Container liner joint optimization method and system, and readable storage medium

Abstract

The invention provides a container liner joint optimization method and system and a readable storage medium, and the method comprises the steps: building a joint optimization model based on heterogeneous fleet deployment number, speed data and fuel filling data, wherein the joint optimization model comprises a dual-target mixed integer nonlinear programming model; solving the joint optimization model according to an epsilon-constraint method to obtain a first solving result; solving the joint optimization model according to a customized multi-objective genetic algorithm to obtain a second solving result; the membership degree of the target function is calculated based on the second solving result and a fuzzy decision method, a solution set of the membership degree of the target function is ranked according to preference of a decision maker, an optimization decision is obtained, and the optimization decision comprises a fleet deployment decision, a navigational speed determination decision and a fuel filling decision. Through the technical scheme of the invention, a container liner fleet deployment decision, a navigational speed determination decision and a fuel filling decision can be optimized, and low-carbon operation is realized.

Description

Technical Field

[0001] The present application relates to the technical field of navigation decision-making, and in particular to a container liner joint optimization method and system, and a readable storage medium. Background Art

[0002] In the global shipping market, liner shipping accounts for 15% of the total global seaborne trade and more than 70% of the total trade value. Container ships account for about 18% of the total deadweight tonnage of the world's fleet, and carbon emissions account for 31% of the entire shipping industry. Therefore, the task of carbon emission reduction in liner shipping is crucial. International shipping emission reduction work is mainly promoted from two aspects at the shipping level: technology and operation. The technical aspect focuses on the installation of power units and the application of alternative fuels, while the operational aspect takes measures such as reduced speed sailing. Previous studies have mainly focused on the heterogeneity of physical characteristics such as ship capacity and age, but lack in-depth discussion on the heterogeneity of carbon emission reduction technologies (such as power units and alternative fuels) of different ships, ignoring the heterogeneity of carbon emission reduction technologies and failing to fully consider the impact of various carbon emission reduction technologies on shipping efficiency and environmental benefits. Most studies focus on the refueling and consumption management of a single fuel, while there is little research on heterogeneous fleets equipped with multi-fuel systems such as dual-fuel engines or fuel cells, failing to fully consider the management complexity of different fuel combinations and their impact on operational efficiency. With the introduction of carbon emission reduction targets by the International Maritime Organization (IMO), the operational goal is no longer just to reduce operating costs, but also to reduce carbon emissions. However, most previous studies have focused on optimizing a single goal, lacking optimization analysis of the trade-offs and coordination between multiple operational goals. Summary of the invention

[0003] This application aims to solve or improve the above technical problems.

[0004] To this end, the first objective of the present application is to provide a container liner joint optimization method.

[0005] The second objective of the present application is to provide a container liner joint optimization system.

[0006] The third objective of the present application is to provide a container liner joint optimization system.

[0007] The fourth object of the present application is to provide a readable storage medium.

[0008] To achieve the first purpose of the present application, the technical solution of the first aspect of the present application provides a container liner joint optimization method, which is applied to a container liner joint optimization system. The container liner joint optimization method is executed by a processor. The container liner joint optimization method includes: based on the number of heterogeneous fleet deployments, speed data, and fuel refueling data, a joint optimization model is constructed through the container liner joint optimization system. The joint optimization model includes a dual-objective mixed integer nonlinear programming model. The dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions; through the container liner joint optimization system, a joint optimization model is constructed. The system solves the joint optimization model according to the ε-constraint method to obtain a first solution result; the container liner joint optimization system solves the joint optimization model according to the customized multi-objective genetic algorithm to obtain a second solution result; the first solution result is compared with the second solution result; the container liner joint optimization system calculates the objective function membership based on the second solution result and the fuzzy decision method, and the solution set of the objective function membership is sorted according to the decision maker's preference to obtain the optimization decision, which includes fleet deployment decision, speed determination decision and fuel refueling decision, and the objective function membership includes the first objective function membership and the second objective function membership.

[0009] According to the container liner joint optimization method provided in this application, firstly, a dual-objective mixed integer nonlinear programming model of container liner fleet deployment, speed optimization and fuel refueling joint optimization is constructed for the green liner transportation problem. Then the ε-constraint method is used to solve the Pareto solution set of the joint optimization model to obtain the first solution result. A customized multi-objective genetic algorithm (Multi-objective Optimization Genetic Algorithm, MOGA) is used to solve the Pareto solution set of the joint optimization model to obtain the second solution result. By comparing the first solution result and the second solution result, it is verified that the MOGA algorithm can solve large-scale problems. Finally, based on the results of MOGA, the membership of the two objective functions is calculated by the fuzzy decision method, and the optimal degree of the selected solution set generated based on the decision maker's preference is ranked to obtain the best solution, so as to achieve the optimal decision of integrated fleet deployment, speed determination and fuel refueling under the background of low-carbon operation. The container liner joint optimization method provided in this application can cope with the complex deployment requirements brought about by the diversity of ship types, improve the operational efficiency and flexibility of liner companies, and integrate the optimization of fuel filling and speed determination. It can not only effectively control operating costs while reducing carbon emissions, but also provide a new decision-making dimension for green liner transportation. By constructing a dual-objective mixed integer nonlinear programming model, it balances fleet deployment, speed determination and fuel filling, achieves low-carbon operation and simultaneous improvement of commercial value, and solves the problem that traditional optimization methods cannot take into account both environmental protection and economic benefits at the same time. By comparing ε-constraints with MOGA, the superiority of MOGA in solving large-scale problems is verified.

[0010] In addition, the technical solution provided by this application may also have the following additional technical features:

[0011] In some technical schemes, optionally, the joint optimization model is solved according to the ε-constraint method to obtain a first solution result, including: combining the cargo hold container capacity type of the ship with the available quantity of the ship, and discretizing it according to the V-type container capacity; linearizing the dual-objective mixed integer nonlinear programming model through the Big-M method; solving the Pareto solution set of the dual-objective mixed integer nonlinear programming model through the two-stage ε-constraint method to obtain the first solution result.

[0012] In this technical solution, the joint optimization model is solved according to the ε-constraint method to obtain the first solution result. Specifically, the dual-objective problem is first linearized, the type of tank capacity of the ship's cargo hold is combined with the available amount of such ships, and discretized according to the type of tank capacity, so that it is transformed into a 0-1 variable, that is, 1 means that the ship with the type of tank capacity equipped with alternative fuel for power plant is deployed on the route, and 0 means that it is not deployed. Then, the Big-M method is used to linearize the model. Finally, the Pareto solution set of the dual-objective mixed integer nonlinear programming model is solved by the two-stage ε-constraint method to obtain the first solution result.

[0013] In some technical schemes, optionally, a two-stage ε-constraint method is used to solve the Pareto solution set of the dual-objective mixed integer nonlinear programming model to obtain a first solution result, including: decomposing the dual-objective mixed integer nonlinear programming model into a first sub-problem and a second sub-problem, the first sub-problem has the minimization of total operating costs as the goal and the total carbon emissions as the constraint, and the second sub-problem has the minimization of total carbon emissions as the goal and the total operating costs as the constraint; by setting an incremental step size to obtain a specific value of ε, multiple single-objective problems are formed, and the Pareto solution set is obtained by solving the single-objective problems.

[0014] In this technical solution, the Pareto solution set of the dual-objective mixed integer nonlinear programming model is solved by a two-stage ε-constraint method to obtain the first solution result, specifically, the dual-objective mixed integer nonlinear programming model is decomposed into a first sub-problem and a second sub-problem, the first sub-problem takes the minimization of total operating costs as the goal and the total carbon emissions as the constraint, and the second sub-problem takes the minimization of total carbon emissions as the goal and the total operating costs as the constraint. Then, the specific value of ε is obtained by setting the incremental step size to form multiple single-objective problems, and the Pareto solution set is obtained by solving the single-objective problems.

[0015] In some technical solutions, optionally, a joint optimization model is solved according to a customized multi-objective genetic algorithm to obtain a second solution result, including: setting ship parameters, fuel and route parameters in the joint optimization model, and setting basic parameters of the inner genetic algorithm and the outer genetic algorithm; initializing individuals of the outer genetic algorithm, optimizing fuel filling schemes and fleet deployment schemes for various types of ships, and calculating the objective function value, the individuals include the speed of the ship in each section; sorting the initialized individuals by a fast non-dominated sorting method; performing crossover mutation on the sorted individuals to generate new individuals, calculating the objective function value, comparing the individuals and updating the species The updated individuals of the population are re-sorted by fast non-dominated sorting. When the outer algorithm reaches the maximum number of iterations, the iteration is stopped and the Pareto solution set is output. On the basis of determining the speed of each segment, the inner genetic algorithm is called. Under the premise of satisfying the constraints of ship fuel combustion energy and inventory level, the fuel filling variables are initialized as individuals to decide the fuel filling amount and fuel filling port. On the basis of determining the speed and fuel filling, the inner genetic algorithm is called to decide the fleet deployment on the basis of satisfying the unique constraint of speed selection for each segment and the constraint that the transportation volume of freight demand does not exceed the container capacity of the cargo hold of the deployed fleet.

[0016] In this technical solution, the joint optimization model is solved according to a customized multi-objective genetic algorithm to obtain the second solution result. Specifically, the parameters such as ships, fuels and routes in the joint optimization model are first set, and the basic parameters of the inner and outer genetic algorithms are set. Then, the individuals are initialized, each of which represents the speed of the ship in each section. For each individual, all routes and all ship types are traversed. Based on the fuel that can be refueled by the ship at the starting port, it is determined whether the ship type is suitable for the route. Based on the speed decision, the fuel refueling plan and fleet deployment plan of various types of ships are optimized, and two objective function values ​​are returned. The initialized individuals are sorted by the fast non-dominated sorting method of NSGA-Ⅱ (nondominated sorting genetic algorithm II, fast non-dominated sorting genetic algorithm with elite strategy). The sorted initialized individuals are cross-mutated to generate new individuals, the objective function value is calculated, the individuals are compared and the population is updated. The updated population individuals are sorted again by the fast non-dominated sorting method. When the outer algorithm reaches the maximum number of iterations, the iteration stops, and the Pareto solution set is output. The optimal solution under different preferences of the decision maker is obtained according to the obtained Pareto solution set and the fuzzy decision method. Specifically, to make decisions on fuel filling, based on the speed of each section, call the genetic algorithm, and initialize the fuel filling variable as an individual under the premise of satisfying the constraints. Calculate the first and second objective functions, and use fast non-dominated sorting to sort the individuals in the population. During the iteration process, perform crossover mutation to determine whether the solution meets the constraints, calculate the first and second objective functions, compare the individuals and update the population, and then use fast non-dominated sorting to sort the individuals in the population, and repeat this process until the maximum number of cycles is reached. To make decisions on fleet deployment, call the genetic algorithm based on the speed and fuel filling, and make decisions on fleet deployment based on satisfying the constraints.

[0017] In some technical solutions, optionally, individuals of the outer genetic algorithm are initialized to optimize the fuel filling schemes and fleet deployment schemes for various types of ships, and calculate the objective function value, including: for each individual, traverse all routes and all ship types, judge whether the ship type is suitable for the route based on the fuel that can be refueled by the ship at the starting port, optimize the fuel filling schemes and fleet deployment schemes for various types of ships based on the speed decision, and return two objective function values, where the individual is the speed of the ship in each section.

[0018] In this technical solution, individuals of the outer genetic algorithm are initialized to optimize the fuel filling schemes and fleet deployment schemes for various types of ships, and the objective function value is calculated. Specifically, each individual represents the speed of the ship in each section of the voyage. For each individual, all routes and all ship types are traversed, and whether the ship type is suitable for the route is determined based on the fuel that can be refueled at the starting port. Based on the speed decision, the fuel filling schemes and fleet deployment schemes for various types of ships are optimized, and two objective function values ​​are returned.

[0019] In some technical solutions, optionally, the joint optimization model includes:

[0020] The first objective function:

[0021]

[0022] in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuel, C represents the total operating cost of deploying a heterogeneous fleet to complete a voyage on a given route network, and C rvkf represents the alternative fuel f∈F used by power unit k on route r A The fixed cost of the V-type tank capacity ship, C if represents the price of fuel f at port i, C if′ represents the price of fuel f′ at port i, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r, represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's alternative fuel f∈F A The filling amount, represents the alternative fuel f∈F used by power plant k at port i on route r A The V-shaped tank capacity of the ship's fossil fuel f′∈F F The filling amount;

[0023] The second objective function:

[0024]

[0025] in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuels, E represents total carbon emissions, and E f represents the carbon emission factor of fuel f, E f′ represents the carbon emission factor of fuel f′, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A V-shaped tank capacity of the ship's alternative fuel f∈F A The consumption, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A The V-shaped tank capacity of the ship's fossil fuel f′∈F F The consumption, F vkf Denotes the use of alternative fuel f∈F for power unit k A V-type tank capacity Fuel consumption rate of the ship in port, T ri represents the stopover time of the ship on route r at port i.

[0026] In this technical solution, the joint optimization model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions.

[0027] In some technical solutions, optionally, the constraints of the joint optimization model are:

[0028] The only constraint on speed selection for each segment is:

[0029]

[0030] Among them, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0. r represents the set of ports of call on route r, indexed by i, and R represents the set of routes, indexed by r;

[0031] The transport volume of freight demand does not exceed the container capacity of the deployed fleet cargo hold:

[0032]

[0033] Among them, I r represents the set of ports of call on route r, with i as the index, R represents the set of routes, with r as the index, q ri represents the freight demand of port i on route r, V is the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuels, Denotes the use of alternative fuel f∈F for power unit k A V-type tank capacity The cargo hold capacity of the ship, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r;

[0034] The constraints to ensure that ships maintain a certain service frequency on each route are:

[0035]

[0036] Among them, I r represents the set of ports of call on route r, indexed by i, R represents the set of routes, indexed by r, V represents the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuel, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank-carrying capacity ships deployed on route r, S represents the speed set, with s as the index, φ r represents the port service frequency of route r, T ris T represents the sailing time of route r at speed s in segment i, ris =D ri / s,T ri represents the stopover time of the ship on route r at port i, D ri represents the sailing distance of segment i on route r, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0.

[0037] In this technical solution, the constraints of the joint optimization model include the unique constraint on the speed selection for each segment, the constraint that the transportation volume of the freight demand does not exceed the container capacity of the cargo hold of the deployed fleet, and the constraint that ensures that the ship maintains a certain service frequency on each route.

[0038] To achieve the second purpose of the present application, the technical solution of the second aspect of the present application provides a container liner joint optimization system, including: a model building module, which is used to build a joint optimization model through the container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data and fuel filling data, the joint optimization model includes a dual-objective mixed integer nonlinear programming model, the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions; a first solution module, which is used to solve the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain the first solution results; a second solving module, used to solve the joint optimization model according to a customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result; a comparison module, used to compare the first solution result and the second solution result; a solution determination module, used to calculate the objective function membership based on the second solution result and the fuzzy decision method through the container liner joint optimization system, sort the solution set of the objective function membership according to the decision maker's preference, and obtain the optimization decision, the optimization decision includes the fleet deployment decision, the speed determination decision and the fuel filling decision, and the objective function membership includes the first objective function membership and the second objective function membership.

[0039] According to the container liner joint optimization system provided in the present application, it includes a model construction module, a first solution module, a second solution module, a comparison module and a solution determination module. Among them, the model construction module is used to construct a joint optimization model through the container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data and fuel filling data. The joint optimization model includes a dual-objective mixed integer nonlinear programming model, and the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions. The first solution module is used to solve the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result. The second solution module is used to solve the joint optimization model according to a customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result. The comparison module is used to compare the first solution result and the second solution result. The scheme determination module is used to calculate the objective function membership based on the second solution result and the fuzzy decision method through the container liner joint optimization system, sort the solution set of the objective function membership according to the decision maker's preference, and obtain the optimization decision. The optimization decision includes fleet deployment decision, speed determination decision and fuel filling decision, and the objective function membership includes the first objective function membership and the second objective function membership. The container liner joint optimization system provided by the present application can cope with the complex deployment requirements brought about by the diversity of ship types, improve the operating efficiency and flexibility of liner companies, integrate and optimize fuel filling and speed determination, and can effectively control operating costs while reducing carbon emissions. It also provides a new decision-making dimension for green liner transportation, and balances fleet deployment, speed determination and fuel filling by constructing a dual-objective mixed integer nonlinear programming model, achieving low-carbon operation and simultaneous improvement of commercial value, and solving the problem that traditional optimization methods cannot take into account both environmental protection and economic benefits at the same time. By comparing ε-constraints with MOGA, the superiority of MOGA in large-scale problem solving is verified.

[0040] To achieve the third purpose of the present application, the technical solution of the third aspect of the present application provides a container liner joint optimization system, including: a memory and a processor, wherein the memory stores programs or instructions that can be run on the processor, and when the processor executes the program or instruction, it implements the container liner joint optimization method of any one of the technical solutions of the first aspect, and thus has the technical effect of any one of the technical solutions of the first aspect above, which will not be repeated here.

[0041] In order to achieve the fourth purpose of the present application, the technical solution of the fourth aspect of the present application provides a readable storage medium on which a program or instruction is stored. When the program or instruction is executed by the processor, the steps of the container liner joint optimization method of any one of the technical solutions of the first aspect are implemented, and thus the technical effect of any one of the technical solutions of the first aspect is obtained, which will not be repeated here.

[0042] Additional aspects and advantages of the present application will become apparent in the following description or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0044] Figure 1 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0045] Figure 2 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0046] Figure 3 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0047] Figure 4 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0048] Figure 5 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0049] Figure 6 This is a schematic structural block diagram of a container liner joint optimization system according to an embodiment of the present application;

[0050] Figure 7 This is a schematic structural block diagram of a container liner joint optimization system according to an embodiment of the present application;

[0051] Figure 8 This is a schematic diagram of the steps of a container liner joint optimization method according to an embodiment of the present application;

[0052] Fig. 9 This is a schematic diagram of the MOGA algorithm flow of a container liner joint optimization method according to an embodiment of the present application;

[0053] Fig.10 A schematic diagram of a Pareto solution set, a Pareto frontier, and a corresponding fleet deployment decision obtained by the first MOGA run of a container liner joint optimization method according to an embodiment of the present application;

[0054] Fig.11 A schematic diagram of a Pareto solution set, a Pareto frontier, and their corresponding fleet deployment decision obtained by the second MOGA run of the container liner joint optimization method according to an embodiment of the present application;

[0055] Fig.12A schematic diagram of a Pareto solution set, a Pareto frontier, and a corresponding fleet deployment decision obtained by the third MOGA run of the container liner joint optimization method according to an embodiment of the present application;

[0056] Fig.13 A schematic diagram of a Pareto solution set, a Pareto frontier, and their corresponding fleet deployment decision obtained by the fourth MOGA run of a container liner joint optimization method according to an embodiment of the present application;

[0057] Fig.14 A schematic diagram of a Pareto solution set, a Pareto frontier, and a corresponding fleet deployment decision obtained by the fifth MOGA run of a container liner joint optimization method according to an embodiment of the present application;

[0058] Fig.15 A schematic diagram of the relationship between the fixed cost of a ship, the fuel filling cost and the total carbon emissions for five operations of a container liner joint optimization method according to an embodiment of the present application;

[0059] Fig.16 A schematic diagram of the relationship between the fuel filling amount, fuel consumption and total carbon emissions of five operations of the container liner joint optimization method according to an embodiment of the present application;

[0060] Fig.17 This is a schematic diagram of the relationship between the average speed of all segments and the total carbon emissions of the container liner joint optimization method according to an embodiment of the present application.

[0061] in, Figure 6 and Figure 7 The corresponding relationship between the reference numerals and component names in the figure is:

[0062] 10: container liner joint optimization system; 110: model building module; 120: first solution module; 130: second solution module; 140: comparison module; 150: solution determination module; 20: container liner joint optimization system; 300: memory; 400: processor. DETAILED DESCRIPTION

[0063] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0064] In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present application is not limited to the specific embodiments disclosed below.

[0065] Refer to the following Figures 1 to 17 Describe the container liner joint optimization method and system and readable storage medium of some embodiments of the present application.

[0066] like Figure 1 As shown, the embodiment of the first aspect of the present application provides a container liner joint optimization method, which is applied to a container liner joint optimization system. The container liner joint optimization method is executed by a processor, and the container liner joint optimization method includes the following steps:

[0067] Step S102: constructing a joint optimization model through a container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data, and fuel refueling data, wherein the joint optimization model includes a dual-objective mixed integer nonlinear programming model, and the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions;

[0068] Step S104: solving the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result;

[0069] Step S106: solving the joint optimization model according to the customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result;

[0070] Step S108: comparing the first solution result and the second solution result;

[0071] Step S110: The objective function membership is calculated by the container liner joint optimization system based on the second solution result and the fuzzy decision method, and the solution set of the objective function membership is sorted according to the decision maker's preference to obtain an optimized decision. The optimized decision includes a fleet deployment decision, a speed determination decision and a fuel filling decision. The objective function membership includes a first objective function membership and a second objective function membership.

[0072] According to the container liner joint optimization method provided in this embodiment, firstly, a dual-objective mixed integer nonlinear programming model for the joint optimization of container liner fleet deployment, speed optimization and fuel refueling is constructed for the green liner transportation problem. Then, the ε-constraint method is used to solve the Pareto solution set of the joint optimization model to obtain the first solution result. A customized multi-objective genetic algorithm MOGA is used to solve the Pareto solution set of the joint optimization model to obtain the second solution result. By comparing the first solution result with the second solution result, it is verified that the MOGA algorithm can solve large-scale problems. Finally, based on the results of MOGA, the membership of the two objective functions is calculated by the fuzzy decision method, and the optimal degree of the selected solution set generated based on the decision maker's preference is ranked to obtain the best solution, so as to achieve the optimal decision of integrated fleet deployment, speed determination and fuel refueling under the background of low-carbon operation. The container liner joint optimization method provided in this application can cope with the complex deployment requirements brought about by the diversity of ship types, improve the operational efficiency and flexibility of liner companies, and integrate the optimization of fuel filling and speed determination. It can not only effectively control operating costs while reducing carbon emissions, but also provide a new decision-making dimension for green liner transportation. By constructing a dual-objective mixed integer nonlinear programming model, it balances fleet deployment, speed determination and fuel filling, achieves low-carbon operation and simultaneous improvement of commercial value, and solves the problem that traditional optimization methods cannot take into account both environmental protection and economic benefits at the same time. By comparing ε-constraints with MOGA, the superiority of MOGA in solving large-scale problems is verified.

[0073] like Figure 2 As shown, according to a container liner joint optimization method according to an embodiment of the present application, the joint optimization model is solved according to the ε-constraint method to obtain a first solution result, including the following steps:

[0074] Step S202: combining the cargo hold tank capacity type of the ship with the available quantity of the ship, and discretizing according to the V-type tank capacity;

[0075] Step S204: linearizing the dual-objective mixed integer nonlinear programming model by Big-M method;

[0076] Step S206: solving the Pareto solution set of the dual-objective mixed integer nonlinear programming model by a two-stage ε-constraint method to obtain a first solution result.

[0077] In this embodiment, the joint optimization model is solved according to the ε-constraint method to obtain the first solution result. Specifically, the dual-objective problem is first linearized, and the type of tank capacity of the ship cargo hold is combined with the available amount of such ships, and discretized according to the v-type tank capacity, so as to transform it into a 0-1 variable, that is, 1 indicates that the ship with a type of tank capacity equipped with a power plant using alternative fuels is deployed on the route, and 0 indicates that it is not deployed. Then, the Big-M method is used to linearize the model. Finally, the Pareto solution set of the dual-objective mixed integer nonlinear programming model is solved by the two-stage ε-constraint method to obtain the first solution result.

[0078] like Figure 3 As shown, according to a container liner joint optimization method according to an embodiment of the present application, a Pareto solution set of a dual-objective mixed integer nonlinear programming model is solved by a two-stage ε-constraint method to obtain a first solution result, including the following steps:

[0079] Step S302: decomposing the dual-objective mixed integer nonlinear programming model into a first sub-problem and a second sub-problem, wherein the first sub-problem takes minimizing the total operating cost as the objective and the total carbon emissions as the constraint, and the second sub-problem takes minimizing the total carbon emissions as the objective and the total operating cost as the constraint;

[0080] Step S304: obtaining a specific value of ε by setting an incremental step size, forming multiple single-objective problems, and obtaining a Pareto solution set by solving the single-objective problems.

[0081] In this embodiment, the Pareto solution set of the dual-objective mixed integer nonlinear programming model is solved by a two-stage ε-constraint method to obtain a first solution result, specifically, the dual-objective mixed integer nonlinear programming model is decomposed into a first sub-problem and a second sub-problem, the first sub-problem takes the minimization of total operating costs as the goal and the total carbon emissions as the constraint, and the second sub-problem takes the minimization of total carbon emissions as the goal and the total operating costs as the constraint. Then, the specific value of ε is obtained by setting the incremental step size to form multiple single-objective problems, and the Pareto solution set is obtained by solving the single-objective problems.

[0082] like Figure 4 As shown, according to a container liner joint optimization method according to an embodiment of the present application, a joint optimization model is solved according to a customized multi-objective genetic algorithm to obtain a second solution result, including the following steps:

[0083] Step S402: setting the ship parameters, fuel and route parameters in the joint optimization model, and setting the basic parameters of the inner genetic algorithm and the outer genetic algorithm;

[0084] Step S404: Initialize individuals of the outer genetic algorithm, optimize the fuel filling scheme and fleet deployment scheme of various types of ships, and calculate the value of the objective function. The individuals include the speed of the ship in each section;

[0085] Step S406: sorting the initialized individuals by a fast non-dominated sorting method;

[0086] Step S408: performing crossover mutation on the sorted individuals to generate new individuals, calculating the objective function value, comparing the individuals and updating the population, and re-sorting the updated population individuals using a fast non-dominated sorting method;

[0087] Step S410: When the outer algorithm reaches the maximum number of iterations, the iteration is stopped and the Pareto solution set is output;

[0088] Step S412: Based on the speed of each section, the inner genetic algorithm is called to initialize the fuel filling variable as an individual, and decide the fuel filling amount and fuel filling port under the premise of satisfying the constraints of ship fuel combustion energy and inventory level.

[0089] Step S414: Based on the speed and fueling, the inner genetic algorithm is called to decide the fleet deployment on the basis of satisfying the unique constraint of speed selection for each segment and the constraint that the transportation volume of the freight demand does not exceed the container capacity of the cargo hold of the deployed fleet.

[0090] In this embodiment, the joint optimization model is solved according to a customized multi-objective genetic algorithm to obtain the second solution result, specifically, the parameters such as ships, fuels and routes in the joint optimization model are first set, and the basic parameters of the inner and outer genetic algorithms are set. Then the individuals are initialized, each individual represents the speed of the ship in each section, and for each individual, all routes and all ship types are traversed, and the fuel that can be refueled by the ship at the starting port is used to determine whether the ship type is suitable for the route. Based on the speed decision, the fuel refueling plan and fleet deployment plan of various types of ships are optimized, and two objective function values ​​are returned. The initialized individuals are sorted using the fast non-dominated sorting method of NSGA-Ⅱ. The sorted initialized individuals are cross-mutated to generate new individuals, the objective function values ​​are calculated, the individuals are compared and the population is updated. The updated population individuals are sorted again using the fast non-dominated sorting method. When the outer algorithm reaches the maximum number of iterations, the iteration stops, and the Pareto solution set is output. The optimal solution under different preferences of the decision maker is obtained according to the obtained Pareto solution set and the fuzzy decision method mentioned in step S2. Specifically, to make decisions on fuel filling, based on the speed of each section, call the genetic algorithm, and initialize the fuel filling variable as an individual under the premise of satisfying the constraints. Calculate the first and second objective functions, and use fast non-dominated sorting to sort the individuals in the population. During the iteration process, perform crossover mutation to determine whether the solution meets the constraints, calculate the first and second objective functions, compare the individuals and update the population, and then use fast non-dominated sorting to sort the individuals in the population, and repeat this process until the maximum number of cycles is reached. To make decisions on fleet deployment, call the genetic algorithm based on the speed and fuel filling, and make decisions on fleet deployment based on satisfying the constraints.

[0091] like Figure 5 As shown, according to a container liner joint optimization method of an embodiment proposed in the present application, individuals of the outer genetic algorithm are initialized, fuel filling schemes and fleet deployment schemes of various types of ships are optimized, and the objective function value is calculated, including the following steps:

[0092] Step S502: For each individual, traverse all routes and all ship types, determine whether the ship type is suitable for the route based on the fuel that can be refueled at the starting port, and optimize the fuel refueling plan and fleet deployment plan for various types of ships based on the speed decision, and return two objective function values. The individual is the speed of the ship in each section.

[0093] In this embodiment, individuals of the outer genetic algorithm are initialized, fuel filling schemes and fleet deployment schemes for various types of ships are optimized, and the objective function value is calculated. Specifically, each individual represents the speed of the ship in each section of the route. For each individual, all routes and all ship types are traversed, and whether the ship type is suitable for the route is determined based on the fuel that can be refueled at the starting port. The fuel filling schemes and fleet deployment schemes for various types of ships are optimized based on the speed decision, and two objective function values ​​are returned.

[0094] In some embodiments, optionally, the joint optimization model includes:

[0095] The first objective function:

[0096]

[0097] in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuel, C represents the total operating cost of deploying a heterogeneous fleet to complete a voyage on a given route network, and C rvkf represents the alternative fuel f∈F used by power unit k on route r A The fixed cost of the V-type tank capacity ship, C if represents the price of fuel f at port i, C if′ represents the price of fuel f′ at port i, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r, represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's alternative fuel f∈F A The filling amount, represents the alternative fuel f∈F used by power plant k at port i on route r A The V-shaped tank capacity of the ship's fossil fuel f′∈F F The filling amount;

[0098] The second objective function:

[0099]

[0100] in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuels, E represents total carbon emissions, and E f represents the carbon emission factor of fuel f, E f′ represents the carbon emission factor of fuel f′, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A V-shaped tank capacity of the ship's alternative fuel f∈F A The consumption, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A The V-shaped tank capacity of the ship's fossil fuel f′∈F F The consumption, F vkf Denotes the use of alternative fuel f∈F for power unit k A V-type tank capacity Fuel consumption rate of the ship in port, T ri represents the stopover time of the ship on route r at port i.

[0101] In some embodiments, optionally, the constraints of the joint optimization model are:

[0102] The only constraint on speed selection for each segment is:

[0103]

[0104] Among them, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0. r represents the set of ports of call on route r, indexed by i, and R represents the set of routes, indexed by r;

[0105] The transport volume of freight demand does not exceed the container capacity of the deployed fleet cargo hold:

[0106]

[0107] Among them, I r represents the set of ports of call on route r, with i as the index, R represents the set of routes, with r as the index, q ri represents the freight demand of port i on route r, V is the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuels, Denotes the use of alternative fuel f∈F for power unit k A V-type tank capacity The cargo hold capacity of the ship, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r;

[0108] The constraints to ensure that ships maintain a certain service frequency on each route are:

[0109]

[0110] Among them, I r represents the set of ports of call on route r, indexed by i, R represents the set of routes, indexed by r, V represents the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuel, x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank-carrying capacity ships deployed on route r, S represents the speed set, with s as the index, φ r represents the port service frequency of route r, T ris T represents the sailing time of route r at speed s in segment i, ris =D ri / s,T ri represents the stopover time of the ship on route r at port i, D ri represents the sailing distance of segment i on route r, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0.

[0111] like Figure 6As shown, an embodiment of the second aspect of the present application provides a container liner joint optimization system 10, including: a model building module 110, which is used to build a joint optimization model through the container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data and fuel filling data, the joint optimization model includes a dual-objective mixed integer nonlinear programming model, the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions; a first solving module 120, which is used to solve the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result; a second solving module 120, which is used to solve the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result; The solution module 130 is used to solve the joint optimization model according to the customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result; the comparison module 140 is used to compare the first solution result and the second solution result; the solution determination module 150 is used to calculate the objective function membership based on the second solution result and the fuzzy decision method through the container liner joint optimization system, sort the solution set of the objective function membership according to the decision maker's preference, and obtain the optimization decision, the optimization decision includes the fleet deployment decision, the speed determination decision and the fuel filling decision, and the objective function membership includes the first objective function membership and the second objective function membership.

[0112] According to the container liner joint optimization system 10 provided in this embodiment, it includes a model construction module 110, a first solution module 120, a second solution module 130, a comparison module 140 and a solution determination module 150. Among them, the model construction module 110 is used to construct a joint optimization model through the container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data and fuel filling data. The joint optimization model includes a dual-objective mixed integer nonlinear programming model, and the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions. The first solution module 120 is used to solve the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result. The second solution module 130 is used to solve the joint optimization model according to a customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result. The comparison module 140 is used to compare the first solution result with the second solution result. The scheme determination module 150 is used to calculate the objective function membership based on the second solution result and the fuzzy decision method through the container liner joint optimization system, sort the solution set of the objective function membership according to the decision maker's preference, and obtain the optimization decision. The optimization decision includes fleet deployment decision, speed determination decision and fuel filling decision, and the objective function membership includes the first objective function membership and the second objective function membership. The container liner joint optimization system provided by the present application can cope with the complex deployment requirements brought about by the diversity of ship types, improve the operating efficiency and flexibility of liner companies, integrate and optimize fuel filling and speed determination, and can effectively control operating costs while reducing carbon emissions. It also provides a new decision dimension for green liner transportation, and balances fleet deployment, speed determination and fuel filling by constructing a dual-objective mixed integer nonlinear programming model, achieving low-carbon operation and simultaneous improvement of commercial value, solving the problem that traditional optimization methods cannot take into account both environmental protection and economic benefits at the same time. By comparing ε-constraints with MOGA, the superiority of MOGA in large-scale problem solving is verified.

[0113] like Figure 7 As shown, an embodiment of the third aspect of the present application provides a container liner joint optimization system 20, comprising: a memory 300 and a processor 400, wherein the memory 300 stores a program or instruction that can be run on the processor 400, and when the processor 400 executes the program or instruction, the steps of the container liner joint optimization method of any one of the embodiments of the first aspect are implemented, so it has the technical effect of any one of the embodiments of the first aspect mentioned above, which will not be repeated here.

[0114] An embodiment of the fourth aspect of the present application provides a readable storage medium on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the container liner joint optimization method of any one of the embodiments of the first aspect are implemented, and thus the technical effects of any one of the embodiments of the first aspect mentioned above are not repeated here.

[0115] like Figure 8 , Fig. 9 , Fig.10 , Fig.11 , Fig.12 , Fig.13 , Fig.14 , Fig.15 , Fig.16 , Fig.17 As shown, according to a specific embodiment of the container liner joint optimization method provided by the present application, the method includes:

[0116] S1: A bi-objective mixed-integer nonlinear programming model is proposed for GLSP. That is, for the green liner shipping problem (GLSP), a bi-objective mixed-integer nonlinear programming model (BO-MINLP, BO-Mixed-Integer Nonlinear Programming) for container liner fleet deployment, speed optimization and fuel refueling joint optimization is constructed;

[0117] S2: Linearize the model through linearization technology and transform the dual-objective problem into a single-objective problem through the epsilon constraint method. That is, the Pareto solution set of the joint optimization model is solved by the ε-constraint method;

[0118] S3: Based on the decision logic of fleet deployment, speed determination and fuel refueling, a multi-objective genetic algorithm (MOGA) was customized to solve BO-MINLP. That is, a customized multi-objective genetic algorithm MOGA was used to solve the Pareto solution set of the joint optimization model;

[0119] S4: The effectiveness of the two methods was verified by comparing their performance. That is, the results of the ε-constraint method and the MOGA algorithm were compared to verify that the MOGA algorithm can solve large-scale problems.

[0120] S5: In order to reflect the practical value of the GLSP model algorithm, it is applied to near-sea routes, revealing the balance between economic efficiency and carbon emission reduction that liner companies are more concerned about, and proposing some targeted decision-making references at the operational level. That is, based on the results of MOGA, the membership of the two objective functions is calculated by fuzzy decision-making method, and the optimal degree of the selected solution set generated based on the decision maker's preference is ranked to obtain the best solution, so as to achieve the optimal decision of integrated fleet deployment, speed determination and fuel refueling in the context of low-carbon operation.

[0121] In step S1, for the GLSP based on heterogeneous fleet deployment, speed determination, and fuel refueling, a container liner fleet deployment, speed optimization, and fuel refueling joint optimization BO-MINLP is constructed;

[0122] Furthermore, the dual objective function of the model is:

[0123]

[0124] in, I r represents the set of ports of call (legs) on route r, with i as the index, i also represents the leg from i to i+1, |I r | is the last port of call before returning to the starting port; R represents the route set, with r as the index; V represents the cargo hold's container capacity set, with v as the index; K represents the set of power devices carried, with k as the index; f∈F A , where F A represents alternative fuel; f′∈F F , where F F represents fossil fuels; C represents the total operating cost of a heterogeneous fleet deployed to complete a voyage on a given route network, C rvkf represents the alternative fuel f∈F used by power unit k on route r A The fixed cost of a V-shaped tank capacity ship (the fixed cost required to complete a voyage), C if represents the price of fuel f at port i; x rvkf Denotes the use of alternative fuel f∈F for power unit k A The number of V-shaped tank capacity ships deployed on route r; represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's alternative fuel f∈F A The filling amount, represents the alternative fuel f∈F used by power plant k at port i on route r A The V-shaped tank capacity of the ship's fossil fuel f′∈F F The filling amount;

[0125]

[0126] Where E represents the total carbon emissions, E f represents the carbon emission factor of fuel f; Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F AV-shaped tank capacity of the ship's alternative fuel f∈F A The consumption of Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A The V-shaped tank capacity of the ship's fossil fuel f′∈F F Consumption of vkf Denotes the use of alternative fuel f∈F for power unit k A V-shaped tank capacity Fuel consumption rate of the ship in port (T / H), where only alternative fuel is consumed; T ri represents the stopover time of the ship on route r at port i;

[0127] Furthermore, the constraints of the model are:

[0128] The only constraint on speed selection for each segment is:

[0129]

[0130] In the formula, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0;

[0131] The transport volume of freight demand does not exceed the container capacity of the deployed fleet cargo hold:

[0132]

[0133] In the formula, q ri represents the freight demand of port i on route r, V is the set of container capacities of cargo holds, indexed by v; Denotes the use of alternative fuel f∈F for power unit k A V-type tank capacity of the ship's cargo hold capacity;

[0134] The constraints to ensure that ships maintain a certain service frequency on each route are:

[0135]

[0136] Where S represents the speed set, with s as the index; φ r represents the port service frequency of route r; T ris represents the sailing time (H) of route r at speed s in segment i, T ris =D ri / s,D ri represents the sailing distance of segment i on route r;

[0137] The combustion energy of dual-fuel ship fuel satisfies the formula:

[0138]

[0139] In the formula, ξ vkf Indicates that the same container ship is equipped with a power unit k and uses alternative fuel f∈F A and the matching of cargo hold capacity v; L f represents the lower calorific value of fuel f; γ vkff′ Indicates that the power unit k is used and the alternative fuel f∈F is used A , the compatibility of container ships with cargo hold capacity v and fossil fuel f′; M vks represents the hourly power requirement of a V-type tank capacity ship equipped with power unit k at speed s; K DF represents a subset of dual-fuel powerplants;

[0140] The combustion energy of fossil fuel as pilot oil for dual-fuel ships meets the formula:

[0141]

[0142] The combustion energy of fuel cell ship fuel satisfies the formula:

[0143]

[0144] In the formula, K FC represents the fuel cell power plant subset, K FC =KK DF ;

[0145] The prerequisite constraints are:

[0146]

[0147] In the formula, z i f is a 0-1 auxiliary parameter, which determines whether port i is a fueling port for fuel f, if yes, it is 1, if not, it is 0; Denotes the use of alternative fuel f∈F for power unit k A V-shaped tank capacity of the ship's alternative fuel f∈F A Tank capacity, Denotes the use of alternative fuel f∈F for power unit k A The V-shaped tank capacity of the ship's fossil fuel f′∈F F Tank capacity; represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's alternative fuel f∈F A inventory levels, represents the alternative fuel f∈F used by power plant k at port i on route r AThe V-shaped tank capacity of the ship's fossil fuel f′∈F F inventory levels;

[0148] The formula for the initial inventory level of fuel that a ship is equipped with when it departs from the port of origin is:

[0149]

[0150] In addition to the departure port, the fuel inventory level constraints that ships should meet are:

[0151]

[0152]

[0153] The number of deployed ships is constrained as:

[0154]

[0155] In the formula, m vkf Denotes the use of alternative fuel f∈F for power unit k A The available amount of V-tank capacity of the vessel;

[0156] The domain of the decision variable is:

[0157]

[0158] x rvkf ∈Z + ∪{0},f∈F A ,k∈K,v∈V,r∈R(18);

[0159] w ris ∈{0,1},v∈V,s∈S,i∈I r :i<|I r |,r∈R(19);

[0160] The hourly power requirement M for a container ship with a V-type tank capacity and a power unit k is vks (MJ / T), that is, the fuel consumption demand, is calculated using the following formula:

[0161]

[0162] In the formula, p vk is the capacity of a V-shaped tank carrying a power unit k at a design speed s vk The reference power under the condition of speed; n is the index showing the relationship between speed and power; η vk is the efficiency of burning fuel in a ship with a V-shaped tank capacity carrying power unit k;

[0163] In step S2, to solve the BO-MINLP constructed for GLSP, the ε-constraint method is adopted. The specific steps are as follows:

[0164] S21: Linearize the dual-objective problem, combine the type of cargo tank capacity of the ship with the available quantity of such ships, discretize it according to the type of tank capacity, and transform it into a 0-1 variable, that is, 1 means that the ship with the type of tank capacity equipped with alternative fuels is deployed on the route, and 0 means that it is not deployed. Then, the Big-M method is used to linearize the model;

[0165] S22: A two-stage ε-constraint method is used to solve the problem. BO-MINLP is decomposed into two sub-problems. The specific value of ε is obtained by setting the Δ step size to form a series of single-objective problems. The Pareto solution set is obtained by solving these problems.

[0166] In step S21, the process of solving the bi-objective problem linearization and model linearization includes:

[0167] S211: Linearization of the dual-objective problem. Introducing the variable n rivkf , o rivkff′ 、p rivkf 、u rivkff′ , the type of cargo hold tank capacity of the ship is combined with the available capacity of this ship, and discretized according to the V-type tank capacity, so that x rvkf Transformed into a 0-1 variable, i.e. 1 means that the power unit k uses alternative fuel f∈F A Vessels with a V-shaped tank capacity of are deployed on route r, and 0 means they are not deployed;

[0168] S212: Use Big-M method to linearize the model. The model can set the M value according to the actual situation. The M value of this model does not need to be too large, it can be greater than the fuel tank capacity, set M1 = M2 = 1700;

[0169] In step S211, variable n is introduced rivkf ,

[0170]

[0171] n rivkf ≤M1xr vkf ,f∈F A ,k∈K,v∈V,i∈Ir,r∈R (22);

[0172]

[0173]

[0174] n rivkf ∈R +∪{0},f∈F A ,k∈K,v∈V,i∈Ir,r∈R (25);

[0175] Introduce variable o rivkff′ :

[0176]

[0177] o rivkff′ ≤M1x rvkf ,f′∈F F ,f∈F A ,k∈K DF ,v∈V,i∈Ir,r∈R (27);

[0178]

[0179]

[0180] o rivkff′ ∈R + ∪{0},f′∈F F ,f∈F A ,k∈K DF ,v∈V,i∈Ir,r∈R (30);

[0181] Introduce variable p rivkf :

[0182]

[0183] p rivkf ≤M2x rvkf ,f∈F A ,k∈K,v∈V,i∈Ir,r∈R (32);

[0184]

[0185]

[0186] p rivkf ∈R + ∪{0},f∈F A ,k∈K,v∈V,i∈Ir,r∈R (35);

[0187] Introduce variable u rivkff′ :

[0188]

[0189] u rivkff′ ≤M2x rvkf ,f′∈F F ,f∈F A,k∈K DF ,v∈V,i∈I r ,r∈R (37);

[0190]

[0191]

[0192] u rivkff′ ∈R + ∪{0},f′∈F F ,f∈F A ,k∈K DF ,v∈V,i∈Ir,r∈R (40);

[0193] In step S212, the model can set the M value according to the actual situation. Set M1 = M2 = 1700. Therefore, the original model can be converted into:

[0194]

[0195]

[0196] st(3)-(20),(21)-(40).;

[0197] In step S22, the two-stage ε-constraint method solution process includes:

[0198] S221: Decompose BO-MINLP into two sub-problems. Sub-problem 1 takes minimizing total operating cost as the objective and total carbon emissions as the constraint, and sub-problem 2 takes minimizing total carbon emissions as the objective and total operating cost as the constraint;

[0199] S222: By setting the step size Δ, a specific value of ε is obtained to form a series of single-objective problems, and the Pareto solution set is obtained by solving these problems.

[0200] In step S221, the formula is:

[0201] Definition of two sub-problems:

[0202] min{C(x)|E(x)≤ε,x∈X} (43);

[0203] min{E(x)|C(x)≤ε,x∈X} (44);

[0204] In the formula, x is the decision vector of all decision variables, and the value of ε is within a certain interval.

[0205] Find the ideal point (f1 I ,f2 I ) and the farthest point (f1N ,f2 N )The formula is:

[0206] f1 I = min{C(x)|x∈X} (45);

[0207]

[0208]

[0209] f2 N =min{E(x)|C(x)=f1 I ,x∈X} (48);

[0210] In step S222, (45)-(46) and (47)-(48) are solved to obtain the lower and upper bounds of the two objective functions, thereby generating approximate ideal points and farthest points, and determining the value range of ε. Furthermore, by setting the step size Δ to obtain the specific value of ε, a series of single-objective problems are formed, and the Pareto solution set is obtained by solving these problems.

[0211] Furthermore, the linear membership function of each Pareto objective function is:

[0212]

[0213] In the formula, represents the membership of the jth objective function of the sth solution, and are the lower and upper bounds of the j-th objective function, respectively. represents the s-th solution to the j-th objective function;

[0214] Total membership θ S The calculation formula is:

[0215]

[0216] In the formula, ω j is the weight of the jth objective function. The specific value is determined by the decision maker’s preference. The best solution is θ S The maximum corresponding solution;

[0217] In step S3, in order to solve the BO-MINLP constructed for GLSP, a customized multi-objective genetic algorithm MOGA method is used. The specific steps are as follows:

[0218] S31: Parameter setting. Set the parameters of the ship, fuel and route in the model; set the basic parameters of the inner and outer genetic algorithms;

[0219] S32: Initialization process. Initialize individuals, each individual (i.e., chromosome) represents the speed of the ship in each section; for each individual, traverse all routes and all ship types, determine whether the ship type is suitable for the route based on the fuel that can be refueled at the starting port, optimize the fuel refueling plan and fleet deployment plan for various types of ships based on the speed decision, and return two objective function values;

[0220] S33: Use the fast non-dominated sorting method of NSGA-Ⅱ to sort the initialized individuals;

[0221] S34: Iteration. Perform crossover mutation on the sorted initialized individuals to generate new individuals, calculate the objective function value, compare individuals and update the population. Use the fast non-dominated sorting method to sort the updated population individuals again;

[0222] S35: When the outer algorithm reaches the maximum number of iterations, the iteration is stopped, and the Pareto solution set is output. The optimal solution under different preferences of the decision maker is obtained according to the obtained Pareto solution set and the fuzzy decision method mentioned in step S2;

[0223] In step S32, the decision making step includes:

[0224] S321: Decision on fueling. After determining the speed of each segment, call the genetic algorithm, and initialize the fueling variable as an individual under the premise of satisfying constraints (6)-(8), constraints (13)-(14) and constraints (16)-(17). Calculate the objective function (1)-(2), and use fast non-dominated sorting to sort the individuals in the population; during the iteration process, perform crossover mutation, and determine whether the solution satisfies the constraints through constraints (6)-(8), constraints (13)-(14) and constraints (16)-(17), calculate the objective function (1)-(2), compare the individuals and update the population, and then use fast non-dominated sorting to sort the individuals in the population; repeat this process until the maximum number of cycles is reached;

[0225] S322: fleet deployment decision. Based on the speed and fuel filling, the genetic algorithm is called to make fleet deployment decisions based on constraints (4)-(5), constraint (15) and constraint (18). The specific cycle process is the same as S321.

[0226] Specifically, Fig. 9As shown in the figure, the MOGA algorithm process principle includes: S602: setting the basic parameters of the model and algorithm; S604: initializing the individuals of the outer genetic algorithm, namely the ship speed; S606: optimizing the fuel filling plan and the fleet deployment plan and calculating the objective function value; S608: sorting the individuals using the fast non-dominated sorting method; S610: judging whether the maximum number of iterations of the algorithm has been reached; S612: if so, outputting the Pareto solution set; S614: if not, generating new individuals through crossover and mutation; S616: optimizing the fuel filling plan, optimizing the fleet deployment plan, and calculating the total operating cost and total carbon emissions.

[0227] This example defines and explores the GLSP problem, which integrates heterogeneous fleet deployment, speed determination, and fuel refueling decisions. This problem is significantly different from previous studies on the deployment of homogeneous fleets using dual-fuel technology on a single route. This example proposes a BO-MINLP model that takes into account the characteristics of heterogeneous fleets (such as cargo capacity, power systems, and alternative fuels), as well as factors such as inherent service frequency in liner shipping, transported cargo volume, various types of refueling ports, and alternative fuel refueling prices.

[0228] The trade-off between low carbon and economic benefits:

[0229] like Fig.10 , Fig.11 , Fig.12 , Fig.13 , Fig.14 As shown in the figure, the distribution of the Pareto front shows that there is an obvious conflict between the two goals pursued by liner companies: minimizing total operating costs and minimizing total carbon emissions. In the process of pursuing carbon emission reduction, liner companies can make trade-offs through preferences and determine the best solution. If carbon emissions can be ignored, the solution with the lowest total operating cost appears in Fig.14 In the case of the 2017 IPO, only US$5.01 million was spent, and the total carbon emissions were 9,822.0 tons. Similarly, if cost control can be ignored, the lowest total carbon emissions are Fig.10 Of these, only 7,846.6 tons were transported, and the total operating cost at this time reached US$5.74 million.

[0230] Optimal decision making for heterogeneous fleet deployment:

[0231] The GLSP proposed in this embodiment comprehensively considers heterogeneous fleets with different capacities, power plants, alternative fuels, and other characteristics, the inherent service frequency of liner shipping, freight demand, and different fuel refueling ports, alternative fuel refueling prices, and other factors, breaking through the limitations of previous solutions to the problem of homogeneous fleets. This comprehensive design enables the fleet to adjust more flexibly and efficiently when facing different freight demands and port conditions, and optimize resource allocation.

[0232] This example emphasizes that dual-fuel LNG ships are the first choice, followed by methanol, ammonia, light diesel and hydrogen fuel. When pursuing lower carbon emissions, it is necessary to increase the deployment of fuel cell ships, especially ammonia fuel cell ships, which greatly increases the potential for carbon emission reduction. Considering a certain degree of scale economy in energy consumption, deploying larger ships on the route network is suitable for the zero-carbon goals of liner companies pursuing low carbon and low cost.

[0233] Fueling and speed optimization decisions:

[0234] like Fig.15 , Fig.16 , Fig.17 As shown, the Pareto frontier analysis of the five runs of MOGA in this embodiment shows that with the reduction of total carbon emissions, the fixed costs of ships and fuel bunkering costs are on an upward trend as a whole, although each fluctuates slightly. In addition, overall, the cost of fuel bunkering is usually about twice the fixed ship cost, accounting for 59%-68% of the total operating cost. As the total carbon emissions decrease, the increase in fuel bunkering costs means that the amount of fuel bunkering will also increase. Usually, in order to ensure that container ships can complete their designated route services, the amount of fuel bunkering always exceeds the fuel consumption. Among all solutions, the maximum difference between the two is 1,498 tons, which is about 25% of the amount of fuel bunkering. The fluctuations mainly come from fleet deployment decisions, which are affected by changes in ship structure and reflect the results of fuel management optimization under various realistic factors.

[0235] The average speed of all sections corresponding to the Pareto front is distributed between 15.8KN and 17.2KN, and the lowest speed is not adopted blindly. There is no obvious pattern between the average speed of all sections and the total carbon emissions. It can be seen that for green heterogeneous fleets, speed determination is an auxiliary decision. Through flexible changes, it can be closely coordinated with fleet deployment and fuel refueling to achieve carbon emission reduction in liner shipping.

[0236] Algorithm innovation and verification:

[0237] To solve the BO-MINLP proposed in this embodiment, the model is firstly linearized by linearization technology and the dual-objective problem is converted into a single-objective problem by the ε-constraint method, so as to obtain results on a commercial solver that implements an exact algorithm; secondly, a multi-objective genetic algorithm (MOGA) is customized based on the decision logic of fleet deployment, speed determination and fuel refueling to solve BO-MINLP; finally, the effectiveness of the two methods is verified by performance comparison.

[0238] The ε-constraint method can only solve the problem of a small-scale single route or a combination of two routes. Since the study considers the deployment of a heterogeneous fleet, the linearization of the original problem needs to be discretized according to the type of cargo hold tank capacity. Each match of tank capacity, power unit and alternative fuel requires 120 traversals. If the route combination is further expanded, as the number of segments to be traversed increases, the solution of Cplex (digital optimization technology) will be extremely difficult, and there will be no solution or memory overflow during the solution process. Therefore, the solution method based on the ε-constraint method and Cplex only obtained solutions to 6 instances.

[0239] However, the proposed MOGA method is able to solve all instances, including large-scale ones. Comparing the instances where both methods can obtain solutions at the same time, MOGA is slightly inferior in solution accuracy. This difference is caused by the initial solution setting of MOGA. Although feasibility is ensured, the minimum total operating cost and carbon emissions of its results are higher than those of the ε-constrained method. In addition, it can be observed that due to the difference in solution methods, the ε-constrained method prioritizes low-carbon and zero-carbon targets, while the solutions of MOGA are more diverse.

[0240] The present embodiment also provides a container liner joint optimization system, including: a dual-objective model construction module for heterogeneous green liner fleet deployment, speed optimization and fuel refueling joint optimization, which is used to construct a dual-objective model for container liner fleet deployment, speed optimization and fuel refueling joint optimization, and the model is a mixed integer nonlinear programming model; a first solution module and a second solution module, which are used to use MOGA to solve the Pareto front and the Pareto solution set; a comparison module 140, which is used to compare the first solution result and the second solution result; a solution determination module, which performs fuzzy decision-making on the generated Pareto frontier, and sorts the selected solutions according to the decision maker's preferences, so as to obtain the best solution and realize the optimization decision of integrated fleet deployment, speed determination and fuel refueling under the background of low-carbon operation.

[0241] In summary, the beneficial effects of the embodiments of the present application are:

[0242] 1. Able to cope with the complex deployment requirements brought about by the diversity of ship types to improve the operational efficiency and flexibility of liner companies.

[0243] 2. Integrated optimization of fuel filling and speed determination can not only effectively control operating costs while reducing carbon emissions, but also provide a new decision-making dimension for green liner shipping.

[0244] 3. By constructing a dual-objective mixed integer nonlinear programming model, balancing fleet deployment, speed determination and fuel refueling, low-carbon operation and simultaneous improvement of commercial value can be achieved, solving the problem that traditional optimization methods cannot take into account both environmental protection and economic benefits at the same time.

[0245] 4. The GLSP was solved by applying the ε-constraint and multi-objective genetic algorithm methods. By comparing the ε-constraint and MOGA, the superiority of MOGA in solving large-scale problems was verified.

[0246] In this application, the terms "first", "second", and "third" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance; the term "plurality" refers to two or more, unless otherwise expressly defined. Terms such as "installed", "connected", "connected", and "fixed" should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; "connected" can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to the specific circumstances.

[0247] In the description of the present application, it should be understood that the terms "upper", "lower", "front", "rear", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or module referred to must have a specific direction, be constructed and operated in a specific orientation, and therefore, should not be understood as a limitation on the present application.

[0248] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0249] The above are only preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A container liner joint optimization method, characterized in that: Applied in a container liner joint optimization system, the container liner joint optimization method is executed by a processor, and the container liner joint optimization method includes: Based on the number of heterogeneous fleet deployments, speed data and fuel filling data, a joint optimization model is constructed through the container liner joint optimization system, wherein the joint optimization model includes a dual-objective mixed integer nonlinear programming model, and the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating the total operating cost and a second objective function for calculating the total carbon emissions; Solving the joint optimization model by the container liner joint optimization system according to the ε-constraint method to obtain a first solution result; Solving the joint optimization model by the container liner joint optimization system according to a customized multi-objective genetic algorithm to obtain a second solution result; comparing the first solution result and the second solution result; The container liner joint optimization system calculates the objective function membership based on the second solution result and the fuzzy decision method, sorts the solution set of the objective function membership according to the decision maker's preference, and obtains an optimization decision, the optimization decision includes a fleet deployment decision, a speed determination decision and a fuel filling decision, and the objective function membership includes a first objective function membership and a second objective function membership.

2. The container liner joint optimization method according to claim 1, characterized in that: Solving the joint optimization model according to the ε-constraint method to obtain a first solution result includes: Combining the cargo hold tank capacity type of the ship with the available quantity of the ship, discretizing according to the V-type tank capacity; Linearizing the bi-objective mixed integer nonlinear programming model by Big-M method; The Pareto solution set of the dual-objective mixed integer nonlinear programming model is solved by a two-stage ε-constraint method to obtain a first solution result.

3. The container liner joint optimization method according to claim 2, characterized in that: The method of solving the Pareto solution set of the dual-objective mixed integer nonlinear programming model by a two-stage ε-constraint method to obtain a first solution result includes: Decomposing the dual-objective mixed integer nonlinear programming model into a first sub-problem and a second sub-problem, wherein the first sub-problem takes minimizing the total operating cost as an objective and the total carbon emissions as a constraint, and the second sub-problem takes minimizing the total carbon emissions as an objective and the total operating cost as a constraint; By setting the incremental step size to obtain the specific value of ε, a plurality of single-objective problems are formed, and the Pareto solution set is obtained by solving the single-objective problems.

4. The container liner joint optimization method according to claim 1, characterized in that: Solving the joint optimization model according to the customized multi-objective genetic algorithm to obtain a second solution result includes: Setting the ship parameters, fuel and route parameters in the joint optimization model, and setting the basic parameters of the inner genetic algorithm and the outer genetic algorithm; Initializing individuals of the outer genetic algorithm, optimizing fuel filling schemes and fleet deployment schemes for various types of ships, and calculating objective function values, wherein the individuals include the speed of the ship in each section; Sort the initialized individuals using the fast non-dominated sorting method; Perform crossover mutation on the sorted individuals to generate new individuals, calculate the objective function value, compare the individuals and update the population, and sort the updated population individuals again using a fast non-dominated sorting method; When the outer algorithm reaches the maximum number of iterations, it stops iterating and outputs the Pareto solution set; On the basis of determining the speed of each section, the inner genetic algorithm is called, and the fuel filling variable is initialized as an individual under the premise of satisfying the constraints of ship fuel combustion energy and inventory level, and the fuel filling amount and fuel filling port are decided; On the basis of determining the speed and fuel filling, the inner genetic algorithm is called to decide the fleet deployment on the basis of satisfying the unique constraint of speed selection for each segment and the constraint that the transportation volume of the freight demand does not exceed the container capacity of the cargo hold of the deployed fleet.

5. The container liner joint optimization method according to claim 4, characterized in that: Initializing the individuals of the outer genetic algorithm, optimizing the fuel filling scheme and fleet deployment scheme of various types of ships, and calculating the objective function value, including: For each individual, all routes and all ship types are traversed, and the ship type is judged whether it is suitable for the route based on the fuel that can be refueled at the starting port. The fuel refueling plan and fleet deployment plan for various types of ships are optimized based on the speed decision, and two objective function values ​​are returned. The individual is the speed of the ship in each section.

6. The container liner joint optimization method according to claim 1, characterized in that: The joint optimization model includes: The first objective function: in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuel, C represents the total operating cost of deploying a heterogeneous fleet to complete a voyage on a given route network, and C rvkf represents the alternative fuel f∈F used by power unit k on route r A The fixed cost of the V-type tank capacity ship, C if represents the price of fuel f at port i, C if′ represents the price of fuel f′ at port i, x rvkf Denotes the use of alternative fuel f∈F with power unit k A The number of V-shaped tank capacity ships deployed on route r, represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's alternative fuel f∈F A The filling amount, represents the alternative fuel f∈F used by power plant k at port i on route r A V-shaped tank capacity of the ship's fossil fuel f′∈F F The filling amount; The second objective function: in, I r represents the set of ports of call on route r, with i as the index, i represents the segment from i to i+1, |I r | is the last port of call before returning to the starting port, R represents the route set, with r as the index. V represents the cargo hold's container capacity set, with v as the index. K represents the set of power devices carried, with k as the index, f∈F A , F A represents alternative fuel, f′∈F F , F F represents fossil fuels, E represents total carbon emissions, and E f represents the carbon emission factor of fuel f, E f′ represents the carbon emission factor of fuel f′, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A V-shaped tank capacity of the ship's alternative fuel f∈F A The consumption, x rvkf Denotes the use of alternative fuel f∈F with power unit k A The number of V-shaped tank capacity ships deployed on route r, Indicates that when flying on route r, segment i, the power unit k is equipped with alternative fuel f∈F A V-shaped tank capacity of the ship's fossil fuel f′∈F F The consumption, F vkf Denotes the use of alternative fuel f∈F with power unit k A V-type tank capacity Fuel consumption rate of the ship in port, T ri represents the stopover time of the ship on route r at port i.

7. The container liner joint optimization method according to claim 6, characterized in that: The constraints of the joint optimization model are: The only constraint on speed selection for each segment is: Among them, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0. r represents the set of ports of call on route r, indexed by i, and R represents the set of routes, indexed by r; The transport volume of freight demand does not exceed the container capacity of the deployed fleet cargo hold: Among them, I r represents the set of ports of call on route r, with i as the index, R represents the set of routes, with r as the index, q ri represents the freight demand of port i on route r, V is the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuels, Denotes the use of alternative fuel f∈F with power unit k A V-type tank capacity The cargo hold capacity of the ship, x rvkf Denotes the use of alternative fuel f∈F with power unit k A The number of V-shaped tank capacity ships deployed on route r; The constraints to ensure that ships maintain a certain service frequency on each route are: Among them, I r represents the set of ports of call on route r, indexed by i, R represents the set of routes, indexed by r, V represents the set of cargo hold container capacities, indexed by v, K represents the set of power units carried, indexed by k, and f∈F A , F A represents alternative fuel, x rvkf Denotes the use of alternative fuel f∈F with power unit k A The number of V-shaped tank-carrying capacity ships deployed on route r, S represents the speed set, with s as the index, φ r represents the port service frequency of route r, T ris T represents the sailing time of route r at speed s in segment i, ris =D ri / s,T ri represents the stopover time of the ship on route r at port i, D ri represents the sailing distance of segment i on route r, w ris is a 0-1 decision variable, which determines whether the ship on route r is sailing at speed s in segment i. If yes, it is 1, otherwise it is 0.

8. A container liner joint optimization system, characterized in that: include: A model building module, for building a joint optimization model through the container liner joint optimization system based on the number of heterogeneous fleet deployments, speed data and fuel filling data, wherein the joint optimization model includes a dual-objective mixed integer nonlinear programming model, and the dual-objective mixed integer nonlinear programming model includes a first objective function for calculating total operating costs and a second objective function for calculating total carbon emissions; A first solving module, used for solving the joint optimization model according to the ε-constraint method through the container liner joint optimization system to obtain a first solution result; A second solving module is used to solve the joint optimization model according to a customized multi-objective genetic algorithm through the container liner joint optimization system to obtain a second solution result; A comparison module, used for comparing the first solution result and the second solution result; A scheme determination module is used to calculate the objective function membership based on the second solution result and the fuzzy decision method through the container liner joint optimization system, sort the solution set of the objective function membership according to the decision maker's preference, and obtain an optimization decision, wherein the optimization decision includes a fleet deployment decision, a speed determination decision, and a fuel filling decision, and the objective function membership includes a first objective function membership and a second objective function membership.

9. A container liner joint optimization system, characterized in that: include: A memory (300) and a processor (400), wherein the memory (300) stores a program or instruction that can be run on the processor (400), and the processor (400) implements the steps of the container liner joint optimization method as described in any one of claims 1 to 7 when executing the program or the instruction.

10. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or the instruction is executed by a processor, the steps of the container liner joint optimization method according to any one of claims 1 to 7 are implemented.

Images