Lane path and navigational speed collaborative optimization method and system considering carbon emission control area

Abstract

The embodiment of the invention discloses a liner path and navigational speed collaborative optimization method and system considering an emission control area. The method comprises the steps that S1, assumed conditions corresponding to liner path selection and navigational speed optimization problems of containers considering the emission control area and perishable products are defined; s2, creating a liner path selection and navigational speed optimization model which corresponds to the optimization target and considers emission control area limitation and perishable product transportation; and S3, solving the liner path selection and navigational speed optimization model through an adaptive hybrid genetic algorithm to generate a corresponding liner path and navigational speed collaborative optimization scheme. According to the method, the policy constraint of an emission control area (ECA) and the transportation time requirement of perishable products are analyzed, a mixed integer nonlinear programming model with the minimum total operation cost as the target is constructed, and an improved self-adaptive hybrid genetic algorithm is adopted for solving; and an optimal ship path and speed cooperation scheme is dynamically generated based on multiple factors such as an ECA region boundary, a fuel price and a port time window.

Description

Technical Field

[0001] This invention relates to the field of ship operation technology, and in particular to a method and system for co-optimizing liner routes and speeds considering carbon emission control zones. Background Technology

[0002] With the continued growth in global demand for perishable goods (such as fresh food and pharmaceuticals) and the deepening implementation of the International Maritime Organization (IMO) Emission Control Area (ECA) policy, the container liner shipping industry faces unprecedented multi-objective coordination challenges in its operation and management. The core of this challenge lies in the lack of a processing solution in existing technologies capable of efficiently handling multiple complex constraints and outputting optimal operational instructions. In other words, although domestic and international scholars have conducted extensive research in the field of shipping optimization and achieved numerous research results, and the relevant theories and methods have certain reference value for this application's research, the implementation of ECA policies has brought new policy requirements to shipping, leading to significant differences between shipping optimization problems under ECA restrictions and those in traditional scenarios. Simultaneously, the refrigerated cargo transportation market continues to grow, the trend towards containerization in transportation is increasingly evident, and the special characteristics of cold chain transportation place higher demands on ship scheduling and route planning. Based on existing research, the following shortcomings still exist: Firstly, the dynamic conflict between environmental regulations and engine fuel system compatibility and emission control is difficult to reconcile: To meet the stringent low-sulfur fuel requirements within the ECA region, ships must dynamically switch to higher-priced, cleaner fuels or adopt route avoidance strategies during operation. However, while traditional decision-making models (such as simply circumventing the ECA) reduce fuel costs, they significantly increase sailing time and distance. In other words, existing operation management systems rely solely on experience-based judgment or single-objective optimization, failing to address the issues of poor engine fuel system compatibility and emission control caused by differences in fuel sulfur content standards between the ECA region and overseas.

[0003] Secondly, there is a contradiction between the transportation needs and timeliness of special transport goods (risk of damage to perishable goods): perishable goods have extremely high requirements for transportation time and temperature control environment. Existing ship scheduling systems have failed to deeply integrate cargo spoilage models with navigation decisions. Any delay in navigation time (such as due to detouring around ECA or improper speed adjustment) will directly lead to the deterioration of cargo quality, that is, it cannot solve the problem of spoilage and deterioration of perishable goods caused by changes in physical environment such as temperature and time during transportation. In other words, traditional route planning systems are applicable to general cargo, and existing models cannot better respond to the unique, time-dependent quality change constraints of perishable goods. Furthermore, how to optimize the navigation scheme of perishable container liner ships under emission control areas is a problem that cannot be ignored. Thirdly, the limitations of existing optimization models and algorithms: Although existing research has focused on multiple factors such as ECA, time window, and fuel cost, it has obvious defects: (1) It decomposes multiple factors into single problems for research, lacks a systematic joint optimization framework, resulting in local optima rather than global optima in decision results, and often assumes that the speed inside and outside the ECA remains consistent throughout the voyage, without fully considering the different arrival time windows and cargo transportation requirements of each port; (2) The constructed mixed integer nonlinear programming (MINLP) model is an NP-hard problem with extremely high computational complexity. Traditional optimization algorithms are prone to getting stuck in local optima or having difficulty converging when solving large-scale practical problems, and cannot meet the needs of liner companies for real-time or near-real-time scheduling decisions; (3) It fails to provide an efficient and stable computing engine to handle uncertainties such as fuel price fluctuations, variable port time windows, and sudden weather events, and the system has poor adaptability. Summary of the Invention

[0004] Based on this, in order to address the shortcomings of existing technologies, a method and system for co-optimizing liner routes and speeds that takes into account emission control zones for the transportation of perishable goods is proposed.

[0005] To achieve the above design objectives, the technical solution of the present invention is as follows: A method for co-optimizing liner routes and speeds within emission control areas, comprising: S1. Define the assumptions for the container liner route selection and speed optimization problem considering emission control areas and perishable goods. S2. Under the given assumptions, set the corresponding optimization objectives and create a liner route selection and speed optimization model that considers emission control area restrictions and perishable goods transportation. S3. Design an adaptive hybrid genetic algorithm to solve the liner route selection and speed optimization model to generate a corresponding liner route and speed co-optimization scheme.

[0006] Implementing the embodiments of the present invention will have the following beneficial effects: To substantially address the issues of high ship operating costs, severe damage to perishable goods, and low decision-making efficiency at the technical level, this invention proposes an intelligent decision-making method that deeply integrates multiple constraints such as ECA policies, perishable goods characteristics, port time windows, and fuel costs. This method analyzes ECA policy constraints and perishable goods transportation time requirements, constructs a mixed-integer nonlinear programming model with the objective of minimizing total operating costs, and solves it using an improved adaptive hybrid genetic algorithm. Based on multiple factors such as ECA boundaries, fuel prices, and port time windows, it dynamically generates the optimal ship route and speed coordination scheme. This effectively helps shipping companies significantly reduce operating costs and improve the adaptability of engine fuel systems and the optimization control of pollutant emissions while meeting environmental protection and cargo preservation requirements. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] in: Figure 1 This is a flowchart of the basic steps corresponding to the solution described in this invention; Figure 2 (a), (b), and (c) are schematic diagrams of the ECA liner routes, segments, and sailing paths; Figure 3 This is a schematic diagram of the solution steps corresponding to the adaptive hybrid genetic algorithm described in this invention; Figure 4 This is a schematic diagram illustrating the construction of each chromosome individual using the multi-layer real number encoding method described in this invention; Figure 5 This is a schematic diagram of the cross operation described in this invention; Figure 6 This is a schematic diagram of the mutation operation described in this invention; Figure 7 This is a diagram showing the convergence effect of the algorithm in the case described in this invention; Figure 8 This is a diagram illustrating the impact of low-sulfur fuel oil prices on total cost in the case described in this invention. Figure 9 This is a diagram illustrating the impact of low-sulfur fuel prices on ECA (Electronic Cruise Control) speeds in the case described in this invention. Figure 10 This is a diagram illustrating the impact of low-sulfur fuel prices on fuel costs and sailing distance in the case described in this invention. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0010] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. It is understood that the terms “first,” “second,” etc., as used herein may be used to describe various elements, but these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, a first element may be referred to as a second element without departing from the scope of this application, and similarly, a second element may be referred to as a first element. Both the first element and the second element are elements, but they are not the same element.

[0011] To address the issue that existing shipping optimization theories do not fully cover the route selection and speed scheduling problems for perishable goods transportation within ECA (Emission Control Area) settings, this invention constructs a route selection and speed optimization scheme for perishable goods transportation in ECA areas. This scheme aims to provide shipping companies with decision-making references for route selection and speed optimization within ECA regions, enabling them to optimize transportation routes and speed decisions while meeting environmental policies, thereby reducing operating costs and ensuring cargo quality.

[0012] Based on the aforementioned design requirements, the overall architecture of this application is as follows: First, the technical problem to be solved is defined and described. Specifically, regarding the container liner shipping system, the basic transportation structure, including port clusters, shipping segments, and ECA regional distribution, needs to be defined, along with the key variables and parameters involved in transportation decisions. Furthermore, the composition of transportation time, fuel usage rules, and weekly liner service frequency are quantified and analyzed in detail to form a complete descriptive framework for the problem. Second, based on the above problem definition, the inter-port shipping segments are divided into emission control zones and non-control zones. By analyzing the impact of emission control zones on ship speed and route selection, and considering the time windows of each port and the transportation time requirements for perishable goods, a mixed-integer nonlinear programming model is constructed with the objective function of minimizing the total weekly cost of liner shipping services. This model, encompassing multiple decision variables, is a joint optimization model. Considering the nonlinearity, nonconvexity, and high-dimensional combination characteristics of this model, which are difficult to solve efficiently using traditional optimization methods, an improved adaptive hybrid genetic algorithm is introduced to solve this problem. Finally, to verify the applicability of the models and algorithms involved in this application, empirical analysis and sensitivity analysis are required. Therefore, this application also selects a container liner route as the research object, conducts numerical analysis and numerical experiments, compares the optimization results with traditional solutions, and provides corresponding experimental results. The technical knowledge involved in this application includes: liner shipping refers to an operational mode in which ships call at various ports in a predetermined order on a specific route and ultimately return to the port of origin, such as... Figure 2 As shown in (a). On a given liner route, the distance between two adjacent ports of call is defined as a segment. Considering the impact of the ECA, each segment is divided into two parts: within the ECA and outside the ECA, as shown in (a). Figure 2 As shown in (b). Because liner ships must use expensive low-sulfur fuel oil within the ECA area, while cheaper conventional fuel oil can be used outside the ECA, shipping companies often choose to circumvent the ECA, resulting in multiple shipping routes between ports, such as... Figure 2 As shown in (c), both Scheme 1 and Scheme 3 bypass the ECA area to reduce the use of low-sulfur fuel; Scheme 2 is the shortest route that directly crosses the ECA area, and must use low-sulfur fuel throughout the entire route.

[0013] Based on the above design framework, this embodiment proposes a method for co-optimizing liner routes and speeds within emission control areas, such as... Figure 1 As shown, the method includes the following steps: S1. Define the assumptions for the container liner route selection and speed optimization problem considering emission control areas and perishable goods; S2. Under the assumptions, set the corresponding optimization objective and create a liner route selection and speed optimization model that considers emission control area restrictions and perishable goods transportation; S3. Design an adaptive hybrid genetic algorithm to solve the liner route selection and speed optimization model to generate a corresponding liner route and speed co-optimization scheme.

[0014] In some specific embodiments, the applicable scenario of this invention refers to a container liner route with N arrival time restrictions, k types of perishable cargo transportation time requirements, and considering multiple navigation schemes between adjacent ports during the planned voyage period. In order to better fit the actual scenario and ensure the feasibility and applicability of the subsequent analysis model, this invention first makes reasonable assumptions about factors such as speed and fuel consumption, ship sailing time, and transportation characteristics of perishable goods, based on the actual transportation situation. That is, the corresponding assumptions in S1 are: (1) the container ships configured on the route are of the same type; (2) the order of the ships calling at ports on the liner route is determined; (3) the ships pass through the ECA (3) The region uses a fuel conversion strategy and the ship's fuel change time is negligible; (4) Port congestion is not considered; (5) Auxiliary engine fuel consumption is included in the ship's fixed costs, and this invention only considers the main engine fuel consumption; (6) During the decision-making period, the freight demand between ports is known and remains basically unchanged; (7) The service frequency of the container liner route is once a week, that is, each ship calls at each designated port once a week.

[0015] In some specific embodiments, although certain assumptions have been set in the previous step, the introduction of Emission Control Area (ECA) restrictions and requirements for the transport of perishable goods makes it impossible to directly apply traditional route selection and speed optimization models. Specifically, ECA areas require ships to use low-sulfur fuel (e.g., sulfur content ≤ 0.1%), which increases fuel costs and limits speed selection; simultaneously, the spoilage risk of perishable goods requires strict control of transport time within a time window. The interaction of these constraints significantly increases the complexity of the problem (e.g., route selection must consider not only route distance but also a dynamic trade-off between ECA avoidance strategies and timeliness). Therefore, before constructing the optimization model, this invention first quantitatively analyzes elements such as the composition of transport time and fuel type switching rules to clarify the technical representation and constraint boundaries of each variable, facilitating the subsequent establishment of a complete and computable problem description framework.

[0016] Specifically, the analysis of the container liner route selection and speed optimization problem considering emission control areas and perishable goods reveals the following conclusions: First, this problem exhibits multi-dimensional coupling characteristics in route selection. In traditional liner shipping scheduling, route selection is typically based on the principle of minimizing geographical distance or sailing time. However, with the introduction of ECA emission regulations and the time constraints of perishable goods, route decision-making is no longer limited to geometric route optimization but evolves into a combined optimization problem constrained by multiple operational and compliance conditions (constraints include whether each segment crosses an ECA area; the set of feasible routes between ports (including detour options); recommended or speed-limited sections on different segments, etc.). Furthermore, due to the mandatory use of low-sulfur fuel oil or the installation of exhaust gas desulfurization devices (such as scrubbers) within ECA areas, ships must perform fuel switching operations when entering and leaving ECA boundaries. This process not only incurs additional fuel costs (low-sulfur fuel oil is significantly more expensive than heavy fuel oil) but may also affect sailing schedules due to operational delays. Therefore, route design must assess the economic and time costs of crossing ECA areas, and some route designs that were originally geographically shorter may become impractical due to frequent crossings of multiple ECA areas. Secondly, this problem possesses nonlinear and multi-objective trade-off attributes in speed optimization. Speed ​​setting directly affects core variables such as fuel consumption, carbon emission levels, transportation cycle, and cargo quality. However, due to multiple constraints and conflicting objectives in speed optimization—namely, speed restrictions within the ECA, timeliness requirements for perishable goods transportation, port time window constraints, and carbon emission quota pressures—the speed optimization problem described in this application is essentially a process of seeking the optimal solution for a nonlinear relationship among multiple objectives such as fuel cost, spoilage loss, on-time performance, and emission compliance. It is difficult to obtain a globally optimal strategy through empirical rules or enumeration methods. Based on the above analysis, to ensure that the optimization model can truly reflect the actual operating environment, this invention needs to introduce the following core constraints during the modeling process and quantify them, so that the various constraints work together to significantly compress the feasible solution space, thus solving the problem that traditional linear programming or static path algorithms cannot effectively solve: the core constraints include at least sailing time-related constraints, time window-related constraints, and constraints on decision variables and their ranges. Therefore, it is necessary to construct a mixed integer nonlinear programming model that can integrate integer and continuous variables to collaboratively optimize the decision variables involved in the selected port navigation scheme, determine the speed inside and outside the emission control area of ​​each segment, and select the container loading and unloading rate of each port, such as route, speed, ship allocation and loading and unloading scheduling.

[0017] Based on the above analysis, for the optimization of container liner routes and speeds considering the transportation of perishable goods under emission control areas, from the perspective of shipping companies, under the premise of meeting transportation demand and port time windows, a corresponding optimization objective is set. That is, the optimization objective is set to minimize the total operating cost during transportation, including fixed costs of ships, costs of perishable goods spoilage and loss, cargo inventory costs, costs of ship delays, loading and unloading costs at each port, and fuel consumption costs of ships.

[0018] The corresponding liner route selection and speed optimization model considering emission control area restrictions and perishable goods transportation is as follows: (1) st (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) Formula (1) represents the objective function for minimizing the operating cost of a liner shipping route within a round trip. The various elements within this objective function comprehensively reflect the important costs involved in the route operation. The first term represents the fixed costs of all vessels serving the route, which mainly include vessel asset costs, employee costs, lubrication costs, and insurance premiums. The second term represents the cost of perishable goods spoilage during a round trip. The third term represents the inventory cost of container cargo for a vessel during a round trip, which is also equivalent to the opportunity cost of the shipping company's vessel operation. The fourth term represents the penalty cost for arriving late at the port during a round trip. The fifth term represents the container loading and unloading cost during a round trip. The sixth term represents the fuel cost during a round trip.

[0019] Specifically, the first item, fixed costs of the vessel, includes vessel asset costs, employee costs, lubrication costs, insurance premiums, etc. If the weekly fixed cost of a vessel for one round trip is... The number of vessels allocated to the route to meet the weekly service frequency is The fixed cost of a ship for a round trip is as shown in formula (21): (twenty one) Regarding the second item: the cost of spoilage and loss of perishable goods, container shipping companies transporting goods on their routes... Perishable goods, No. Perishable goods from the port of origin Transport to the port of destination for From the moment of containerization, as transportation time (including the total sailing time from the port of origin to the port of destination, and the loading and unloading time at each port) increases, its quality will be affected. The spoilage loss function for perishable goods is: .

[0020] If the first The unit decay cost of perishable products is The total cost of spoilage is as shown in formula (22): (twenty two) For the third item: inventory cost, specifically container inventory cost, it is equivalent to the shipping company's opportunity cost and is a significant component of total operating costs. Total container inventory cost can be calculated based on the transit time of containers transported on a given liner route segment, assuming a unit inventory cost of... Flight segment The number of containers transported is The total inventory cost is as shown in formula (23): (twenty three); The fourth item, penalty cost, is the cost of vessel delay. It represents the penalty cost for a vessel to arrive late to port for a round trip. It is mainly related to the sailing time of each segment, the port loading and unloading time, and the port time window. Assuming the vessel leaves the port... The delay time is Ships in port The cost of late arrival penalties is The total delay cost for a ship's round trip is given by formula (24): (twenty four) Regarding the fifth item: container handling costs, shipping companies sign cooperation agreements with various port operators to provide container handling services for inbound vessels. According to the agreement, each port operator provides the shipping company with a container handling rate plan. Different loading and unloading rates correspond to different loading and unloading costs, and the loading and unloading time in port is related to the total volume of containers handled at the port. It is related to the loading and unloading rate, assuming the unit loading and unloading cost is... , The 0-1 variable represents the ship in port. Should a loading / unloading rate scheme be selected? The total loading and unloading cost for a round trip is as shown in formula (25): (25) Regarding the sixth item: fuel cost, i.e., fuel consumption cost, based on the functional relationship between fuel consumption and speed, the hourly fuel consumption of the ship in each leg of the journey can be calculated as follows: Because the ship navigates through the ECA area and uses different speeds inside and outside the ECA area, the ship will be in the following segment of the journey. The fuel consumption function is shown in formula (26): (26) in This is the fuel consumption coefficient. 0-1 variables represent the ship's position on the voyage. Should a flight plan be selected? For navigation, assuming the fuel prices inside and outside the emission control area are respectively The fuel consumption cost of a ship during a round trip is given by formula (27): (27); The corresponding constraints are formulas (2) to (20); they can be divided into three parts according to the type of constraint: time-related constraints, time-window related constraints, and decision variables and their range constraints. Among them, Formula (2), Formula (8)-Formula (15) belong to the navigation time related constraints; Formula (2) represents the total navigation time of the ship on each segment, which is obtained by adding the navigation time inside / outside the emission control area; Formula (8) represents the loading and unloading time of a ship in port, which describes the time a ship spends at each port of call. This time is mainly determined by the time required for container loading and unloading operations. Formula (9) limits the earliest and latest allowed time for a ship to leave the port. Formula (10) is a constraint on the time a ship is late at each port. Formulas (11) and (12) represent the time constraints for a ship to arrive at the port, where 168 is the number of hours in a week. Formulas (13) and (14) represent the waiting time constraints for a ship at the next port. Formula (15) represents the service frequency constraints for a ship, which means that the time for a liner's round trip is an integer multiple of a week. Formulas (5), (6), and (7) belong to time window related constraints; Formula (5) represents the time window constraint for ships arriving at each port; Formula (6) represents the first... The total transport time for perishable goods from the port of origin to the port of destination includes sailing time and loading / unloading time at the port; Formula (7) represents the total transport time for perishable goods from the port of origin to the port of destination. Allowable sailing time constraints; Among them, formulas (3), (4), (16)-(20) are decision variables and their range constraints; formula (3) represents the ship's position on the voyage. The ship must choose a navigation plan; Formula (4) indicates that the ship is in port. Only one loading and unloading rate scheme can be selected for loading and unloading operations; Formula (16) represents the maximum number of ships on the route; Formulas (17) and (18) represent the speed limits of ships within / outside the emission control area; Formulas (19) and (20) represent 0-1 variable constraints.

[0021] In some specific implementations, given that the liner route selection and speed optimization problem considering emission control zones is essentially an NP-hard problem, and further requires consideration of factors such as the selection of loading and unloading rates at various ports, port time window constraints, and the timeliness requirements for perishable goods transportation; and involving multiple decision variables and complex nonlinear constraints, particularly the nonlinear relationship between speed and fuel consumption, port time window limitations, and the timeliness requirements for perishable goods transportation, the problem becomes enormous and difficult to solve. With the increase in the number of ports and available routes, the combinatorial space grows exponentially, making it difficult for traditional linear programming or integer programming methods to obtain high-quality solutions within a reasonable timeframe. Therefore, a genetic algorithm (GA) was chosen as the solution tool. A GA is an optimization search algorithm that simulates the biological heredity and evolutionary processes in nature. Its basic idea originates from the principles of natural selection and gene mutation in biological genetics: in nature, chromosomes carry genetic information, and individuals transmit and optimize traits through gene replication and recombination; while chromosome mutations may introduce new traits, thereby driving the population to evolve towards a better direction. In other words, the genetic algorithm has strong global search capabilities, is suitable for nonlinear and discrete optimization problems, and improves the solution efficiency through population parallel evolution. Therefore, this algorithm is a suitable choice for solving problems.

[0022] However, the standard genetic algorithm (GA), as a general global search algorithm, has the following inherent defects when dealing with complex optimization problems with high constraints and multiple peaks (i.e., multiple local optima), which makes it unsuitable for solving the technical problems of this application or it will have the problem of low efficiency. Therefore, when the standard genetic algorithm is directly applied to the mixed integer nonlinear programming model described in this application, the optimization effect will be poor. The main reasons are: (1) It is easy to get stuck in local optima: Since the set of feasible solutions in this application has a large number of local optima due to its complex constraints (ECA region, time window, corruption cost), the crossover and mutation operations of the standard genetic algorithm are random. After the population diversity decreases in the later stage of evolution, the algorithm is easy to get stuck around a local optimum and cannot get out and find a globally satisfactory optimization solution. For example, the standard genetic algorithm may have a fixed route with a low total cost, but since it cannot finely adjust the path of a single route or the loading and unloading rate of a specific port, it may miss the globally better solution that can significantly reduce the total cost by fine-tuning these local variables (such as avoiding high fuel cost routes or reducing corruption losses). (2) Insufficient local search capability and slow convergence speed: The mutation operation of the standard genetic algorithm is usually a blind, small-amplitude random perturbation. For the solution in this application, the quality strongly depends on local variables, such as continuous variables that need fine adjustment, such as speed and loading / unloading rate. It requires a large number of iterations to find a local improvement direction, resulting in low search efficiency and slow convergence speed. (3) Insufficient ability to handle complex constraints: The standard genetic algorithm is difficult to search efficiently under the premise of meeting the hard constraints such as ECA regulations and the timeliness of perishable goods. It is easy to generate a large number of infeasible solutions and the optimization process is inefficient. For example, the speed value generated by a random mutation may cause the ship to travel at high speed outside the ECA area but use high-sulfur fuel oil in the ECA area, thus violating the regulations; or cause the transportation time to exceed the shelf life of perishable goods.

[0023] Based on the above analysis, it is evident that the standard genetic algorithm needs to be improved. An improved adaptive hybrid genetic algorithm is designed by introducing a local search mechanism on the basis of the traditional genetic algorithm. Specifically, based on the classic genetic algorithm, a multi-layered real-number encoding of the design path, speed, and loading / unloading rate is used to enhance the expression ability of the chromosomes. By combining crossover and mutation operations with the local search mechanism, the quality of the current solution is improved, the convergence performance of the algorithm is enhanced, and trapping in local optima is avoided, thereby effectively improving the search efficiency of the algorithm.

[0024] like Figure 3 The specific solution steps corresponding to the adaptive hybrid genetic algorithm include: S1: Initialize the parameters of the genetic algorithm and generate an initial population. The method for generating the initial population includes constructing each individual using multi-layer real number encoding and setting the current generation number. k =0, population size is N; S2: Determine the fitness function and calculate the fitness value of all individuals in the population. The fitness function is the reciprocal of the objective function, that is, the reciprocal of the total operating cost is used as the fitness evaluation basis; S3: Perform a selection operation on the population, selecting some individuals as individuals for the next generation of the population. The selection operation adopts a roulette wheel strategy; S4: According to the set crossover probability Perform crossover on the individuals selected in S3; S5: Perform crossover according to the set mutation probability. Perform mutation operations on the individuals after crossover in S4 to generate new individuals; S6: Define a local search mechanism and optimize the new individuals according to the local search mechanism; S7: Replace the corresponding individuals in the current population with the individuals optimized by local search in S6, complete the population update, and recalculate the fitness values ​​(i.e., total operating costs) of all individuals in the new generation population; S8: Determine whether the preset termination conditions are met: if the current iteration number has reached the preset maximum iteration number, or the objective function value (total operating cost) of the best individual in the population for M consecutive generations has not exceeded the set change threshold, then terminate the iteration and output the current optimal solution as the final optimization scheme; otherwise, return to S3 (perform selection, crossover, and mutation operations) and continue the iteration optimization of the next generation; where M is the upper limit of the number of generations in which the objective function value has not been improved.

[0025] In some more specific embodiments, chromosome encoding is crucial in the genetic algorithm solution process because genetic algorithms cannot directly handle the various parameters in the problem. Therefore, before computation, these parameters must be converted into chromosomes in the genetic space according to certain rules; this process is called encoding. Common encoding methods currently include binary encoding, character encoding, and real number encoding. Considering the characteristics of the aforementioned decision variables, the initial population generation method in S1 uses multi-layer real number encoding to construct each chromosome individual, setting the current generation number... k=0, population size is N; for example, if we assume there are 8 ports of call on a route, the shipping company needs to decide on the navigation plan between ports, the ship speed inside and outside the emission control area, and the loading and unloading rate plan at each port; then a four-layer coding method with a chromosome size of four rows and eight columns is used to generate the initial population. The chromosome of the first layer represents the selection of a specific navigation plan, which is encoded using natural numbers. The numbers 1, 2, and 3 represent the three navigation plans between ports. For example, the chromosome of the first layer represents that port 1 to port 2 adopts plan 1, port 2 to port 3 adopts plan 3, and so on. The chromosomes of the second and third layers represent the ship speed inside and outside the ECA, respectively, and correspond to the selection of the navigation plan in the first layer. In order to ensure the rationality of the coding, if the distance inside or outside the emission control area in the corresponding navigation plan is 0, then the speed is also 0. The chromosome of the fourth layer represents the loading and unloading rate plan provided by each port of call, which is also encoded using natural numbers. Each integer corresponds to a specific combination of loading / unloading rate and unit cost, such as Figure 4 As shown.

[0026] In some more specific embodiments, in order to improve the convergence efficiency of the population algorithm and ensure the feasibility of the initial solution, an initial population generation strategy is also designed to optimize the initial population generation method in S1 above. The initial population generation strategy includes the following steps: Step 1: Initialize parameters, set the corresponding population size, and import raw data such as port information, ECA internal and external route plans, and cargo OD (including perishable goods), and set the number of individuals n=0; Step 2: Use multi-layer real number coding to encode the selected scheme to construct the corresponding chromosome individual. That is, starting from the port of origin, based on the OD pair of cargo between ports, combined with the ECA area, select the navigation route, the speed inside and outside the ECA, and the loading and unloading rate scheme to construct each individual. Step 3: Determine whether the arrival time of the ship under the selected route and speed meets the preset constraints. If it does, proceed directly to Step 4; if it does not, adjust the relevant genes in the chromosome, i.e., perform gene repair treatment, and then proceed to Step 4, such as appropriately increasing the speed or adjusting the loading and unloading plan. The preset constraints include, but are not limited to: (1) port time window restrictions; (2) perishable goods delivery time requirements. Step 4: After completing the construction of the current individual, verify whether all time nodes from the port of origin to the final port (including the arrival time of each port and the total transportation time of perishable goods) meet the preset constraints; if they do, add the individual to the initial population and set the individual count n=n+1; if they do not meet the constraints, return to Step 3 for secondary repair or discard the individual. Step 5: Iteration control: Set the individual counter n = n + 1; then, determine whether n has reached the preset population size; if not, return to Step 2 and start the next iteration to generate new individuals; if it has reached the preset population size, output the complete initial population and end the initial stage.

[0027] Preferably, during the initial population generation and evolution, individuals represented by chromosomes may violate port time window constraints or the timeliness requirements for transporting perishable goods due to random encoding or crossover mutation operations. To ensure the feasibility of the solution, this invention designs the following gene repair rules to adjust chromosomes that do not meet the constraints, so that they satisfy the constraints set by the model. Specifically, the corresponding gene repair rules include: for individuals that violate the port time window, i.e., the ship's arrival time is earlier than the earliest acceptable time or later than the latest time, adjustments are made based on the difference between the actual arrival time and the time window. For example, if the ship arrives earlier, the route remains unchanged, and the arrival time is delayed by reducing the speed of the previous segment or extending the loading and unloading time at the previous port; if the ship arrives later, the speed of the corresponding segment is increased first, or a faster loading and unloading rate scheme is selected at the previous port to reduce the time spent in port.

[0028] For individuals whose delivery time for perishable goods exceeds the prescribed limit, the repair rule is to prioritize increasing the speed outside the ECA and selecting a high loading and unloading rate to compress the overall transportation time. If the time requirement still cannot be met, a shorter voyage can be selected as an alternative.

[0029] Additionally, to ensure better repair results, after cross-mutation, a mismatch may occur between the speed and the path selection scheme. That is, in a navigation scheme, the ship may navigate within the ECA, but the speed outside the ECA may not be 0. In this case, the individual needs to set the corresponding speed to 0. Finally, replace the individual with the repaired, normal individual.

[0030] In some more specific embodiments, to ensure the feasibility of individuals and improve the search efficiency of the genetic algorithm, given that the optimization objective of this application is to minimize the operating cost of container liner shipping while satisfying ECA policies and the timeliness requirements for transporting perishable goods, the smaller the objective function, the better the solution. Therefore, the fitness evaluation of all individuals in the population is based on selecting the reciprocal of the objective function as the fitness function to evaluate the quality of individuals. The objective function is to minimize operating costs; the smaller the objective function, the better the quality of the solution and the higher the fitness. The specific formula is as follows: (28) In the formula: For the fitness function, The objective function is the total cost. This represents the iteration number.

[0031] In some more specific embodiments, the design of the selection operator has a crucial impact on the algorithm's performance. A reasonable selection strategy can not only improve the algorithm's global search capability but also significantly accelerate the convergence speed. Considering the complexity of the joint speed optimization and path selection model constructed in this invention, and the parameter space characteristics brought about by the multi-layer chromosome structure, a roulette wheel selection strategy is adopted in S3. The core idea of ​​this method is: based on the fitness value of each individual, the probability of being selected is proportionally allocated, so that individuals with higher fitness values ​​have a greater probability of being selected and inheriting into the next generation, while individuals with lower fitness values ​​also have a certain probability of being selected, thereby maintaining population diversity while preserving partial solution inheritance. Specifically, this refers to performing a roulette wheel selection operation on the population to select some individuals as individuals for the next generation; that is, the higher the fitness value, the greater the probability of entering the next generation. The calculation method is as follows: Step 1: Obtain the fitness values ​​of all individuals in the population; Step 2: Calculate the sum of fitness values ​​for all individuals in the current population. Then, calculate the ratio of each individual's fitness value to the sum of fitness values ​​to obtain the selection probability of that individual. The corresponding formula is: (29) In the formula: i Describing the kth generation i Individual, For population size, Chromosomes The probability of being selected.

[0032] Step 3: Construct a cumulative probability interval based on the selection probability, and simulate the rotation of the "roulette" by randomly generating numbers to complete the individual selection.

[0033] In some more specific embodiments, in order to maintain population diversity and accelerate convergence, this invention adopts a two-point crossover method for the encoding of multi-layer chromosomes, combined with an adaptive crossover probability mechanism; the crossover probability is dynamically adjusted according to the distribution of individual fitness, thereby more effectively balancing the global and local aspects of the search.

[0034] S4: According to the set crossover probability After S3 selection, a crossover operation is performed on the individuals. To improve the algorithm's convergence, in the genetic operation, for each pair of parent chromosomes before crossover, the crossover probability is adaptively and dynamically adjusted based on the current population's fitness distribution, i.e., based on the fitness value. The calculation method is as follows: (30) in: The initial crossover probability, This represents the average fitness of the population. For individuals with higher fitness among the crossover individuals, It is a constant between 0 and 1.

[0035] The specific operation involves two-point crossover on the second and third chromosome layers. First, two crossover points are randomly selected within the length range of each chromosome layer. Then, gene segments located within the crossover interval between the two parent chromosomes are exchanged (the corresponding positions of the parent chromosome segments are swapped with the corresponding positions of the offspring chromosomes), thus generating two new offspring individuals. This process not only maintains the continuity of the original structure but also effectively enhances information exchange and diversity within the population. Assuming that in a certain crossover process, the randomly selected crossover points are the 3rd and 6th positions, the diagram of the crossover operation is as follows: Figure 5 The diagram illustrates the process of exchanging corresponding segments between chromosomes, where the chromosome before crossing over represents the parent chromosome, and the chromosome after crossing over represents the offspring chromosome.

[0036] Simultaneously, to avoid the population getting trapped in local optima, a mutation operation is performed on the individuals after crossover. That is, according to a set mutation probability... Perform mutation operations on the individuals after S4 crossover to generate new individuals; the specific formula and rules are as follows: Similarly, the mutation probability also adopts an adaptive dynamic adjustment method, as described above for the crossover probability. The calculation method is as follows: (31) in: This represents the initial mutation probability. This represents the average fitness of the population. For individuals with higher fitness among those that have mutated, It is a constant between 0 and 1.

[0037] That is, this mutation operation is mainly performed on the first and fourth layers of the multi-layered real-number encoded chromosome structure; The first layer of chromosomes represents the navigation scheme between segments. The mutation operation performs mutation operations on all gene positions in this layer according to the preset mutation probability p. The new values ​​are randomly selected from the interval [1, 3] according to a uniform distribution, which means that the navigation scheme may be changed to scheme 1, 2 or 3.

[0038] The fourth layer of chromosomes represents the loading and unloading rate scheme for each port. In the mutation operation, based on the set mutation probability, all gene positions in this layer are mutated and changed to uniform integer random values ​​in the interval [1, 5] to reflect the operation selection under different loading and unloading efficiencies.

[0039] A diagram illustrating the mutation operation is shown below. Figure 6 As shown, during this mutation process, multiple genes in the first and fourth layers underwent mutations. In the first layer, the first gene changed from 1 to 2, the third gene changed from 2 to 3, and the sixth and seventh genes changed from 2 and 1 to 3 and 3 respectively, indicating that the ship's navigation path between ports was adjusted. In the fourth layer, the third, fourth, and fifth genes changed from 1, 5, and 4 to 4, 2, and 1, reflecting a reselection of the loading and unloading scheme. The second and third layers did not participate in the mutation process.

[0040] In some more specific embodiments, to improve the local search capability of the genetic algorithm and avoid it getting stuck in local optima during the search process, this invention introduces a local search mechanism. Specifically, a route segment is randomly selected, its navigation scheme is changed, and the fitness is recalculated based on the new navigation scheme or by increasing or decreasing the speed. For port loading and unloading schemes, the loading and unloading rate of a specific port is selected for adjustment, the scheme is adjusted, and the total transportation cost is recalculated. By comparing the fitness of the adjusted solution with that of the current solution, if the adjusted solution is better, the current solution is replaced with the new solution, and the genetic operation continues. For speed search, based on the speed value of the current solution, the step size is dynamically adjusted to adjust the speed, ensuring that the adjusted speed still complies with ECA regulations, and a new fitness is calculated. Specifically, the local search mechanism is defined, and the specific steps for optimizing new individuals according to the local search mechanism are as follows: Step 1: In the current iteration, first perform the basic operations of the genetic algorithm on the previous generation population to generate new candidate individuals; Step 2: Define three local search operations; Operation 1: Increase or decrease the speed of a certain segment by 0.5 knots; Operation 2: Replace the current segment's navigation plan, i.e., select other alternative plans from the candidate paths (e.g., switch from plan 2 to plan 1); Operation 3: Change the current port's loading and unloading rate plan; Step 3: Randomly select one of the three predefined operations in Step 2, and perform a local search operation (i.e., a perturbation operation) on the corresponding gene position of the individual to generate a new individual, and re-evaluate the fitness value of the new individual; Step 4: Fitness comparison and solution update: Calculate the total operating cost (i.e., fitness value) corresponding to the new individual, and compare it with the total operating cost of the original individual. If the total operating cost of the new individual is lower than that of the original individual (i.e., the fitness value of the new individual is better than the original solution), then the new individual replaces the original individual in the current population. Otherwise, the original individual is retained to avoid the transmission of inferior solutions to the next generation and to ensure that the population quality does not decline. In addition, during the update process, it is determined in real time whether the termination condition has been met. That is, when the number of algorithm iterations exceeds the preset maximum number of iterations Max Gen, or when the objective function value obtained by the algorithm has not changed in multiple iterations (the upper limit of the number of generations in which the objective function value has not been improved is set to "Maxnone"). Specifically, after each round of iteration, the best individual in the current generation of the population and its corresponding objective function value are recorded, and the current generation's best solution is compared with the recorded historical best solution. If the current generation's best solution is better than the historical best solution, the historical best solution is updated to the current generation's best solution, and the number of generations in which the objective function value has not been improved is reset to 0; otherwise, the number of generations in which the objective function value has not been improved is increased by 1. If the current number of iterations is greater than the maximum number of iterations "Max Gen", or when the number of generations in which the objective function value has not been improved is greater than "Max none", then the algorithm terminates and outputs the best individual. By introducing local search, this genetic algorithm not only retains its powerful global search capability, but also allows for fine-tuning in local regions, thereby improving the overall performance and search efficiency of the algorithm.

[0041] Based on the same inventive concept, this invention also proposes a collaborative optimization system for liner routes and speeds considering emission control areas, comprising: The system includes a preprocessing unit for defining the assumptions for optimizing the route selection and speed of container liner shipping, taking into account emission control areas and perishable goods; a model creation unit for setting the corresponding optimization objective under the assumptions and creating a liner route selection and speed optimization model that considers emission control area restrictions and perishable goods transportation; a model solving unit for designing an adaptive hybrid genetic algorithm to solve the liner route selection and speed optimization model; and a scheme output unit for outputting the liner route and speed co-optimization scheme determined by the optimized liner route selection and speed optimization model.

[0042] Based on the same inventive concept, the present invention also proposes a computer-readable storage medium including computer instructions that, when executed on a computer, cause the computer to perform the method described thereon.

[0043] The following case study uses actual shipping routes as a background and combines collected and randomly generated relevant data for analysis. The effectiveness and applicability of the constructed model and designed algorithm are evaluated through verification of their solution process. Based on this, theoretical support and practical guidance are provided for container liner companies' operational decisions regarding the transportation of perishable goods under the ECA framework.

[0044] Data Collection: Assume a shipping company operates liner services on a Southeast Asian route. The port calls and segments of the vessels on this route are as follows: Qingdao—Shanghai—Hong Kong—Saigon—Laem Chabang—Hong Kong—Xiamen—Shanghai—Qingdao, a total of 8 port calls, forming a circular route. The four ports marked within the dashed box in the diagram—Qingdao, Shanghai, Hong Kong, and Xiamen—are all within the ECA (Emission Control Area). Therefore, vessels must choose appropriate routes and switch to low-sulfur fuel when entering or leaving these ports. Taking the Qingdao to Shanghai segment as an example, since both the origin and destination ports are within the ECA, the entire voyage is affected by ECA policy, and there are three optional navigation routes. Each navigation option represents the distance within / outside the ECA; 0 indicates no meaning, as shown in the table below.

[0045] Table of sailing distances (nautical miles) for each leg of the journey

[0046] Based on the above, this route has six ports of call and is operated by vessels of the same type with a capacity of 14,000 TEU. Assuming the shipping company is responsible for transporting 15 types of perishable goods, the vessel's maximum and minimum speeds are 25 knots and 15 knots respectively, the fuel consumption coefficient is 0.012, the vessel's weekly fixed cost is $150,000, and the unit inventory cost of containers is $0.5 per TEU.-1 ·h -1 The price of low-sulfur fuel oil is $650 per ton. -1 The price of regular fuel is $350 per ton. -1 Each port of call offers shipping companies four selectable container handling rate options, assuming that the handling rate and unit cost are the same for refrigerated and regular containers. Major ports such as Qingdao and Shanghai offer handling rate options of 50, 75, 100, and 125 TEU·h. -1 The loading and unloading rate schemes for the other ports are 50, 60, 75, and 100 TEU·h. -1 The corresponding loading and unloading costs for different loading and unloading schemes at various ports are US$475,550,625,750 per TEU. -1 Meanwhile, the following parameters are randomly generated in this case: the penalty cost incurred by a vessel for late port calls follows a uniform distribution. US Dollar -1 The container volume carried by a ship in each leg of its voyage follows a uniform distribution. TEU, the loading and unloading volume of ships at each port follows a uniform distribution. TEU. Assuming the start time of the port of origin is 0, the start times (in hours) of other ports are determined according to the formula... Generate, where the time window length for each port is... Follows uniform distribution h, the time limit for the transportation of perishable goods is based on the formula. The unit decay cost of perishable goods follows a uniform distribution. US Dollar TEU -1 ·% -1 The decay rate of perishable products changed from a uniform distribution %·h -1 Randomly generated. To facilitate the generation of OD streams, two new variables are introduced. : No. Port of origin for perishable goods : The port of destination for the k-th category of perishable goods. Then the transport volume of the k-th category of perishable goods will increase from the original... Become and obeys a uniform distribution TEU, port of origin for Class k perishable goods From uniform distribution The destination port for the generation of the kth type of perishable goods From uniform distribution This generates the number of refrigerated containers transported for each flight segment.

[0047] To ensure the rationality and effectiveness of OD pairs between ports, specific constraints are set: avoid the origin port and destination port being the same port, especially when considering repeated calls, strictly exclude such situations, and prevent the generation of invalid or unrealistic transport routes.

[0048] Simultaneously, the optimal cooperative scheme is determined using the method described in this invention. In this example, the hardware configuration is a computer equipped with an Intel Core i7 processor (2.50GHz) and 16GB of memory. The proposed adaptive hybrid genetic algorithm is implemented using MATLAB R2021b, with the following algorithm parameters: population size of 100, maximum number of iterations of 500 generations, upper limit of 100 generations for which the objective function value remains unchanged, initial crossover probability of 0.8, and initial mutation probability of 0.1. The calculation results obtained through the above scheme are shown in the table below.

[0049] Example Optimization Results Table

[0050] Note 1): A speed of 0 indicates that it is meaningless.

[0051] As shown in the table above, the dwell time of ships varies across ports. This difference is primarily due to variations in container loading and unloading volumes at each port, and the different loading and unloading rates chosen by shipping companies for different ports, thus affecting the ship's port operation time. Regarding speed, it can be observed that speeds outside the ECA area are generally higher than those within the ECA area. This is because low-sulfur fuel oil is required within the ECA area, and its price is significantly higher than that of conventional fuel oil used outside the ECA area. To control fuel costs, ships typically reduce speed within the ECA area to minimize fuel consumption, while increasing speed outside the ECA area to ensure on-time delivery. Therefore, the speed pattern exhibits a "lower inside, higher outside" characteristic.

[0052] To verify the effectiveness of this invention, a comparison was made between a traditional genetic algorithm and an adaptive hybrid genetic algorithm. The vertical axis represents the total cost, and the horizontal axis represents the number of algorithm iterations. The algorithm convergence curve is shown in Figure [Figure number missing]. Figure 7 As shown. By Figure 7 As can be seen, the algorithm described in this invention achieves rapid cost reduction in the initial stage, demonstrating strong convergence ability. Furthermore, with the increase in the number of iterations, the overall cost reduction trend becomes more significant, eventually converging to a better solution, indicating its strong global search capability. In contrast, traditional genetic algorithms converge more slowly during iteration and are prone to getting trapped in local optima, resulting in a significantly lower quality solution than the algorithm described in this invention.

[0053] Meanwhile, considering the perishable goods liner shipping within emission control areas, changes in fuel prices significantly impact speed and sailing time, while port time windows have a substantial effect on vessel time in port. Since vessel sailing time and port time directly relate to the cost of perishable goods spoilage and the total cost of liner shipping, this invention conducts a sensitivity analysis on fuel prices and port time windows. The specific analysis results are as follows: Analysis of the impact of fuel prices To analyze the impact of low-sulfur fuel oil price fluctuations on the total operating cost of liner shipping, it is assumed that the price of regular fuel oil remains constant, while the price of low-sulfur fuel oil fluctuates within a certain range. The results are as follows: Figure 8 As shown, the original plan assumes the shipping company chooses the shortest route and uses the same speed both inside and outside the ECA, calculating the operating cost; Figure 8 As can be seen, with the continuous rise in low-sulfur fuel oil prices, the total operating costs of both schemes show an upward trend, reflecting that fuel prices, as an important component of transportation costs, have a significant impact on overall operating costs. However, compared to the original scheme, the scheme of this invention demonstrates superior cost performance under various price fluctuation scenarios. This result indicates that by synergistically optimizing speed, route, and loading / unloading rate, the cost pressure brought about by rising fuel prices can be mitigated to a certain extent. In addition, the original scheme selects the shortest distance between each pair of ports in terms of route selection, corresponding to a total voyage of 6449 nautical miles in this example, while the total voyage of the scheme of this invention is 6687 nautical miles, an increase of 238 nautical miles compared to the original scheme. This means that the ship will choose a route with a longer total distance but a shorter distance within the ECA, minimizing the voyage distance within the ECA area to reduce total costs.

[0054] Figure 9 This further demonstrates the impact of fuel price differences on speed decisions. When the price difference between low-sulfur fuel and regular fuel is small, the speed difference between ships inside and outside the ECA (Emission Control Area) is not significant. However, as the fuel price difference widens, ship speeds within the ECA decrease significantly to reduce the consumption of high-priced fuel, while speeds outside the ECA increase accordingly to compensate for the extended transit time, thereby balancing overall transportation costs. Therefore, under the trend of continuously rising fuel prices, using different speeds inside and outside the ECA can bring considerable economic benefits to shipping companies. The fundamental reason for this phenomenon is that changes in the price of low-sulfur fuel alter the structure of the unit-time operating cost for each segment. When the price of low-sulfur fuel is significantly higher than that of regular fuel, the unit-time cost within the ECA increases significantly. Therefore, shipping companies tend to reduce speed within the ECA to extend the period of using regular fuel, and increase speed outside the ECA to compensate for the transit time, thereby reducing overall transportation costs and meeting port service time windows.

[0055] Fluctuations in fuel prices not only affect a ship's operating costs and speed decisions, but also influence its route selection strategies. Figure 10 This study demonstrates the impact of low-sulfur fuel oil prices on ship fuel costs and sailing distances under different changing scenarios.

[0056] Depend on Figure 10 It can be seen that as the price of low-sulfur fuel oil gradually increases from a lower level, fuel costs exhibit a trend of "first rising, then falling, and then rising again," while the total sailing distance of ships also changes accordingly. This is because when the price difference between low-sulfur fuel oil and regular fuel oil is small, fuel costs within the emission control area have little impact on navigation plans. However, when the price of low-sulfur fuel oil is significantly higher than that of regular fuel oil, the choice of navigation plan changes. Shipping companies tend to choose navigation plans with shorter distances within the emission control area to reduce the use of low-sulfur fuel oil, while accelerating outside the emission control area to meet port time window requirements. But when the price of low-sulfur fuel oil continues to rise to even higher levels, the room for route adjustment is further compressed. At the same time, to ensure on-time arrival at port to meet port time window constraints, ships need to increase speed in areas outside the ECA to compensate for time losses, resulting in increased fuel consumption and a rise in fuel costs.

[0057] Analysis of the impact of arrival time window: The present invention sets the time window interval range from [24, 29] to [69, 72] and gradually increases it. Ten sets of calculation examples are constructed to analyze the impact of the time window interval change on various costs of ship operation. The results are shown in the table below.

[0058] Table of cost sensitivity analysis results under different time windows

[0059] As shown in the table above, with the increase in port time windows, the total weekly service cost of perishable goods liner services, the penalty cost for vessels arriving late, loading and unloading costs, and speed within the Emission Control Area (ECA) all show a downward trend, while the spoilage cost of perishable goods, inventory costs, and speed outside the ECA show an upward trend. The fundamental reason for this change is that the larger the time window range offered by the port, the less likely the vessel is to arrive late, allowing shipping companies more flexibility in choosing vessel speed and container loading / unloading schemes. For example, when port time restrictions are more lenient, vessels do not need to travel at high speeds within the ECA to catch the time window, thus choosing to reduce speed within the ECA to minimize the consumption of expensive low-sulfur fuel. Simultaneously, to compensate for transportation timeliness, vessels will accelerate outside the ECA, thereby optimizing overall transportation costs. Furthermore, due to more flexible loading and unloading arrangements, port throughput pressure is alleviated, and loading and unloading costs decrease accordingly. Excessive transportation time can also lead to a decline in cargo freshness, causing spoilage costs and inventory costs to increase.

[0060] Based on the simulation results above, it can be determined that the solution described in this invention has significant advantages in reducing total transportation costs, optimizing speed and port call sequence, reducing delay rates, and controlling fuel consumption. Furthermore, through sensitivity analysis of key factors such as port time window length and fuel price fluctuations, the model's adaptability to changes in fuel prices and port time windows is further verified, providing theoretical reference and strategic support for responding to policy adjustments and market changes in actual operations.

[0061] Implementing the embodiments of the present invention will have the following beneficial effects: This invention aims to optimize ship route selection, speed allocation, fuel strategy, and port operation planning by constructing a mixed-integer nonlinear programming model and designing an efficient solution algorithm. This method fully considers multiple factors, including fuel cost differences across different voyage segments, the relationship between cargo spoilage rate and time, and port time window constraints. It provides shipping companies with scientific decision support, enabling them to effectively reduce fuel consumption and cargo spoilage risks while strictly adhering to environmental regulations, thereby improving transportation efficiency and service reliability, and ultimately enhancing their market competitiveness. Furthermore, the implementation of this invention can significantly reduce the total operating costs of liner companies and improve the quality of perishable goods transportation services. In addition, this invention provides scientific decision support for shipping companies, promotes the development of green and intelligent shipping, and has significant practical application value and industry promotion significance.

[0062] In addition, a table of relevant symbols is attached to this application, as shown in the table below.

[0063]

[0064]

[0065] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for co-optimizing liner routes and speeds considering emission control areas, characterized in that, include: S1. Define the assumptions for the container liner route selection and speed optimization problem considering emission control areas and perishable goods. S2. Under the given assumptions, set the corresponding optimization objectives and create a liner route selection and speed optimization model that considers emission control area restrictions and perishable goods transportation. S3. Design an adaptive hybrid genetic algorithm to solve the liner route selection and speed optimization model to generate a corresponding liner route and speed co-optimization scheme.

2. The liner route and speed co-optimization method according to claim 1, characterized in that, The corresponding model formula for the liner route selection and speed optimization model that considers emission control area restrictions and perishable goods transportation is: （1） st （2） （3） （4） （5） （6） （7） （8） （9） （10） （11） （12） （13） （14） （15） （16） （17） （18） （19） （20） Formula (1) represents the objective function for minimizing the operating cost of a liner shipping route within a round trip. The first term represents the fixed cost of all vessels serving the route; the second term represents the cost of perishable goods spoilage during a round trip; the third term represents the inventory cost of container cargo during a round trip; the fourth term represents the penalty cost for late arrival at a port during a round trip; the fifth term represents the container loading and unloading cost during a round trip; and the sixth term represents the fuel cost during a round trip. Specifically, for the first item: the fixed costs of the vessel, which include the vessel's asset costs, employee costs, lubrication costs, and insurance premiums, if the weekly fixed cost of a vessel for one round trip is... The number of vessels allocated to the route to meet the weekly service frequency is The fixed cost of a ship for a round trip is as shown in formula (21): （21） Regarding the second item: the cost of spoilage and loss of perishable goods, container shipping companies transporting goods on their routes... Perishable goods, No. Perishable goods from the port of origin Transport to the port of destination for From the beginning of packing The spoilage loss function for perishable goods is: ; If the first The unit decay cost of perishable products is The total cost of spoilage is as shown in formula (22): （22） Regarding the third item: inventory cost, specifically container inventory cost, let's assume the unit inventory cost is... Flight segment The number of containers transported is The total inventory cost is as shown in formula (23): （23）； Regarding the fourth item: penalty cost, i.e., ship delay cost, assuming the ship leaves the port... The delay time is Ships in port The cost of late arrival penalties is The total delay cost for a ship's round trip is given by formula (24): （24） For item five: container loading and unloading costs, assuming the unit loading and unloading cost is... , 0-1 variables: representing ships in port Should a loading / unloading rate scheme be selected? The total loading and unloading cost for a round trip is as shown in formula (25): （25） Regarding item six: fuel cost, i.e., fuel consumption cost, the cost of fuel consumed by the ship during the voyage. The fuel consumption function is shown in formula (26): （26） in This is the fuel consumption coefficient. 0-1 variables: representing the ship's position on the voyage. Should a flight plan be selected? For navigation, assuming the fuel prices inside and outside the emission control area are respectively The fuel consumption cost of a ship during a round trip is given by formula (27): （27） The corresponding constraints are formulas (2) to (20); they are divided into three parts according to the type of constraint: time-related constraints, time-window related constraints, and decision variables and their range constraints. Formulas (2), (8)-(15) are constraints related to sailing time; Formula (2) represents the total sailing time of the ship on each segment; Formula (8) represents the loading and unloading time of the ship in port, which describes the stay time of the ship at each port of call; Formula (9) limits the earliest and latest allowed time for the ship to leave the port; Formula (10) is the constraint on the late arrival time of the ship at each port; Formulas (11) and (12) represent the time constraints for the ship to arrive at the port, where 168 is the number of hours in a week; Formulas (13) and (14) represent the waiting time constraints for the ship at the next port; Formula (15) represents the service frequency constraint of the ship, that is, the time of a round trip of a liner is an integer multiple of a week; Formulas (5), (6), and (7) belong to time window related constraints; Formula (5) represents the time window constraint for ships arriving at each port; Formula (6) represents the first... The total transport time for perishable goods from the port of origin to the port of destination includes sailing time and loading / unloading time at the port; Formula (7) represents the total transport time for perishable goods from the port of origin to the port of destination. Permissible sailing time constraints; Among them, formulas (3), (4), (16)-(20) are decision variables and their range constraints; formula (3) represents the ship's position on the voyage. The ship must choose a navigation plan; Formula (4) indicates that the ship is in port. Only one loading and unloading rate scheme can be selected for loading and unloading operations; Formula (16) represents the maximum number of ships on the route; Formulas (17) and (18) represent the speed limits of ships within / outside the emission control area; Formulas (19) and (20) represent 0-1 variable constraints; The parameters involved in the above formula have the following meanings: For port calls and assembly, and the route i, j ∈ ; Perishable The type set, and ∈ ; For the first The unit deterioration cost of perishable goods; For the first Perishable goods from the port to port The number of containers; Inventory cost per unit container of goods; For the ship in the voyage section The sailing time; The time taken for the ship to travel on segment N; For the ship in the voyage section The total number of containers transported; For ships in port The cost of late arrival penalties; For ships to leave port The delay time; This is the set of loading and unloading rates for the port. For the port The total volume of container loading and unloading; For ships in port Use loading and unloading rate scheme Loading and unloading costs at the time; It is a 0-1 variable, that is, when the ship is in port. Select loading and unloading rate scheme Its value is 1 when loading and unloading operations are being performed, and 0 otherwise. For the segment The set of paths; This refers to the fuel consumption of a ship at its design speed. The design speed of the ship; These are the prices of fuel used by ships within and outside the ECA area; For the flight segment Choose a navigation plan The speed of the vessel within the ECA during navigation; For the ship in the voyage section Choose a navigation plan Distance within the ECA region; For the ship in the voyage section Choose a navigation plan Distance outside the ECA region; For the flight segment Choose a navigation plan The speed of a vessel outside the ECA during navigation; It is a 0-1 variable, that is, when the ship is in the voyage segment Choose a navigation plan Its value is 1 during navigation, and 0 otherwise. Ports The start and end times of the time window; Perishable Total transportation time; , Perishable goods Port of origin and port of destination; For ships in port Loading and unloading time; For goods In the flight segment The maximum permitted sailing time; For ships in port The selected number Loading and unloading rates; For ships to leave port Time; For ships to arrive at the port Time; For ships in port Waiting time; This is the maximum number of ships that can be deployed. These are the ship's maximum and minimum speeds.

3. The liner route and speed co-optimization method according to claim 1, characterized in that, The solution steps corresponding to the adaptive hybrid genetic algorithm include: S1: Initialize the parameters of the genetic algorithm and generate an initial population. The method for generating the initial population includes constructing each individual using multi-layer real number encoding and setting the current generation. k =0, population size is N; S2: Determine the fitness function and calculate the fitness value of all individuals in the population. The fitness function is the reciprocal of the objective function, that is, the reciprocal of the total operating cost is used as the fitness evaluation basis. S3: Perform a selection operation on the population to select a portion of individuals as individuals for the next generation of the population. The selection operation adopts a roulette wheel strategy. S4: According to the set crossover probability Perform a crossover operation on the individuals selected by S3; S5: Based on the set mutation probability Perform mutation operations on the individuals after S4 crossover to generate new individuals; S6: Define a local search mechanism, and optimize the new individual based on the local search mechanism; S7: Replace the corresponding individuals in the current population with the individuals optimized by local search in S6, complete the population update, and recalculate the fitness values ​​of all individuals in the new generation population. S8: Determine if the preset termination condition is met: If the current iteration count has reached the preset maximum iteration count, or the objective function value of the best individual in the population for M consecutive generations has not exceeded the set change threshold, then terminate the iteration and output the current optimal solution as the final optimization scheme; otherwise, return to S3 and continue the iteration optimization for the next generation; where M is the upper limit of the number of generations in which the objective function value has not been improved.

4. The liner route and speed coordinated optimization method according to claim 3, characterized in that, A predefined initial population generation strategy is used to optimize the initial population generation method in S1 above. The initial population generation strategy includes the following steps: Step 1: Initialize parameters, set the corresponding population size, import the original data, and set the number of individuals n=0; Step 2: Use multi-level real number encoding to encode the selected scheme to construct the corresponding chromosome individual; Step 3: Determine whether the arrival time of the ship under the selected route and speed meets the preset constraints. If it does, proceed directly to Step 4; if it does not, adjust the relevant genes in the chromosome, i.e., perform gene repair treatment, and then proceed to Step 4. The preset constraints include, but are not limited to: (1) port time window restrictions; (2) perishable goods delivery time requirements. Step 4: After completing the construction of the current individual, verify whether all time nodes from the port of origin to the port of destination satisfy the preset constraints; if they are satisfied, add the individual to the initial population and set the individual count n=n+1; if they are not satisfied, return to Step 3 for secondary repair or discard the individual. Step 5: Iteration control: Set the individual counter n = n + 1; then, determine whether n has reached the preset population size; if not, return to Step 2 and start the next iteration to generate new individuals; if it has reached the preset population size, output the complete initial population and end the initial stage.

5. The liner route and speed co-optimization method according to claim 3, characterized in that, The fitness evaluation of all individuals in the population is based on the reciprocal of the objective function, with the specific formula as follows: (28) In the formula: For the fitness function, The objective function is the total cost. This represents the iteration number.

6. The liner route and speed coordinated optimization method according to claim 3, characterized in that, The specific steps for optimizing a new individual based on a local search mechanism are as follows: Step 1: In the current generation iteration, first perform the basic operations of the genetic algorithm on the previous generation population to generate new candidate population individuals; Step 2: Define three local search operations, specifically including: Operation 1: Increase or decrease the speed of a certain segment by 0.5 knots; Operation 2: Replace the current navigation plan of the segment, that is, select other alternative plans from the candidate paths; Operation 3: Change the loading and unloading rate plan of the current port. Step 3: Randomly select one of the three predefined operations in Step 2, and perform a local search operation (perturbation operation) on the corresponding gene location of the individual to generate a new individual, and re-evaluate the fitness value of the new individual. Step 4: Fitness Comparison and Solution Update: Calculate the total operating cost, i.e., the fitness value, of the new individual and compare it with the total operating cost of the original individual; if the total operating cost of the new individual is lower than that of the original individual, i.e., the fitness value of the new individual is better than the original solution, then replace the original individual in the current population with the new individual; otherwise, retain the original individual.

7. A collaborative optimization system designed based on the collaborative optimization method for liner routes and speeds according to any one of claims 1-6, characterized in that, include: The preprocessing unit is used to define the assumptions for the container liner route selection and speed optimization problem that takes into account emission control areas and perishable goods. The model creation unit is used to set the corresponding optimization objective under the assumed conditions, and to create a liner route selection and speed optimization model that takes into account emission control area restrictions and perishable goods transportation, corresponding to the optimization objective. The model solving unit is used to design an adaptive hybrid genetic algorithm to solve the liner route selection and speed optimization model. The scheme output unit is used to output the liner route and speed co-optimization scheme determined by the optimized liner route selection and speed optimization model.

Images