Container ship stowage optimization method based on improved genetic algorithm

By improving the genetic algorithm and multi-objective optimization model, the problems of multi-objective and multi-constraints in container ship loading are solved, the operating efficiency of the dock and the seaworthiness of the ship are improved, and a more efficient loading and unloading process is achieved.

CN118916781APending Publication Date: 2024-11-08HEBEI UNIVERSITY
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410961777.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing research on loading of container ships is mainly single-target optimization, which fails to effectively solve the combination optimization problems of multiple goals and multiple constraints, resulting in insufficient loading and unloading efficiency and airworthiness.

Method used

Using improved genetic algorithms, a multi-objective multi-constraint mathematical model with the minimum number of container inversions and ship tilt torque as the objective function, and a Pareto optimal solution and elite selection strategy are added to the genetic algorithm to perform multi-objective optimization to improve Operational efficiency of the dock.

Benefits of technology

By optimizing the loading plan, the operating efficiency of the dock is significantly improved, the number of container reversals is reduced, the tilt torque of the ship is controlled, and the loading and unloading speed and seaworthiness are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916781A_ABST
    Figure CN118916781A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of automatic manufacturing system control, and particularly relates to a container ship stowage optimization method based on an improved genetic algorithm, and the method comprises the steps: building a multi-target multi-constraint mathematical model with the minimum container falling amount and the transverse tilting moment of a ship as targets on the basis of meeting the ship seaworthiness; the method comprises the following steps: firstly, creating a population comprising a plurality of individuals in a population initialization stage, wherein each individual comprises layout, port and weight information; secondly, new individuals are formed through selection, intersection and variation to replace individuals in the original population, and a new population is formed; and finally, finding an optimal solution by calculating an objective function value of each population. Container ship stowage optimization is carried out based on the target function model considered by the wharf side and the improved genetic algorithm, and the obtained stowage result can effectively improve the operation efficiency of the wharf.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of automatic manufacturing system control, and in particular relates to a container ship loading optimization method based on an improved genetic algorithm. Background Art

[0002] Ship stowage planning is an NP-Hard problem. Existing literature mainly consists of the study of pre-stowage planning of ships and actual stowage planning of terminals. In the actual operation process, stowage of container ships is a multi-objective and multi-constrained combinatorial optimization problem. The quality of stowage planning is related to the loading and unloading efficiency of container ships at ports, the seaworthiness of ships, etc.

[0003] For the research on container loading, the current research mainly focuses on single-objective optimization, while there are few literatures on multi-objective optimization. Based on the shortcomings of the above research, the present invention studies the multi-objective loading optimization problem of container terminals from the perspective of overall balance of container ship loading. Summary of the invention

[0004] The purpose of the present invention is to provide a container ship loading optimization method based on an improved genetic algorithm, which performs container ship loading optimization based on an objective function model considered by the terminal and an improved genetic algorithm, and the obtained loading results can effectively improve the operating efficiency of the terminal.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A container ship loading optimization method based on an improved genetic algorithm comprises the following steps:

[0007] Step 1: Model establishment: Establish a multi-objective and multi-constraint mathematical model with the minimum number of container overturning and the ship's heeling moment as the objective function;

[0008] Step 2: Use the improved genetic algorithm to solve the mathematical model: Add the Pareto optimal solution to the genetic algorithm for multi-objective optimization, and use the elite selection strategy in the genetic algorithm selection function.

[0009] Furthermore, the mathematical model includes an objective function and constraints;

[0010] The objective function is:

[0011]

[0012] The constraints are:

[0013]

[0014] w(r,t)-w(r,t+1)≥0 (5)

[0015] a(r,t)≤1 (6)

[0016] a(r, t)-a(r, t+1)≥0 (7)

[0017]

[0018] Z(r, t, l) = 0 or 1 (9)

[0019] a(r, t) = 0 or 1 (10)

[0020] Where R is the maximum number of rows in the bay; r is the row number (r∈R); T is the highest layer in the bay; t is the layer number (t∈T); N is the number of containers; w kc is the maximum weight that the ship can bear; c(r, t) represents the container slot at the rth column and tth layer in the bay; w(r, t) represents the weight type of the container placed in the container slot c(r, t), and its values ​​are 3, 2, and 1, representing heavy container, medium container, and light container, respectively; w r represents the total weight of the container in the rth row (r = 1, 2, ..., R); L B Indicates the width of a 20ft container;

[0021] Formula (1) indicates that the number of containers overturned is the minimum; Formula (2) indicates the heeling moment of the ship; Formula (3) indicates the sum of the weight of each row of containers in the bay; Formula (4) indicates that all containers are guaranteed to be loaded on the ship; Formula (5) indicates the constraint that the heavy does not crush the light, indicating that the container on the current layer must be heavier than the container on the upper layer; Formula (6) indicates that each slot is occupied at most once; Formula (7) indicates the constraint that the slot is not suspended, that is, if there is no container on the slot c(r, t), then containers are not allowed to be placed on the upper layer; Formula (8) indicates that the weight of the container must be within the range that the ship can bear; Formulas (9) and (10) indicate that the decision variables take values ​​0 or 1.

[0022] Furthermore, the step 2 includes the following steps:

[0023] S 1. Population initialization: According to the number of containers and the ship's bay parameters, the initial population is randomly generated. In the population initialization stage, a population containing multiple individuals is created, each of which contains the layout, port and weight matrix;

[0024] S2: Calculate the objective function values ​​of all individuals in the population and retain the optimal solution;

[0025] S3: Retain more excellent individuals through elite selection strategy;

[0026] S4: Use the single-point crossover method to cross the selected parent generation to generate new offspring individuals;

[0027] S5: Perform bit flip mutation operation on offspring individuals;

[0028] S6: Update the Pareto optimal solution set: Update the Pareto optimal solution set according to the fitness evaluation results;

[0029] S7: If the output result meets the termination condition of the algorithm; if not, return to S3 and continue iterating until the result of the iteration meets the termination condition.

[0030] Furthermore, S3 adopts the elite selection strategy method. When the population evolves, the individual with the largest fitness value in the nth generation population is The individual with the largest fitness value in the n+1 generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individual with the smallest fitness value in the n+1 generation population;

[0031] When the population evolves, the individual with the second largest fitness value in the nth generation population is When The fitness value is greater than

[0032] The fitness value of Copy to the next generation, and eliminate the individual with the second smallest fitness value in the n+1 generation population;

[0033] By analogy, when the population evolves, the individual with the largest fitness value in the nth generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individuals with the smallest fitness value m in the n+1 generation population, until there are no individuals with fitness values ​​greater than m in the n generation population. Individuals with the best fitness are selected from the population as parents.

[0034] The technical effects achieved by the present invention are:

[0035] The container ship loading optimization method based on the improved genetic algorithm of the present invention performs container ship loading optimization based on the objective function model considered by the terminal and the improved genetic algorithm, and the obtained loading result can effectively improve the operation efficiency of the terminal. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of the improved genetic algorithm of the present invention;

[0037] Figure 2 is a schematic diagram of a loading scheme in Embodiment 2 of the present invention;

[0038] Figure 3 is the number of times the container is turned over in Embodiment 2 of the present invention;

[0039] Figure 4 is the heeling moment in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0041] Embodiment 1:

[0042] like Figure 1 As shown in the figure, a container ship loading optimization method based on an improved genetic algorithm is established. On the basis of satisfying the seaworthiness of the ship, a multi-objective and multi-constraint mathematical model with the minimum amount of container dumping and the heeling moment of the ship as the objectives is established; and the improved genetic algorithm is used to solve it: first, in the population initialization stage, a population containing multiple individuals is created, and each individual contains layout, port and weight information; secondly, new individuals are formed through selection, crossover and mutation to replace the individuals in the original population to form a new population; finally, the optimal solution is found by calculating the objective function value of each population.

[0043] Specifically, a container ship loading optimization method based on an improved genetic algorithm comprises the following steps:

[0044] Step 1: Model establishment: Based on the relevant information of the container ship, a multi-objective and multi-constrained mathematical model is established with the minimum number of container overturning and the ship's heeling moment as the objective function, and the constraints are that the heavy does not press the light, the container space is not placed in the air, and the total weight of the container is within the allowable range of the ship;

[0045] When loading container ships, the terminal mainly considers to ensure that the container ships can leave the port on time according to the berthing plan, so the loading and unloading speed and efficiency are the key factors. This technology, taking into account the research significance and practical significance, decided to establish a mathematical model that only unloads and does not load, and takes the minimum amount of container turnover and heeling moment as the goal, and solves it with an improved genetic algorithm.

[0046] When constructing the mathematical model, the following assumptions are made:

[0047] (1) Before the actual loading at the terminal, the pre-loading diagram must be provided by the ship to the terminal to ensure that all information about the containers to be loaded, such as size, is known; and this technical solution is based on the perspective of the terminal, so the main consideration is the optimization of the terminal's loading and loading plan.

[0048] (2) This technical solution considers the loading operation of a single ship, and in order to avoid the increase of equipment conflicts and the increase of the number of container turnovers when loading and unloading are carried out simultaneously, loading and unloading at the same time are not considered; considering the convenience of modeling and the implementation conditions of problem solving, the following reasonable assumptions are made:

[0049] (a) Assume that the ship starts loading at the initial port. At this time, all the loading bays are empty. The ship only unloads but does not load at the ports along the way. At the final port, all the containers in the ship's bays are unloaded.

[0050] (b) Assume that the container size is 20ft, the cargo types are the same, and no special containers are considered;

[0051] (c) The impact of hatch covers on container ship's shell tipping is not considered;

[0052] (d) Assume that the number of containers to be loaded at the bay does not exceed the bay’s maximum capacity.

[0053] The parameters are defined as follows:

[0054] R represents the maximum number of columns of bay positions;

[0055] r is the column number (r∈R);

[0056] T represents the highest layer number of the bay position;

[0057] t is the layer number (t∈T);

[0058] N is the number of containers;

[0059] w kc is the maximum bearing weight of the ship;

[0060] c(r, t) represents the box position of the rth column and the tth layer in the bay;

[0061] P is the total number of ports of call along the way, d(r, t) represents the destination port number of the container placed in slot c(r, t) (d(r, t)∈P); w(r, t) represents the weight type of the container placed in slot c(r, t), and the values ​​are 3, 2, and 1, representing heavy container, medium container, and light container respectively; w r represents the total weight of the container in the rth row (r = 1, 2, ..., R); L B Indicates the width of a 20ft container.

[0062] In order to accurately describe the model, decision variables Z(r, t, l) and a(r, t) are introduced, and their values ​​are 0 or 1.

[0063]

[0064] Wherein, the mathematical model includes an objective function and constraint conditions;

[0065] The objective function is:

[0066]

[0067] Wherein, formula (1) indicates the minimum number of container overturning; formula (2) indicates the heeling moment of the ship;

[0068] The constraints are:

[0069]

[0070] w(r,t)-w(r,t+1)≥0 (5)

[0071] a(r,t)≤1 (6)

[0072] a(r, t)-a(r, t+1)≥0 (7)

[0073]

[0074] Formula (3) represents the sum of the weights of containers in each row of the bay; Formula (4) ensures that all containers are loaded on the ship; Formula (5) represents the constraint that heavy containers do not crush light containers, which means that the containers on the current layer must be heavier than the containers on the upper layer; Formula (6) represents that each slot is occupied at most once; Formula (7) represents the constraint that the slot is not vacant, that is, if there is no container on the slot c(r, t), then containers are not allowed to be placed on the upper layer; Formula (8) represents that the weight of the container must be within the range that the ship can bear; Formulas (9) and (10) represent that the decision variables take values ​​of 0 or 1.

[0075] Step 2: Use the improved genetic algorithm to solve the mathematical model: Add the Pareto optimal solution to the genetic algorithm for multi-objective optimization, and use the elite selection strategy in the selection function of the genetic algorithm, which can retain more excellent individuals, help accelerate the optimization process and improve the convergence of the algorithm.

[0076] Traditional genetic algorithms have the following advantages:

[0077] (1) Strong parallel search capability: Genetic algorithms can search multiple solution spaces simultaneously, so they have certain advantages in parallel search;

[0078] (2) Strong global search capability: Genetic algorithms are suitable for solving complex multimodal and multimodal problems and can globally search the solution space of the problem;

[0079] (3) Strong adaptability: Genetic algorithms can adaptively adjust the search space to adapt to the characteristics of different problems.

[0080] But there are some disadvantages:

[0081] (1) Slow convergence: When solving problems, genetic algorithms usually take a long time to converge to the optimal solution;

[0082] (2) Difficulty in parameter selection: The parameter selection in genetic algorithms has a great impact on the performance of the algorithm, but it is usually difficult to select appropriate parameters;

[0083] (3) Easy to fall into local optimality: Genetic algorithms have a certain probability of falling into local optimality during the search process, and it is difficult to escape from the local optimal solution.

[0084] Therefore, improvements are needed. The concept of Pareto optimal solution set is added to the genetic algorithm to enable multi-objective optimization. In addition, the selection function in the genetic algorithm uses an elite selection strategy, which can retain more excellent individuals, help accelerate the optimization process and improve the convergence of the algorithm.

[0085] Specifically, Figure 1 As shown, the step 2 includes the following steps:

[0086] S1. Population initialization: According to the number of containers and the ship's bay parameters, the initial population is randomly generated. In the population initialization stage, a population containing multiple individuals is created, each of which contains a layout, port and weight matrix;

[0087] S2: Calculate the objective function values ​​of all individuals in the population and retain the optimal solution;

[0088] S3: Retain more excellent individuals through elite selection strategy;

[0089] Specifically, S3 adopts the elite selection strategy method. When the individual with the largest fitness value in the nth generation population during the population evolution process is The individual with the largest fitness value in the n+1 generation population is When The fitness value is greater than The fitness value of The next generation is copied and the individual with the smallest fitness value in the n+1 generation population is eliminated. When the population evolves, the individual with the second largest fitness value in the nth generation population is When The fitness value is greater than The fitness value of The next generation is copied to the next generation, and the second smallest individual in the n+1 generation population is eliminated. Similarly, when the population evolves, the individual with the mth largest fitness value in the nth generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individuals with the smallest fitness value m in the n+1 generation population, until there are no individuals with fitness values ​​greater than m in the n generation population. Individuals with the best fitness are selected from the population as parents.

[0090] S4: Use the single-point crossover method to cross the selected parent generation to generate new offspring individuals;

[0091] S5: Perform bit flip mutation operation on offspring individuals;

[0092] S6: Update the Pareto optimal solution set: Update the Pareto optimal solution set according to the fitness evaluation results;

[0093] S7: If the output result meets the termination condition of the algorithm; if not, return to S3 and continue iterating until the result of the iteration meets the termination condition.

[0094] Embodiment 2:

[0095] In order to verify the superiority of the algorithm model proposed in this paper, this paper is based on the loading problem of the full container ship "Ginghe". The "Ginghe" is a full container ship owned by Shanghai Ocean Shipping Company and put into use at the end of the last century. The maximum container capacity of the Glacier is 1696 TEUs, and one of the 94 pre-allocated ship container slots with 35 bays is used for the experiment. For the convenience of statistics and simulation calculations, it is assumed that the size of all containers waiting to be loaded in all yards is 20ft.

[0096] To verify the superiority of the present invention, MATLAB is used for simulation verification. The output images of the objective function are respectively Figure 3 and Figure 4 ,from Figure 3 It can be seen from the results that the results output by the improved genetic algorithm are obviously better than those before the improvement; Figure 4 It can be seen that the heeling moments output by the improved genetic algorithm and the traditional genetic algorithm are both within a certain range, and the output results of the improved genetic algorithm are better. Therefore, the improved genetic algorithm has certain advantages over the traditional genetic algorithm.

[0097] In summary, this technical solution optimizes container ship loading based on the objective function model considered by the terminal and the improved genetic algorithm, and the obtained loading results can effectively improve the operating efficiency of the terminal.

[0098] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A container ship loading optimization method based on an improved genetic algorithm, characterized in that: The following steps are involved: Step 1: Model establishment: Establish a multi-objective and multi-constraint mathematical model with the minimum number of container overturning and the ship's heeling moment as the objective function; Step 2: Use the improved genetic algorithm to solve the mathematical model: Add the Pareto optimal solution to the genetic algorithm for multi-objective optimization, and use the elite selection strategy in the genetic algorithm selection function.

2. The container ship loading optimization method based on improved genetic algorithm according to claim 1 is characterized by: The mathematical model includes an objective function and constraint conditions.

3. The container ship loading optimization method based on improved genetic algorithm according to claim 2 is characterized by: The objective function is: Where R represents the maximum number of columns in the bay; r is the column number (r∈R); T represents the highest layer in the bay; t is the layer number (t∈T); w r represents the total weight of the container in the rth row (r = 1, 2, ..., R); L B Indicates the width of a 20ft container; Formula (1) indicates that the number of container overturning is the minimum; Formula (2) indicates the heeling moment of the ship.

4. The container ship loading optimization method based on improved genetic algorithm according to claim 3 is characterized by: The constraints are: w(r,t)-w(r,t+1)≥0 (5) a(r,t)≤1 (6) a(r, t)-a(r, t+1)≥0 (7) Z(r, t, l) = 0 or 1 (9) a(r, t) = 0 or 1 (10) Where N is the number of containers; w kc is the maximum weight that the ship can bear; c(r, t) represents the container slot at the rth column and the tth layer in the bay; w(r, t) represents the weight type of the container placed in the container slot c(r, t), and its values ​​are 3, 2, and 1, representing heavy container, medium container, and light container, respectively; Formula (3) represents the sum of the weights of containers in each row of the bay; Formula (4) ensures that all containers are loaded on the ship; Formula (5) represents the constraint that heavy containers do not crush light containers, which means that the containers on the current layer must be heavier than the containers on the upper layer; Formula (6) represents that each slot is occupied at most once; Formula (7) represents the constraint that the slot is not vacant, that is, if there is no container on the slot c(r, t), then containers are not allowed to be placed on the upper layer; Formula (8) represents that the weight of the container must be within the range that the ship can bear; Formulas (9) and (10) represent that the decision variables take values ​​of 0 or 1.

5. The container ship loading optimization method based on improved genetic algorithm according to claim 1 is characterized by: The step 2 comprises the following steps: S1. Population initialization: According to the number of containers and the ship's bay parameters, the initial population is randomly generated. In the population initialization stage, a population containing multiple individuals is created, each of which contains a layout, port and weight matrix; S2: Calculate the objective function values ​​of all individuals in the population and retain the optimal solution; S3: Retain more excellent individuals through elite selection strategy; S4: Crossover the selected parent generation to generate new offspring individuals; S5: Perform bit flip mutation operation on offspring individuals; S6: Update the Pareto optimal solution set: Update the Pareto optimal solution set according to the fitness evaluation results; S7: If the output result meets the termination condition of the algorithm; if not, return to S3 and continue iterating until the result of the iteration meets the termination condition.

6. The container ship loading optimization method based on improved genetic algorithm according to claim 5 is characterized by: S3 adopts the elite selection strategy method. When the population evolves, the individual with the largest fitness value in the nth generation population is The individual with the largest fitness value in the n+1 generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individual with the smallest fitness value in the n+1 generation population; When the population evolves, the individual with the second largest fitness value in the nth generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individual with the second smallest fitness value in the n+1 generation population; By analogy, when the population evolves, the individual with the largest fitness value in the nth generation population is When The fitness value is greater than The fitness value of Copy to the next generation, and eliminate the individuals with the smallest fitness value m in the n+1 generation population, until there are no individuals with fitness values ​​greater than m in the n generation population. Individuals with the best fitness are selected from the population as parents.

7. The container ship loading optimization method based on improved genetic algorithm according to claim 5 is characterized by: In S4, a single-point crossover method is used to crossover the selected parent generation.

8. The container ship loading optimization method based on improved genetic algorithm according to claim 1 is characterized by: The mathematical model in step 1 is based on the constraints that the heavy does not press the light, the containers are not placed in the air, and the total weight of the containers is within the range allowed by the ship.

Citation Information

Cited By

  • Structural optimization method for modifying container into mobile laboratory based on finite element analysis

    CN119397659A