River-sea combined transportation optimization method for large goods

Abstract

The invention relates to a large cargo river-sea combined transportation optimization method. The method comprises the following steps: S1, constructing a large cargo river-sea combined transportation space-time dynamic hyper-network model; s2, an objective function of the model is constructed, and the objective function aims at minimizing the sum of the transportation time, the transportation cost, the carbon emission and the safety risk; s3, constraint conditions of the model are generated, and the constraint conditions comprise a flow conservation constraint, a transport capacity limitation constraint, a channel transportability constraint, a node loading and unloading capacity constraint, a transfer time window constraint, a transfer frequency constraint, a safety redundancy transport time constraint and a carbon emission limitation constraint; and S4, solving the model by adopting a two-dimensional coding improved genetic algorithm of a partial matching cross strategy to obtain an optimal transportation scheme of river-sea combined transportation of the large goods. Compared with the prior art, the method provided by the invention has the advantages of providing an optimized decision scheme with economy, safety and environmental protection for the complex intermodal network and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of transportation, in particular to a large cargo river-sea intermodal optimization method. BACKGROUND

[0002] In the field of logistics transportation involving port group intermodal transportation, especially in the comprehensive transportation network with multiple port forms such as sea ports, river ports and river ports, the existing technology faces a series of bottleneck problems affecting overall efficiency and sustainable development. The current network presents the typical characteristics of "scattered port points and fragmented hinterland delivery", and each main port area mainly relies on its own independent port access channel to realize single connection with the hinterland. This scattered layout and operation mode leads to a lack of effective coordination and connection between port areas, and there is significant overlap in functional positioning, for example, two sea ports both undertake part of the grain or chemical logistics functions, resulting in repeated allocation of resources and internal competition.

[0003] At the same time, the existing intermodal network also has deficiencies in infrastructure connection and transportation mode coordination. The construction density of the railway branch lines matched with some ports in the region is insufficient, and the transport capacity is limited, which cannot effectively share the pressure of river transportation, resulting in excessive load on the key transportation channels, and the overall transportation capacity of the region is facing severe challenges. When goods need to be transported from the starting block to the terminal block through multiple transit city nodes, due to the different transportation modes such as river transportation and river transportation between adjacent nodes, and the significant differences in key indicators such as transportation distance, transportation time, unit transportation cost and carbon emissions between different modes, the existing technology cannot realize reasonable path allocation and transportation mode combination optimization of freight volume. This situation directly restricts the improvement of the overall efficiency of the network, making the regional logistics face great challenges in reducing the overall transportation cost, controlling the total carbon emissions and improving the intermodal efficiency (reducing cost, improving quality and increasing efficiency), and a systematic network optimization method is urgently needed to solve the problem. SUMMARY

[0004] The purpose of the present application is to provide a large cargo river-sea intermodal optimization method.

[0005] The purpose of the present application can be achieved by the following technical solutions: A large cargo river-sea intermodal optimization method, comprising: Step S1: constructing a large cargo river-sea intermodal space-time dynamic super network model , wherein V is a node layer composed of all ports, T is discrete time, is an edge set, E m is a transportation path m of transportation mode, and M is the number of transportation modes; Step S2: Construct the objective function of the model, wherein the objective function aims to minimize the sum of transportation time, transportation cost, carbon emissions, and safety risks; Step S3: Generate the constraints of the model, wherein the constraints include flow conservation constraints, transport capacity limitation constraints, corridor suitability constraints, node loading and unloading capacity constraints, transshipment time window constraints, transshipment frequency constraints, safe redundancy transport time constraints, and carbon emission limitation constraints. Step S4: Use a two-dimensional encoding improved genetic algorithm with a partial matching crossover strategy to solve the model and obtain the optimal transportation scheme for river-sea intermodal transport of large cargo.

[0006] The objective function of step S2 is: in: T For the sub-objective of transportation time, C For the sub-objective of transportation cost, E For carbon emission sub-targets, R For the sub-objective of security risk, α The weighting coefficients for the transportation time sub-objective. The weighting coefficients for the transportation cost sub-objective. For the weighting coefficients of carbon emission sub-targets, The weighting coefficients for the sub-objectives of safety risk; The transportation time sub-target is: Where: D is the total transportation distance of all routes. d ij For nodes i To node j Path distance, Let m be the transit time from mode m to mode n. For path Adopting the first The number of transport days for each mode of transport; The transportation cost sub-objective is: in: For nodes i To node j The unit cost of transportation for the route and mode of transport. for t Time-of-use goods g By transportation m From node i To node j The volume of transportation, For nodesi The transit cost from the transportation mode m to the transportation mode n , p i The port penalty unit price of the node i ; The carbon emission sub-target is: Wherein: The unit carbon emission (kgCO2 / ton·km) of the mode ; The safety risk sub-target is: Wherein: The risk coefficient of the node, The parallel operation capacity of the node , The working time window length of the node , The total processing capacity of the node , The utilization rate of the node resource, if >1, indicating that the task exceeds the node capacity, causing congestion risk, Whether the path from the node i to the node j is suitable for large goods traffic.

[0007] The flow conservation constraint is: Wherein: The transportation volume of the node t to the node g by the transportation mode m in the time period j ; i The transportation capacity limit constraint is: Wherein: The maximum number of shipping batches of the path from the node to the node i by the transportation mode j in unit time, m The maximum carrying capacity of a single vehicle of the path from the node to the node i ; j The channel suitability constraint is: The node loading and unloading capacity constraint is:​ wherein: is a node i by transportation mode m the ability to transfer goods; the transfer time window constraint is: wherein: is a node i from transportation mode m to transportation mode n transfer cost; the transfer number constraint is: wherein: is a node whether it is a transfer node, Z max is the maximum allowed number of transfers; the safety redundancy transportation time constraint is: wherein: T max is the total transportation time of the whole transportation of the large goods from river to sea, which should not exceed the maximum acceptable time, and is is the transportation safety redundancy time; the carbon emission limit constraint is: wherein: is the emission of each path of transportation mode m , E max is the environmental upper limit.

[0008] In the step S4, the two-dimensional coding improved genetic algorithm of the partial matching cross strategy, the coding length of each individual gene is determined by the available path from the starting port to the destination of the task, the coding of each individual gene is composed of multiple transportation task units, each transportation task unit adopts an upper and lower double-layer two-dimensional coding structure, and the coding rule of the double-layer two-dimensional coding structure is: (1) the upper layer is a path selection vector: the path selection vector is expressed by binary coding, 1 represents that the path is selected, and 0 represents that the path is not selected, and each unit includes at least one selectable path; (2) the lower layer is a flow allocation vector: the selected path in each unit is allocated in proportion to the flow, the flow allocation proportion of the path not selected is 0, and the total of the flow allocation proportion in each unit is 1.

[0009] In the step S4, the two-dimensional coding improved genetic algorithm adopting the partial matching crossover strategy is used to solve the model, and the genetic factors are decoded, the decoding method comprising: initializing the genetic factors, then performing dimensionless processing on the model, and using the obtained target function as the fitness function to calculate the target value of each genetic factor.

[0010] The process of performing dimensionless processing on the model comprises the following steps: Step W01: designing task unit and task demand quantity: according to the freight demand of large goods river-sea intermodal transport, designing the task unit and demand quantity data of individual genes; Step W02: performing freight flow distribution optimization: specifically comprising: Step W021: determining path candidate set: generating multiple candidate paths for each task unit according to the shortest path algorithm, or the random path algorithm, or the comprehensive evaluation path algorithm; Step W022: using a normalization method to perform dimensionless processing on the constructed multi-objective path optimization model.

[0011] The normalization method in the step W022 specifically comprises: Step W0221: respectively obtaining the maximum values maxC, maxT, maxE and the minimum values minC, minT, minE; Step W0222: when a feasible solution is obtained by the hybrid algorithm, respectively calculating C, T and E, then using a normalization method to perform linear change on C, T and E, and mapping the results to (0, 1); Step W0223: respectively giving the weight parameters values, and then performing weighted summation on the normalized multi-objective function values.

[0012] The solving method of the two-dimensional coding improved genetic algorithm adopting the partial matching crossover strategy in the step S4 is: Step S4-1: two-dimensional coding of individual genes; Step S4-2: randomly generating an initial population; Step S4-3: calculating the population target value according to the target function; Step S4-4: obtaining the river-sea intermodal transport path and the freight flow distribution result of each path; Step S4-5: performing line capacity and port traffic capacity check; Step S4-6: capacity overrun penalty term calculation; Step S4-7: returning the target value; Step S4-8: judging whether the maximum iteration number is reached, if yes, ending, and if not, entering step S4-9; Step S4-9: sorting according to the target value size, selecting and copying individuals, PMX strategy crossover, random mutation, returning to step S4-4 for the processing result after random mutation, and implementing the solution of the double-layer coding genetic algorithm by cycling.

[0013] A large cargo river-sea intermodal optimization device, comprising a memory, a processor, and a program stored in the memory, wherein the processor implements the method as described above when executing the program.

[0014] A storage medium having a program stored thereon, wherein the program implements the method as described above when executed.

[0015] Compared with the prior art, the present application has the following beneficial effects: 1. Comprehensive optimization of transportation efficiency and cost, by constructing a multi-objective optimization model, simultaneously minimizing transportation cost C , time T , and carbon emissions E , and introducing safety risk cost R as a key decision-making index, realizing efficient integration of transportation resources. Linear weighting method is used to unify the multi-objective function, combined with transfer time window, node loading and unloading capacity and other constraints, to ensure that the comprehensive logistics cost of large cargo is significantly reduced and the overall economic benefit of intermodal network is improved under the premise of meeting the actual operation demand.

[0016] 2. Improve resource utilization efficiency and network synergy, based on space-time dynamic super network modeling, integrate node resources such as seaports, inland terminals, and estuary terminals, solve the resource mismatch problem caused by scattered port locations and repeated functions through flow conservation constraints and transportation capacity restrictions. The optimized freight flow allocation scheme can effectively coordinate the collaborative operation of multiple paths and multiple transportation modes, avoid single channel overload, and significantly improve the carrying capacity and resource utilization rate of regional transportation network.

[0017] 3. Strengthen the safety and controllability of large cargo transportation, design channel suitability constraints (such as bridge and tunnel height limits, channel restrictions), node-specific loading and unloading equipment capacity constraints, and transfer frequency restrictions (such as a maximum of one transfer), and strictly avoid transportation risks. Add safety redundancy transportation time constraints to reserve buffer for unpredictable factors such as weather and equipment failure, ensure timely delivery of goods and reduce damage rate.

[0018] 4. Realize green and low-carbon transportation, through carbon emission limitation constraints, forcibly control the total carbon emissions of each path not to exceed the environmental upper limit, combine with the normalization method to quantify the unit carbon emission cost, guide the model to preferentially select low-carbon transportation paths. This mechanism promotes the green transformation of the river-sea intermodal system and reduces the impact of logistics activities on the environment.

[0019] 5. Realize efficient solution of complex network, adopt two-dimensional coding improved genetic algorithm (PMXIGA) of partial matching crossover strategy, solve NP-hard problem of multimodal transport path planning through joint optimization of upper path selection vector and lower flow allocation vector. Introduce partial matching crossover strategy (PMX), enhance global optimization ability and convergence efficiency of the algorithm, and provide optimal scheme with economy and feasibility for large-scale freight transport demand.

[0020] 6. Realize dynamic supply and demand balance and strategic adaptation, based on service hinterland-destination (task unit) freight demand prediction data, combine with transport route parameters (cost, time, transport capacity, etc.), dynamically generate path candidate set and optimize freight flow allocation proportion. The model can flexibly adapt to port development planning and policy requirements, provide data-driven decision support for long-term layout adjustment of regional intermodal network, and ensure accurate matching of transport supply and demand.

[0021] 7. Through systematic optimization model and intelligent algorithm, significantly improve the operation efficiency, safety level and environmental benefits of large cargo river-sea intermodal transport, and provide core technical support for cost reduction, quality improvement and efficiency increase of multimodal transport network. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 Flow chart of double-layer coding genetic algorithm solution of the present application; Figure 2 Large cargo river-sea intermodal transport network line diagram of a certain coastal multimodal transport port; Figure 3 Large cargo river-sea intermodal transport network line diagram after flow allocation optimization; Figure 4 Figure 3 Large cargo river-sea intermodal transport network flow carrying capacity distribution diagram of each port; Figure 5 Main step flow chart of the method of the present application. DETAILED DESCRIPTION

[0023] The present application will be described in detail below in combination with the drawings and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present application, and detailed implementation mode and specific operation process are given, but the protection scope of the present application is not limited to the following examples.

[0024] The current situation of large cargo river-sea intermodal transport network of the present application is that in the multimodal transport network, the traffic distance, transport time, unit transport price, cargo damage rate and unit carbon emission of different transfer nodes, different transport lines and different transport modes are generally different.

[0025] The present application constructs a large cargo river-sea intermodal transport space-time dynamic super network model, and the following assumptions are made for the large model:​ (1) All shipments are large, bulky goods that can be split or cannot be split into batches; (2) Different routes and different modes of transportation have different traffic capacity, loading capacity and traffic adaptability; (3) The loading capacity of all modes of transportation can always meet the user's freight transportation needs; (4) Transfer can only be carried out at transfer nodes, and at most one change of transportation mode can be carried out at each node; (5) Each transfer point has sufficient transfer capacity, and the waiting time and waiting cost during the transfer process are not considered; (6) The impact of force majeure on the transportation process is not considered, such as severe weather, epidemics, equipment failures, etc.

[0026] Table 1 shows the parameter settings for various transport routes at a certain canal-sea multimodal transport port. Table 1 Figure 2 This is the route map of the river-sea intermodal transport network corresponding to Table 1.

[0027] Table 2 shows the freight demand for river-sea intermodal transport along the canal.

[0028] Table 2 Based on the multimodal transport port routes in Tables 1 and 2 above, and the freight demand for canal-sea intermodal transport, an optimization method for river-sea intermodal transport of oversized cargo is proposed, such as... Figure 5 As shown, it includes: Step S1: Construct a spatiotemporal dynamic hypernetwork model for the intermodal transport of large cargo by river and sea. Where V is the node layer, consisting of all ports, and T is the discrete time. For edge set, E m For transportation methods m The transportation route, M is the number of transportation modes; Step S2: Construct the objective function of the model, wherein the objective function aims to minimize the sum of transportation time, transportation cost, carbon emissions, and safety risks; Step S3: Generate the constraints of the model, wherein the constraints include flow conservation constraints, transport capacity limitation constraints, corridor suitability constraints, node loading and unloading capacity constraints, transshipment time window constraints, transshipment frequency constraints, safe redundancy transport time constraints, and carbon emission limitation constraints. Step S4: Use a two-dimensional encoding improved genetic algorithm with a partial matching crossover strategy to solve the model and obtain the optimal transportation scheme for river-sea intermodal transport of large cargo.

[0029] As shown in Table 3, an individual gene double-layer coding structure is designed, wherein OD pair is called task unit (abbreviation: OD pair), and n represents task number.

[0030] Table 3 The cargo demand of each OD pair (such as the cargo demand from service hinterland 1 to destination 1) can be completed through multiple paths. First, multiple candidate paths need to be generated for each OD pair, and the paths can be connected through different transportation modes (such as sea transportation, inland water transportation, and river transportation). The transportation cost, time, and emission of each path and each transportation mode are comprehensively considered to obtain the final path selection and flow distribution scheme.

[0031] In order to generate a multimodal transport scheme, the genetic factor needs to be decoded. After initializing the genetic factor, the model is dimensionless, and the obtained multi-objective function is used as the fitness function to calculate the target value of each genetic factor. By constantly updating the genetic information of the genetic factor, if the target function value is smaller, the corresponding genetic factor is closer to the optimal solution. At the same time, the PMX crossover strategy and elite selection operation are introduced to increase the diversity of the population and improve the convergence and global optimization ability of the algorithm.

[0032] The dimensionless processing includes the following steps: Step W01: Design OD pair and task demand: according to the large cargo river-sea combined transport freight demand in Table 2, design the OD pair and demand data of the individual gene; Step W02: Perform freight flow distribution optimization: specifically including: Step W021: Determine the path candidate set: generate multiple candidate paths for each OD pair according to the shortest path algorithm, or the random path algorithm, or the comprehensive evaluation path algorithm; Step W022: Use the normalization method to perform dimensionless processing on the constructed multi-objective path optimization model. The normalization method specifically includes: Step W0221: respectively find the maximum values maxC, maxT, maxE and the minimum values minC, minT, minE; Step W0222: When a feasible solution is obtained by the hybrid algorithm, calculate C, T, and E, and then use the normalization method to perform linear transformation on C, T, and E, and map the results to (0, 1), that is: ​Step W0223: give the value of the weight parameter respectively , and then weight the normalized multi-objective function values, that is: to obtain the dimensionless comprehensive index; in the formula: The value of the weight parameter is set according to the user's preference for the weights of cost, time and carbon emissions.

[0033] As shown in FIG. 1, the solution flowchart of the PMXIGA genetic algorithm mainly includes the following steps: Figure 1 Step S041: two-dimensional coding is performed on individual genes; Step S042: an initial population is randomly generated; Step S043: the population target value is calculated according to the target function formula 1; Step S044: the river-sea intermodal transport path and the cargo flow distribution result of each path are obtained; Step S045: line capacity and port traffic capacity check is performed; Step S046: capacity overrun penalty term calculation; Step S047: return the target value; Step S048: determine whether the maximum number of iterations is reached? If yes, end; if not, go to step S049; Step S049: sort according to the target value size, select and copy individuals, PMX strategy crossover, random mutation, and return to step S044 for the processing result after random mutation, and cycle to realize the solution of the double-layer coding genetic algorithm. By using the above method, combined with Table 1 and Table 2, the river-sea intermodal transport path selection and cargo flow distribution optimization result of a certain canal can be obtained, as shown in Table 4.

[0034] Table 4

[0035] As shown in FIG. 2, it is a network line diagram after the network flow distribution optimization of the large cargo river-sea intermodal transport network in Table 4. Figure 3 As can be seen from Table 4 and FIG. 2, after the optimization of the large cargo river-sea intermodal transport model, the best transportation scheme of the transportation path selection from the service hinterland to the destination (i.e. the starting point to the terminal) and the transportation capacity distribution of each path can be given.

[0036] Figure 3

[0037] ​​​​For example, as shown in Table 4, the transportation demand from service hinterland 1 to destination 2 is 1.04 million tons. Through the optimization method of this application, three route selection schemes are given: (1) Service hinterland 1 → seaport 1 → destination 2, the transportation volume allocation ratio on this route is 61.38%; (2) Service hinterland 1 → seaport 4 → destination 2, the transportation volume allocation ratio on this route is 23.65%; (3) Service hinterland 1 → seaport 3 → destination 2, the transportation volume allocation ratio on this route is 14.98%. The sum of the transportation volume allocation ratios is 1.

[0038] like Figure 4 As shown, Figure 3 The map shows the port capacity distribution for medium and heavy cargo river-sea intermodal transport. Port 1 accounts for 43% of the total transport demand, Port 2 for 10.5%, Port 3 for 28%, and Port 4 for 18.5%. This optimized port capacity distribution effectively avoids overloading at any single port and significantly improves network synergy among seaports. Specifically, Port 1's capacity does not exceed its loading capacity limit (see the capacity column in Table 1). Ports 2 and 3 share the pressure on key routes, especially alleviating the high-demand path from Port 1 to the destination. Port 4 provides fault tolerance for unforeseen circumstances, conforming to the redundancy safety mechanism. This allocation mechanism makes the utilization of port resources more balanced and efficient.

[0039] The optimization method described in this embodiment can dynamically coordinate resources across multiple ports, achieving reduced transportation costs, lower carbon emissions, and improved time efficiency while meeting safety and environmental constraints. By dynamically coordinating resources through the dual-layer coding PMXIGA algorithm, it provides an optimized decision-making scheme that combines economy, safety, and environmental protection for complex intermodal transport networks.

[0040] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for optimizing river-sea intermodal transportation of large cargoes, characterized in that, The application relates to a method for solving an optimal transport scheme of large cargo river-sea intermodal transport. Step S1: Constructing a space-time dynamic supernetwork model for intermodal transportation of large cargos between river and sea where V is the node layer consisting of all ports, T is discrete time, is the edge set, E m is the transportation mode m of the transportation path, and M is the number of transportation modes; The method comprises the following steps: Step S2: constructing a target function of the model, wherein the target function aims to minimize the sum of transport time, transport cost, carbon emission and safety risk; Step S3: generating constraint conditions of the model, wherein the constraint conditions comprise a flow conservation constraint, a transport capacity limitation constraint, a channel suitability constraint, a node loading and unloading capacity constraint, a transshipment time window constraint, a transshipment frequency constraint, a safety redundant transport time constraint and a carbon emission limitation constraint; 2. The method according to claim 1, wherein, Step S4: solving the model by using a two-dimensional coding improved genetic algorithm with a partial matching crossover strategy to obtain the optimal transport scheme of the large cargo river-sea intermodal transport. wherein: T is a transportation time sub-target, C is a transportation cost sub-target, E is a carbon emission sub-target, R is a safety risk sub-target, α is a weight coefficient of the transportation time sub-target, is a weight coefficient of the transportation cost sub-target, is a weight coefficient of the carbon emission sub-target, is a weight coefficient of the safety risk sub-target; The target function of the step S2 is as follows: Where: D is the total transportation distance of all routes. d ij For nodes i To node j Path distance, Let m be the transit time from mode m to mode n. For path Adopting the first The number of days for each mode of transport; The transport time sub-target is as follows: in: For nodes i To node j The unit cost of transportation for the route and mode of transport. for t Time-of-use goods g By transportation m From node i To node j The volume of transportation, For nodes i From the mode of transportation m To transportation method n The cost of transshipment p i For nodes i The unit price of the detention penalty in Hong Kong; The transport cost sub-target is as follows: wherein: is the specific carbon emission (kg CO2 / ton·km) of the way . The carbon emission sub-target is as follows: wherein: is the risk coefficient of each node, is the node parallel job capacity, is the node working time window length, is the node total processing capacity, is the usage rate of the node resource, if >1, it indicates that the task exceeds the node capacity, causing congestion risk, is whether the path from node i to node j is suitable for large goods to pass.

3. The method according to claim 2, wherein, The safety risk sub-target is as follows: wherein: is t period goods g by mode of transport m from node j to node i transportation volume; The flow conservation constraint is as follows: in: For path self-node i To node j The route and mode of transportation per unit time m Maximum number of shipment batches, For self-node i To node j The maximum carrying capacity of a single vehicle along the route; The transport capacity limitation constraint is as follows: The channel suitability constraint is as follows: wherein: is a node i by transport means m the ability to transfer goods; The node loading and unloading capacity constraint is as follows: wherein: is a node i from transportation mode m to transportation mode n transfer cost; The transshipment time window constraint is as follows: wherein: is whether the node is a transit node, Z max is the maximum allowed number of transits; The transshipment frequency constraint is as follows: Wherein: T max The total of the overall transportation time of the ocean-river intermodal transportation of large cargo shall not exceed the latest acceptable time, which is, The transportation safety redundancy time; The safety redundant transport time constraint is as follows: wherein: the mode of transport for each route m the emissions for each route, E max is the environmental upper limit.

4. The method according to claim 1, wherein, The carbon emission limitation constraint is as follows: In the step S4, the coding length of each individual gene in the two-dimensional coding improved genetic algorithm with the partial matching crossover strategy is determined by the available path from the starting hinterland to the destination of the task, the coding of each individual gene is composed of multiple transport task units, each transport task unit adopts an upper and lower two-dimensional coding structure, and the coding rule of the two-dimensional coding structure is as follows: (1) the upper layer is a path selection vector: the path selection vector is expressed by binary coding, 1 represents that a path is selected, and 0 represents that a path is not selected, and each unit includes at least one selectable path; 5. The method according to claim 4, wherein, (2) the lower layer is a flow allocation vector: flow proportion allocation is carried out on the selected paths in each unit, the flow allocation proportion of the paths not selected is 0, and the flow allocation proportion sum in each unit is 1.

6. The method according to claim 5, wherein, In the step S4, when the two-dimensional coding improved genetic algorithm with the partial matching crossover strategy is used to solve the model, the genetic factors are decoded, and the decoding method comprises the following steps: the genetic factors are initialized, the model is subjected to dimensionless processing, the obtained target function is used as an adaptability function, and the target value of each genetic factor is calculated. The dimensionless processing process of the model comprises the following steps: Step W01: designing a task unit and task demand: according to the cargo transport demand of the large cargo river-sea intermodal transport, a task unit and demand data of an individual gene are designed; Step W02: carrying out cargo transport flow allocation optimization: specifically comprising the following steps: Step W021: determining a path candidate set: a plurality of candidate paths are generated for each task unit according to a shortest path algorithm, a random path algorithm or a comprehensive evaluation path algorithm; 7. The method according to claim 6, wherein, The normalization method in the step W022 specifically comprises the following steps: Step W0221: respectively calculating a maximum value maxC, maxT, maxE and a minimum value minC, minT, minE; Step W0222: When the hybrid algorithm obtains a feasible solution, C, T and E are calculated respectively, and then normalized to linearly change C, T and E and map the results to (0, 1); Step W0223: give the weight parameters values, and then weight-sum the normalized multi-objective function values.

8. The method according to claim 4, wherein, In the step S4, the solving method of the two-dimensional coding improved genetic algorithm with partial matching crossover strategy is as follows: Step S4-1: two-dimensional coding is performed on individual genes; Step S4-2: an initial population is randomly generated; Step S4-3: the target value of the population is calculated according to the objective function; Step S4-4: the river-sea intermodal transportation path and the cargo flow distribution result of each path are obtained; Step S4-5: line capacity and port traffic capacity check is performed; Step S4-6: capacity overrun penalty term calculation is performed; Step S4-7: the target value is returned; Step S4-8: whether the maximum iteration number is reached is judged, if yes, the process is ended, and if not, the process goes to step S4-9; Step S4-9: the target values are sorted according to their sizes, and individuals are selected and copied, PMX crossover is performed, random mutation is performed, and the processing result after the random mutation is returned to step S4-4, and the process is repeated to realize the solving of the double-layer coding genetic algorithm. 9.A device for optimizing river-sea intermodal transportation of large cargos, comprising a memory, a processor, and a program stored in the memory, wherein, The program is executed by the processor to realize the method of any one of claims 1-8.

10. A storage medium having stored thereon a program, characterized by The program is executed to realize the method of any one of claims 1-8.

Images