, the penalty cost of shortage, the variable cost of bulk-to-container, the environmental cost and the port inventory cost; Optimization methods include: Construct a deterministic model, define the set required for constructing the model, define the parameters required for constructing the model, define the decision variables required for constructing the model, define the dependent variables required for constructing the model, and obtain the objective function of the deterministic model based on the set, parameters, decision variables and dependent variables; The set includes S: the supply port set, s, i∈S; K: the transit port set, k, i∈K; P: the demand node set, p, i∈P; T: the cycle number set in the entire planning period, cycle t, t , ∈T; M: domestic transport mode set, inland waterway transport, railway transport, road transport, m∈M; G: capacity type set of carrier ships, ship type number g∈G; The parameters include: it : In period t, the fixed cost of purchasing bulk-to-collector equipment for port node i, i∈S∪K,t∈T; The fixed cost of chartering a bulk carrier with g capacity on the (s, k) route in period t, s∈S, k∈K, t∈T, g∈G; R it : The maintenance cost of bulk-to-concentrator equipment at port i in period t, t∈T,i∈S∪K; The unit cost of transporting bulk cargo from port s to port k in period t, s∈S, k∈K, t∈T; The unit bulk cargo transportation cost from port s to port k in period t, s∈S, k∈K, t∈T; The cost of transporting bulk cargo from k to i using m mode of transport in period t, k∈K,i∈K∪P,m∈M,t∈T; The unit bulk cargo transportation cost from k to i in period t using mode m, k∈K,i∈K∪P,m∈M,t∈T; A it : The unit operation cost of converting scattered goods into concentrated goods at port i in period t, i∈S∪K,t∈T; In period t, the storage cost of bulk cargo stored in port i is, i∈S∪K,t∈T; In period t, the storage cost of converting bulk storage units into bulk storage units at port i is i∈S∪K,t∈T; E bul : Unit carbon emission cost of ocean transport of bulk cargo; Carbon emission cost of transporting bulk cargo by domestic transport mode m, m∈M; E con : Carbon emission costs of converting bulk cargo into container cargo in ocean transport; The carbon emission cost of domestically transporting bulk cargoes by transport mode m, m∈M; The decision variables include z it : 0, 1 variable, whether port i purchases distributed-to-collective equipment within period t, i∈S∪K, t∈T; l sktg : 0,1 variable, whether to rent a bulk carrier of g capacity on the (s,k) route in period t, s∈S, k∈K, t∈T, g∈G; x skt : Continuous variable, the quantity of bulk cargo transported from port s to port k in period t, s∈S, k∈K, t∈T; Note: Considering the form of cross-period transportation, it is: x kimt : continuous variable, the quantity of bulk cargo transported from k to i by m transportation mode in period t, k∈K,i∈K∪P,m∈M,t∈T; y skt : Continuous variable, the number of bulk cargoes converted from port s to port k in ocean transport within period t, s∈S, k∈K, t∈T, Note: Considering the form of cross-period transportation, it is: y kimt : continuous variable, the number of bulk cargo transported from k to i by m transportation mode in period t, k∈K,i∈K∪P,m∈M,t∈T; The dependent variables include y kt : continuous variable, the number of k-port scattered groups in t period, k∈K,t∈T; Continuous variable, the quantity of bulk cargo stored at port k in period t, k∈K,t∈T; Continuous variable, the storage quantity of bulk cargo converted to consolidated cargo at port k in period t, k∈K,t∈T; The objective function of the deterministic model is: By means of a stochastic optimization method, a problem model is constructed according to the deterministic model, main problem parameters are defined, sub-problem parameters are defined, main problem decision variables are defined, sub-problem decision variables are defined, and dependent variables are defined, and an objective function of the main problem model and an objective function of the sub-problem model are obtained according to the main problem parameters, sub-problem parameters, main problem decision variables, sub-problem decision variables and dependent variables; Among them, the main problem parameters include d ptω : In the ω scenario, the demand of the p demand node in period t, p∈P,t∈T,ω∈Ω; F it : In period t, the fixed cost of purchasing bulk-to-container equipment at port i, i∈S∪K,t∈T; The fixed cost of chartering a bulk carrier with g capacity on the (s, k) route in period t, s∈S, k∈K, t∈T, g∈G; R it : The maintenance cost of bulk-to-concentrator equipment at port i in period t, t∈T,i∈S∪K; The sub-problem parameters include Under the ω demand scenario, the unit bulk-to-container transportation cost from port s to port k in period t is s∈S,k∈K,t∈T,ω∈Ω; Under the ω demand scenario, the unit bulk cargo transportation cost from port s to port k in period t is s∈S,k∈K,t∈T,ω∈Ω; In the ω demand scenario, the unit bulk-to-contained cargo transportation cost from k to i using m transportation mode in period t, k∈K, i∈K∪P, m∈M, t∈T, ω∈Ω; Under the ω demand scenario, the unit bulk cargo transportation cost from k to i using m transportation methods in period t, k∈K,i∈K∪P,m∈M,t∈T,ω∈Ω; A itω : Under the ω demand scenario, in period t, the unit operation cost of converting scattered data into centralized data at port i, i∈S∪K,t∈T,ω∈Ω; In the ω demand scenario, the storage cost of storing unit bulk cargo at port i in period t is, i∈S∪K,t∈T,ω∈Ω; In the ω demand scenario, in the t cycle, the storage cost of converting bulk storage units into bulk storage units at port i is, i∈S∪K,t∈T,ω∈Ω; E bul : Unit carbon emission cost of ocean transport of bulk cargo; Carbon emission cost of transporting bulk cargo by domestic transport mode m, m∈M; E con : Carbon emission costs of converting bulk cargo into container cargo in ocean transport; The carbon emission cost of domestically transporting bulk cargoes by transport mode m, m∈M; The main problem decision variables include z it : 0, 1 variable, whether port i purchases distributed-to-collective equipment in period t, i∈S∪K, t∈T; l sktg : 0,1 variable, whether to rent a bulk carrier of g capacity on the (s,k) route in period t, s∈S, k∈K, t∈T, g∈G; The sub-problem decision variables include x sktω : Continuous variable, under the ω demand scenario, the quantity of bulk cargo transported from port s to port k in period t, s∈S,k∈K,t∈T,ω∈Ω, Note: Considering the form of cross-period transportation is: x kimtω : continuous variable, under the ω demand scenario, the quantity of bulk cargo transported from k to i using m transportation mode within t period, k∈K,i∈K∪P,m∈M,t∈T,ω∈Ω; y sktω : Continuous variable, under the ω demand scenario, within the t period, the number of bulk cargoes converted from s to k ports in ocean transport, s∈S, k∈K, t∈T, ω∈Ω, Note: Considering the form of cross-period transportation, it is: y kimtω : continuous variable, under the demand scenario ω, within the t period, the number of bulk cargo transported from k to i by m transportation mode, k∈K,i∈K∪P,m∈M,t∈T,ω∈Ω; The dependent variables include y ktω : continuous variable, under the ω demand scenario, within the t period, the number of k-port scattered changes, k∈K,t∈T,ω∈Ω; Continuous variable, the quantity of bulk cargo stored at port k in period t under the ω demand scenario, k∈K,t∈T,ω∈Ω; Continuous variable, under the ω demand scenario, the storage quantity of bulk cargo converted to consolidated cargo at port k in period t, k∈K,t∈T,ω∈Ω; The objective function of the main problem model is: Solve the problem model based on the precise algorithm to obtain the optimization solution of the multimodal transport network. The first iteration needs to initialize the binary decision variables in the main problem model and set the upper and lower limits of the objective function. The objective function of the subproblem model is: In each scenario, the feasible solution of the main problem model is brought into solving the sub-problem; For the solution of the sub-problem in each scenario, if the sub-problem model has an optimal solution, the optimal cut is generated and added to the constraints of the main problem model; otherwise, a feasible cut is generated and added to the constraints of the main problem model; Solve the main problem model after one iteration. If the upper and lower limits are equal, the algorithm terminates and the optimization solution is obtained. Otherwise, continue to solve the sub-problems in each scenario. If the sub-problem model has an optimal solution, generate the optimal cut and add it to the constraints of the main problem model. Otherwise, generate a feasible cut and add it to the constraints of the main problem model. This process continues until all sub-problem models have been solved. Then go back to each scenario and bring the feasible solution of the main problem model into solving the sub-problems.

2. The multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport as claimed in claim 1, characterized in that: The optimization method also includes a planning period, which includes multiple cycles. Each cycle requires decisions on the purchase of bulk-to-container equipment and bulk cargo chartering. The supply port and the transit port can only carry out the bulk-to-container process after purchasing the bulk-to-container equipment. Each cycle also requires decisions on the bulk cargo and bulk-to-container freight volumes for each mode of transportation in the domestic transportation.

3. The multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport as claimed in claim 1, characterized in that: The objective function of the deterministic model includes: Equipment purchase and maintenance costs, intermodal transport costs of bulk cargo and containers, bulk cargo chartering costs, labor costs for converting bulk cargo to container cargo, warehousing costs and environmental costs.

4. The multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport as claimed in claim 1, characterized in that: The objective function of the deterministic model also includes: Out-of-stock penalty constraints, unmet freight demand constraints, inventory balance constraints, bulk cargo and bulk-to-container cargo constraints and cargo supply limit constraints, port capacity limit constraints, bulk carrier capacity limit constraints, and port bulk-to-container restriction constraints.

5. The multimodal transport network optimization method considering demand uncertainty and scattered-to-centralized transport as claimed in claim 1, characterized in that: The method of solving the problem model based on an accurate algorithm to obtain an optimization solution for the multimodal transport network also includes: Put all the sub-problems into the main problem model at the same time to reduce the number of iterations and quickly narrow down the solution space.

Images