Method and device for optimizing ship transport capacity structure

Abstract

The present invention discloses a method and device for optimizing the ship transport capacity structure, belonging to the technical field of water transportation management and dispatching optimization. The method includes: obtaining a set of feasible configuration schemes for the ship transport capacity structure, clustering the set of feasible configuration schemes according to the Gaussian convolution clustering model to obtain a feasible solution set; constructing an optimization model for the transport capacity structure, and iteratively searching the feasible solution set in combination with the whale optimization algorithm to determine the optimal configuration scheme. By constructing an optimization model for the transport capacity structure to screen the set of feasible configuration schemes for the ship transport capacity structure, a configuration scheme that meets the requirements is obtained, ensuring the rationality of the transport capacity structure; by iteratively searching the configuration scheme through the whale optimization algorithm, with the maximum utilization rate of the lock chamber as the guide, and selecting the best through simulation, the optimal configuration scheme of the transport capacity structure is obtained, greatly reducing resource waste.

Description

Technical Field

[0001] The present invention relates to the technical field of water transportation management and dispatching optimization, and particularly to a method and device for optimizing the ship capacity structure. Background Art

[0002] With the rapid development of the global economy and the increasing frequency of trade activities, the status of inland waterway transportation in the cargo transportation system is becoming more and more important. As a low-cost and high-benefit transportation mode, inland waterway shipping is widely used in various bulk cargoes and long-distance transportation. With the increase in the ship flow on inland waterways, the problems of traffic congestion and resource waste are becoming increasingly prominent. The main manifestations are: (1) Unreasonable ship capacity structure: The distribution of ship types and sizes on inland waterways is unbalanced, and the capacity structure does not match the actual transportation demand. (2) Low utilization rate of passing facilities: The arrangement and dispatching of ships of different sizes and types in passing facilities are unreasonable, resulting in low space utilization rate of passing facilities.

[0003] Currently, most inland waterway shipping management systems rely on fixed dispatching rules and manual experience for the allocation and dispatching of the capacity structure, lacking a systematic optimization model and a scientific and reasonable management mechanism, and it is difficult to effectively cope with the complex and changeable shipping environment and the constantly changing transportation demand. Especially in the case of mixed passage of multiple types of ships, dispatching errors and resource waste are likely to occur.

[0004] Therefore, in the process of water transportation management and dispatching in the prior art, there is a problem of resource waste caused by unreasonable capacity structure. Summary of the Invention

[0005] In view of this, it is necessary to provide a method and device for optimizing the ship capacity structure to solve the problem of resource waste caused by unreasonable capacity structure in the process of water transportation management and dispatching in the prior art.

[0006] To solve the above problems, the present invention provides a method for optimizing the ship capacity structure, including:

[0007] Obtaining a set of feasible configuration schemes for the ship capacity structure;

[0008] Constructing a capacity structure optimization model, where the capacity structure optimization model includes an optimization objective, constraint conditions, and a multi-objective optimization function;

[0009] Performing multi-objective optimization on the feasible configuration schemes based on the capacity structure optimization model, and iteratively searching in combination with the whale optimization algorithm to determine the optimal configuration scheme;

[0010] Among them, the optimization objectives include the minimum capacity adjustment degree, the maximum lock chamber utilization rate, and the minimum number of lock passages.

[0011] In a possible implementation, multi-objective optimization is performed on the feasible configuration solutions according to the transport capacity structure optimization model, including:

[0012] Construct a multi-objective optimization function according to the optimization objectives and constraints;

[0013] Set the objective weights of the optimization objectives respectively, and adjust the multi-objective optimization function according to the objective weights to obtain the target transport capacity structure optimization model;

[0014] Screen out the feasible solution set of the feasible configuration solutions according to the target transport capacity structure optimization model.

[0015] In a possible implementation, the calculation formula of the multi-objective optimization function is:

[0016]

[0017] ）

[0018] Where is the objective function of the number of gate releases, is the maximum number of gate releases within the specified time, is the objective function of the progressive performance of transport capacity optimization, is the objective function of the utilization rate of the lock chamber, are the weights corresponding to the three different objective functions.

[0019] In a possible implementation, the optimal configuration solution is determined by iterative search in combination with the whale optimization algorithm, including:

[0020] Initialize the parameters of the feasible solution set, and set the population size and the maximum number of iterations;

[0021] Set the fitness function based on the maximum utilization rate of the lock chamber and the minimum number of gate releases;

[0022] Calculate the fitness value of each whale individual in the population, and iteratively update it to the maximum number of iterations based on the whale optimization algorithm, and determine the feasible solution corresponding to the maximum fitness value as the optimal configuration solution.

[0023] In a possible implementation, setting the objective weights of the optimization objectives includes:

[0024] Calculate the standard deviation of the first objective according to the feasible solution set;

[0025] Calculate the distance correlation between the first objective and the remaining objectives respectively;

[0026] Determine the objective weight of the first objective according to the standard deviation and the distance correlation.

[0027] In a possible implementation, obtaining a set of feasible configuration solutions for the ship capacity structure includes:

[0028] Obtaining the cargo transportation demand for each route according to historical data;

[0029] Generating multiple ship capacity structure configuration solutions that meet the cargo transportation demand to obtain a set of feasible configuration solutions;

[0030] Among them, the ship capacity structure configuration solution includes combinations of ships of different types, different tonnages, and different quantities.

[0031] In a possible implementation, before performing multi-objective optimization on the feasible configuration solutions according to the capacity structure optimization model, it also includes:

[0032] Clustering the set of feasible configuration solutions according to the Gaussian convolution clustering model to obtain a set of feasible solutions;

[0033] Performing multi-objective optimization on the set of feasible solutions according to the capacity structure optimization model.

[0034] In a possible implementation, clustering the set of feasible configuration solutions according to the Gaussian convolution clustering model to obtain a set of feasible solutions includes:

[0035] Calculating the density convolution estimate values of all data points of the feasible configuration solutions, and determining the data points with density convolution estimate values within the first preset range as candidate central points;

[0036] Taking the data point corresponding to the maximum density convolution estimate as the first central point, respectively calculating the conditional probability ratios between the remaining candidate central points and the first central point, and removing the candidate central points with conditional probability ratios greater than the preset conditional probability threshold to obtain updated central points;

[0037] Taking the data point corresponding to the second largest density convolution estimate of the updated central points as the second central point, respectively calculating the conditional probability ratios between the remaining updated central points and the second central point, and removing the candidate central points with conditional probability ratios greater than the preset conditional probability threshold to obtain secondarily updated central points;

[0038] Repeating the iteration until the density convolution estimate minimum value among the candidate central points is used as the central point to obtain the initial principal component central point, optimizing the parameters of the Gaussian convolution clustering model, and then determining the final principal component central point based on the initial principal component central point, and determining the feasible configuration solutions corresponding to the final principal component central point as the set of feasible solutions.

[0039] In a possible implementation, after determining the optimal configuration solution, it also includes:

[0040] Selecting the index measurement factors for ship congestion;

[0041] Obtain the factor set and evaluation set for the optimal configuration plan based on the index measurement factors;

[0042] Perform a composite operation on the factor set and the evaluation set through the interval average method to obtain the comprehensive index.

[0043] To solve the above problems, the present invention also provides a ship transport capacity structure optimization device, including:

[0044] A feasible configuration plan set acquisition module for obtaining the feasible configuration plan set of the ship transport capacity structure;

[0045] A transport capacity structure optimization model construction module for constructing a transport capacity structure optimization model, where the transport capacity structure optimization model includes an optimization objective, constraint conditions, and a multi-objective optimization function;

[0046] An optimal configuration plan determination module for performing multi-objective optimization on the feasible configuration plans based on the transport capacity structure optimization model, and iteratively searching in combination with the whale optimization algorithm to determine the optimal configuration plan;

[0047] Among them, the optimization objectives include the minimum transport capacity adjustment degree, the maximum chamber utilization rate, and the minimum number of lockage times.

[0048] The beneficial effects of adopting the above embodiments are as follows: The present invention provides a ship transport capacity structure optimization method. First, by constructing a transport capacity structure optimization model, the feasible configuration plan set of the ship transport capacity structure is screened to obtain the required configuration plans, ensuring the rationality of the transport capacity structure; then, through the whale optimization algorithm, iterative search is performed on the configuration plans, guided by the maximum chamber utilization rate, and the best is selected through simulation, realizing the acquisition of the optimal configuration plan for the transport capacity structure and greatly reducing resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic flowchart of an embodiment of the ship transport capacity structure optimization method provided by the present invention;

[0050] Figure 2 It is a schematic flowchart of an embodiment of obtaining the feasible solution set provided by the present invention;

[0051] Figure 3 It is a schematic flowchart of an embodiment of performing multi-objective optimization on the feasible configuration plans provided by the present invention;

[0052] Figure 4 It is a schematic flowchart of an embodiment of determining the first objective weight provided by the present invention;

[0053] Figure 5 It is a schematic flowchart of an embodiment of determining the optimal configuration plan provided by the present invention;

[0054] Figure 6Schematic flow chart of an embodiment of the whale optimization algorithm provided by the present invention;

[0055] Figure 7 Schematic flow chart of an embodiment of evaluating the congestion state of ships provided by the present invention;

[0056] Figure 8 Schematic flow chart of an embodiment of obtaining the optimal allocation of shipping capacity structure and evaluating it provided by the present invention;

[0057] Figure 9 Block diagram of an embodiment of the shipping capacity structure optimization device provided by the present invention;

[0058] Figure 10 Block diagram of an embodiment of the electronic device provided by the present invention. Detailed implementation manners

[0059] The preferred embodiments of the present invention will be specifically described below with reference to the accompanying drawings, in which the accompanying drawings form a part of the present invention and are used together with the embodiments of the present invention to explain the principles of the present invention, rather than to limit the scope of the present invention.

[0060] The Gaussian mixture clustering model, especially the Gaussian Mixture Model (GMM), is a clustering algorithm based on probability distribution. Specifically, GMM assumes that the data comes from multiple different Gaussian distributions, and each distribution represents a cluster. It can not only handle the uncertainty of the data, but also consider the probability that a data point belongs to different clusters, so it is called soft clustering. Compared with hard clustering algorithms such as K-means, GMM may be more suitable for dealing with data with irregular cluster shapes or correlated relationships between clusters.

[0061] The shipping capacity structure refers to the composition and configuration of different types and specifications of ships in a shipping company or fleet, which usually depends on multiple factors such as the operation strategy of the shipping company, market demand, ship technical characteristics, and operation costs.

[0062] The whale optimization algorithm can efficiently solve complex optimization problems by simulating the strategies of whales such as encirclement, spiral predation, and collision avoidance. This algorithm regards the position of each humpback whale as a feasible solution, and the process of searching for the solution of the problem is regarded as a process in which several whale individuals continuously update their individual positions until the target prey (i.e., the optimal solution) is tightly surrounded.

[0063] The shipping capacity structure determines the effect of water transportation management and scheduling. Due to the unreasonable shipping capacity structure, the utilization rate of passing facilities is low, and the overall efficiency of the inland waterway shipping system is severely restricted. Therefore, in the process of water transportation management and scheduling in the prior art, there is a problem of resource waste caused by the unreasonable shipping capacity structure.

[0064] To solve the above problems, the present invention provides a method and device for optimizing the ship capacity structure, which will be described in detail below.

[0065] Figure 1 As shown in the flowchart of an embodiment of the method for optimizing the ship capacity structure provided by the present invention, Figure 1 the method for optimizing the ship capacity structure includes:

[0066] S101: Obtain a set of feasible configuration schemes for the ship capacity structure;

[0067] S102: Construct an optimization model for the capacity structure, where the optimization model for the capacity structure includes an optimization objective, constraint conditions, and a multi-objective optimization function;

[0068] S103: Perform multi-objective optimization on the feasible configuration schemes based on the optimization model for the capacity structure, and iteratively search in combination with the whale optimization algorithm to determine the optimal configuration scheme;

[0069] Among them, the optimization objectives include the minimum capacity adjustment degree, the maximum chamber utilization rate, and the minimum number of lockage times.

[0070] In this embodiment, first, a set of feasible configuration schemes for the ship capacity structure is screened by constructing an optimization model for the capacity structure to obtain the configuration schemes that meet the requirements, ensuring the rationality of the capacity structure; then, the configuration schemes are iteratively searched by the whale optimization algorithm, guided by the maximum chamber utilization rate, and the best is selected through simulation, realizing the acquisition of the optimal configuration scheme for the capacity structure and greatly reducing resource waste.

[0071] As a preferred embodiment, in S101, to obtain a set of feasible configuration schemes for the ship capacity structure, first, the cargo transportation demand for each route is obtained according to historical data; then, multiple ship capacity structure configuration schemes that meet the cargo transportation demand are generated to obtain a set of feasible configuration schemes.

[0072] Among them, the ship capacity structure configuration scheme includes combinations of ships of different types, different tonnages, and different quantities.

[0073] It should be noted that the OD line (Origin-Destination Line) refers to the line connecting the origin and the destination, which is used to visually represent the connection between two geographical locations on the map. OD data is important information in traffic planning and can be used to analyze traffic demand, predict traffic flow, plan traffic facilities, etc.

[0074] In order to generate multiple ship capacity structure configuration plans that meet the requirements of cargo transportation, first, by collecting historical transportation data, transportation plans submitted by shipping companies, and actual transportation records, statistical analysis is carried out on the transportation demands of different types of goods between different origins and destinations (ODs), and the fluctuations of cargo transportation demands in different seasons and specific time periods are analyzed to form a cargo transportation demand model for each OD route in different periods. Then, according to the cargo transportation demands obtained from the statistical analysis, ship capacity structure configuration plans that meet the transportation demands are generated. These plans include combinations of different types (such as bulk carriers, dangerous goods carriers, container ships, etc.), different tonnages (such as small-tonnage ships, medium-tonnage ships, large-tonnage ships), and different numbers of ships.

[0075] That is to say, based on each OD route, first obtain the cargo transportation demand according to historical data, and then, based on the cargo transportation demand, automatically generate multiple ship capacity structure configuration plans that meet the requirements in combination with basic information such as the type, transportation capacity, and quantity of ships, which are recorded as the set of feasible configuration plans.

[0076] Furthermore, the set of feasible configuration plans only represents that it can meet the requirements, but there will definitely be a situation of a large amount of data redundancy. Therefore, in order to reduce the subsequent data comparison work, before performing multi-objective optimization on the feasible configuration plans according to the capacity structure optimization model, cluster the set of feasible configuration plans according to the Gaussian convolution clustering model to obtain the feasible solution set; then, perform multi-objective optimization on the feasible solution set according to the capacity structure optimization model.

[0077] In a specific embodiment, as Figure 2 shown, Figure 2 is a schematic flowchart of an embodiment for obtaining the feasible solution set provided by the present invention, including:

[0078] S201: Calculate the density convolution estimated values of all data points of the feasible configuration plan, and determine the data points with density convolution estimated values within a preset first range as candidate central points;

[0079] S202: Take the data point corresponding to the maximum density convolution estimate as the first central point, calculate the ratio of the conditional probabilities between the remaining candidate central points and the first central point respectively, and remove the candidate central points with the ratio of conditional probabilities greater than the preset conditional probability threshold to obtain updated central points;

[0080] S203: Take the data point corresponding to the second largest density convolution estimate of the updated central point as the second central point, calculate the ratio of the conditional probabilities between the remaining updated central points and the second central point respectively, and remove the candidate central points with the ratio of conditional probabilities greater than the preset conditional probability threshold to obtain the secondarily updated central points;

[0081] S204: Repeat the iteration until the minimum value of the density convolution among the candidate center points is used as the center point, obtaining the initial principal center point. After optimizing the parameters of the Gaussian convolution clustering model, determine the final principal center point based on the initial principal center point, and determine the feasible configuration solution corresponding to the final principal center point as the feasible solution set.

[0082] In this embodiment, the density of the configuration solution is estimated through the Gaussian convolution model, and similar solutions are clustered to screen out a representative set of transport capacity structure configurations as the feasible solution set, effectively reducing the number of candidate solutions and improving the optimization and solution efficiency.

[0083] It should be noted that the calculation formula of the Gaussian mixture model is as follows:

[0084] (1)

[0085] (2)

[0086] Among them, is the normal distribution density function, x is the data point, is the k th mean value of the Gaussian component, is the k th covariance of the Gaussian component, is the k th weight coefficient of the Gaussian component, ≥0, and satisfies , M is the number of Gaussian components, π is the periodic law of elements, exp represents taking the exponential function of the formula in the brackets, represents a d -dimensional data matrix, d represents the dimension of the data point, T represents taking the transpose of the vector .

[0087] The calculation formula of the Gaussian kernel function is as follows:

[0088] (3)

[0089] is the Gaussian kernel function, H is the covariance matrix of the Gaussian kernel function.

[0090] The calculation formula of the Gaussian mixture model based on Gaussian convolution is as follows:

[0091] (4)

[0092] Density convolutionf*H (x) The estimation calculation formula is as follows:

[0093] (5)

[0094] n is the number of data points, represents the i th data point.

[0095] As a preferred embodiment, in S201, in order to calculate the density convolution estimation values of all data points of the feasible configuration solutions, first, the estimation of the covariance matrix H of the Gaussian kernel function is performed. Specifically, first, the eigenvectors of each transport capacity structure configuration solution are extracted, including information such as ship type, tonnage distribution, quantity, etc. Each transport capacity structure configuration solution is represented as an eigenvector matrix , and the solution set has a dimension of d , and the number of solutions is n , represents the eigenvector of the i th solution.

[0096] Secondly, two eigenmatrices of the solution set are calculated, where is a data range matrix composed of the ranges of each dimension of the solution set , reflecting the size of the data distribution range, where d，= , diag means converting the vector into a diagonal matrix, is a diagonal matrix constructed with as the Gaussian kernel function and the normalized values of the densities of each sample point, represents the inverse matrix of the data range matrix, represents the estimation of the density value of the n th data point, and the calculation method is as follows:

[0097] (6)

[0098] (7)

[0099] (8)

[0100] Furthermore, in order to better perform clustering analysis and density estimation, Gaussian convolution data transformation (GCDT) is performed on the solution set . Specifically, each eigenvector matrix X is converted into a new data set Y, which is expressed as:

[0101] (9)

[0102] In the formula, , is the data range matrix of is the inverse matrix of which can be regarded as the compression ratio matrix between the data sets Y and X in the overall sense, while represents the compression of each data point according to its density value.

[0103] G is the Gaussian convolution data transformation of the solution set , that is , G the weighted covariance matrix of (WCM) is calculated as follows:

[0104] (10)

[0105] In the formula, is the covariance matrix of the data set , is G the data range matrix of

[0106] The calculation formula of the covariance matrix H of the Gaussian kernel function is as follows:

[0107] (11)

[0108] (12)

[0109] Among them, r is the correction coefficient, is the identity matrix of order d indicating the dimension of the data point.

[0110] According to the above formula, for any feasible configuration solution, its corresponding density convolution estimate value can be calculated. The density convolution estimate value represents the distribution of the feasible configuration solution to a certain extent. Then, calculate the density convolution estimate values of all data points { } of all feasible configuration solutions. Sort the convolution estimate values from large to small, and select the points within the first preset range as candidate central points (the excluded points are regarded as isolated points), denoted as , realizing the preliminary screening. The isolated points are the deleted and poor-quality feasible configuration solutions.

[0111] In a specific embodiment, the first preset range is selected as 80%. In other embodiments, the first preset range can also be adaptively adjusted according to actual needs, which will not be elaborated here.

[0112] As a preferred embodiment, in S202, the conditional probability calculation formula for the i-th data point generated by the k-th principal component center is as follows:

[0113] (13)

[0114] (14)

[0115] Where is the weight parameter for each as the principal component center, is the density convolution estimate value of s is the number of candidate center points, H is the covariance matrix of the Gaussian kernel function, is the i th data point generated by the k th principal component center.

[0116] Calculate the ratio of the conditional probabilities that the i th principal component center is generated by the k th principal component and by the i th principal component respectively:

[0117] (15)

[0118] Then, screen the initial principal component center points: Sort the candidate center points in descending order according to the convolution estimate value, and select the corresponding to the maximum value of the convolution estimate value as the first principal component center point, denoted as .

[0119] Calculate the ratio of the conditional probabilities of the remaining points and as . When , delete from the maximum point sequence, update the sequence, denoted as , and update to , where is the preset conditional probability threshold; when , retain , and do not update the sequence.

[0120] Select As the second principal component center point, denoted as , when ） , delete from the sequence of candidate centers, update the sequence, denoted as , and update to .

[0121] Repeat the iteration until the last point , and obtain M initial principal component center points, denoted as the sequence .

[0122] Set the initial preset covariance and initial preset weight of each initial center point of the Gaussian mixture model as follows:

[0123] (16)

[0124] (17)

[0125] are the data points in the sequence of M initial principal component center points.

[0126] ⑥ Use the EM algorithm to solve the initial estimated values of the mean, covariance, and weight coefficient of each initial center point, and calculate as follows:

[0127] (18)

[0128] (19)

[0129] (20)

[0130] (21)

[0131] Among them, represents the probability that the i th data point is generated by the k th principal component center, is the mean of the k th principal component center, and also represents the k th initial center point, Similarly, represents the initial pre-weight of the k th principal component center, represents the initial preset weight of the j th principal component center, represents the initial mean estimated value of the k th initial center point, nis the number of all data points, is the i th data point, represents the initial covariance estimate of the k th initial center point, represents the initial weight estimate of the k th initial center point.

[0132] 4) Optimize each parameter of the Gaussian mixture model to determine the final parameters and the principal component centers of the model

[0133] ① Calculate the conditional probability matrix of the initial center points using the initial estimated values of each parameter , where is the conditional probability that the th principal component center is generated by the th principal component. The calculation formula is as follows:

[0134] (22)

[0135] is the mean value calculated by formula (19), is the covariance calculated by formula (20), is the weight calculated by formula (21), M is the number of initial principal component center points selected, is the normal distribution density function.

[0136] ② Iteratively optimize and select the principal component center point sequence: Calculate the density convolution estimate value of each initial center point , and sort in descending order according to the convolution estimate value . Given the significance level , select as the first clustering center point of the model and represent it as . When , remove from the center point sequence. Use to represent the new center point sequence, update the conditional probability matrix to , select as the second clustering center point of the model and represent this center point as . When , remove from the center point sequence . Update the center point sequence and update the conditional probability matrix. Repeat the above process until the last initial center point to obtain principal component center points, and the new sequence is .

[0137] ③Set the initial covariance of each center point of the new sequence to be the corresponding original covariance. The initial weight coefficients of each center point of the new sequence are calculated as follows:

[0138] (23)

[0139] where is the data point in the new sequence of the principal component center points.

[0140] ④Use equations (18) - (21) to update the estimated values of the mean, covariance, and weight coefficients of each center point of the new center point sequence, and recalculate the conditional probability matrix using the new parameter estimated values .

[0141] ⑤Repeat steps ① - ④ until , is a threshold for stopping iteration. Obtain the final principal component center points and the parameter estimated values of each principal component center. Use the final center points and parameter estimated values to cluster the solutions, screen out the representative configuration solutions, and form the initial set of transport capacity structures.

[0142] In this embodiment, by performing clustering calculations on the set of feasible configuration solutions, the reliability of the feasible solution set is greatly improved, and the data redundancy is reduced.

[0143] As a preferred embodiment, in S102, in order to perform multi - objective optimization on the feasible configuration solutions according to the transport capacity structure optimization model, as Figure 3 shown Figure 3 is a schematic flowchart of an embodiment for performing multi - objective optimization on the feasible configuration solutions provided by the present invention, including:

[0144] S301: Construct a multi - objective optimization function according to the optimization objectives and constraint conditions;

[0145] S302: Respectively set the objective weights of the optimization objectives, and adjust the multi - objective optimization function according to the objective weights to obtain the target transport capacity structure optimization model;

[0146] S303: Screen out the feasible solution set of the feasible configuration solutions according to the target transport capacity structure optimization model.

[0147] In this embodiment, with the minimum degree of transport capacity adjustment, the maximum utilization rate of the lock chamber, and the minimum number of lock discharges as the optimization objectives, by taking the ship operation conditions as the constraint conditions and guiding by the three optimization objectives, a multi-objective optimization function is constructed; in addition, in order to meet the needs of different environments, the objective weights of the optimization objectives are respectively set, and the multi-objective optimization function is adjusted according to the objective weights to obtain the target transport capacity structure optimization model, realizing the combination of the three optimization objectives for subsequent optimization steps.

[0148] Specifically, the degree of transport capacity adjustment is represented by the difference between the optimized ship transport capacity structure and the current ship transport capacity structure, that is, the transport capacity gradualness. The calculation method of the gradualness objective function for the transport capacity structure optimization of any OD line is as follows:

[0149] (24)

[0150] In the formula, is the number of i type ships for transporting the j th type of goods; is the total number of ships for transporting the i th type of goods; is the number of i type ships for transporting the j th type of goods in the existing transport capacity structure; is the total number of ships for transporting the i th type of goods in the existing transport capacity structure, is the gradualness target value of the i type ships for transporting the j th type of goods.

[0151] The utilization rate of the lock chamber is quantified by the utilization rate of the lock chamber area. The utilization rate of the lock chamber area is the ratio of the projected area of the ships passing through the lock on the horizontal plane of the ship lock to the effective area of the ship lock. The greater the demand for ships passing through the lock and the higher the standardization degree of the ships, the greater the corresponding utilization rate of the lock chamber area, indicating that the economic benefits created by the ship lock are higher.

[0152] The calculation method of the utilization rate of the lock chamber area is as follows:

[0153] (25)

[0154] In the formula, is the area occupied by a certain ship in a certain ship lock, is the area of a certain ship lock, is the utilization rate of the lock chamber.

[0155] The number of lockage arrangements refers to the minimum number of lockages required for the ship capacity structure to pass through the lock under the premise of meeting the cargo transportation demand. The more the required number of lockage arrangements, the more likely it is to cause ship congestion when exceeding the maximum number of lockage arrangements of the lock. On the contrary, the fewer the required number of lockage arrangements, the greater the potential for improving ship capacity. Under a certain ship capacity structure, the number of lockage arrangements required within a certain period is related to information such as the number of ships, ship dimensions, and lock dimensions. The calculation method is as follows:

[0156] (26)

[0157] In the formula, is the number of lockage arrangements under a certain ship capacity structure; is the number of ships under this capacity structure; is the ship length under this capacity structure; is the ship width under this capacity structure; is the area of a certain lock.

[0158] Regarding the constraint conditions, the capacity structure optimization model needs to meet the following constraint limitations:

[0159] 1) The cargo transportation demand means that the number of ships required for each type of cargo on each OD line within a certain period is not less than the number of ships of this type to be configured during this period.

[0160] (27)

[0161] In the formula, is the total amount of cargo to be transported; is the number of ships of a certain type; is for transporting the i th type of cargo j the maximum deadweight tonnage of the th type of ship; i is for transporting the j th type of cargo

[0162] 2) The ship waiting time at the lock refers to the time delay caused by the need for ships to queue up to pass through the lock due to lock traffic congestion.

[0163] (28)

[0164] In the formula, is the ship waiting time at the lock, which does not exceed 72 hours.

[0165] 3) The lock means that the sum of the areas of all ships passing through the lock once should not exceed the lock chamber area of the lock.

[0166] (29)

[0167] (30)

[0168] (31)

[0169] Formula (29) indicates that the sum of the areas of the ships arranged in the lock cannot exceed the ship area, For ships i The area of is the lock area. In formula (30), For ships i Length; is the length of the ship lock. In formula (31), For ships i Width; The width of the lock.

[0170] 4) Ship loading means that the sum of the cargo tonnage of all cargo ships must be greater than the total weight of the cargo, taking into account the loading rate.

[0171] (32)

[0172] In the formula, Indicates i The load of the ship, Indicates the total weight of the cargo.

[0173] 5) The lock passing rules mean that ordinary ships and dangerous goods ships cannot pass through the locks at the same time, and dangerous goods ships must be arranged to pass through the locks separately.

[0174] (33)

[0175] Formula (33) represents the lock rule constraints for different ship types. Indicates that the ship is passing through the lock once. Indicates an ordinary ship. Indicates a dangerous goods vessel.

[0176] In S302, under the premise of satisfying the above constraints, the capacity adjustment degree and the number of gate discharges are minimized, the utilization rate of the gate chamber is maximized, and a multi-objective optimization function is constructed based on the above optimization objectives and various sub-objectives. The multi-objective optimization function is described as follows:

[0177]

[0178] In S303, in order to simplify the form of the multi-objective optimization function, the target weights of the optimization targets are set respectively, and the multi-objective optimization function is adjusted according to the target weights, and the expression of the target capacity structure optimization model is obtained as follows:

[0179] (34)

[0180] In the formula, is the objective function of the gate opening and closing times, is the maximum number of gate openings and closings within a specified time, is the objective function of the progressive performance of transport capacity optimization, is the objective function of the utilization rate of the lock chamber, are the weights corresponding to the three different objective functions.

[0181] In order to standardize the objective weights for setting optimization objectives, in S302, it is also necessary to adjust the objective weights. Taking the first objective as an example, as Figure 4 shown, Figure 4 is a schematic flowchart of a specific embodiment for determining the weight of the first objective provided by the present invention, including:

[0182] S401: Calculate the standard deviation of the first objective according to the feasible solution set;

[0183] S402: Calculate the distance correlation between the first objective and the remaining objectives respectively;

[0184] S403: Determine the weight of the first objective according to the standard deviation and the distance correlation.

[0185] In a specific embodiment, first, the data is standardized, and the calculation formula is as follows:

[0186] (35)

[0187] Wherein, is the original data, is the data after standardization, is the original value of the th configuration scheme for the i th objective, is the objective value after standardization of the th configuration scheme for the i th objective, is the minimum value of the th configuration scheme for the i th objective, is the maximum value of the th configuration scheme for the i th objective.

[0188] For each objective i, calculate its standard deviation:

[0189] (36)

[0190] Wherein, is the th configuration scheme for thei The value of a target is the i average value of the n th target,

[0191] For each pair of targets and , calculate the distance correlation between them :

[0192] ① Calculate the distance matrix: For each target, calculate the Euclidean distance between configuration schemes , where and are the values of the target X in the configuration schemes and correspondingly.

[0193] ② Perform double centering on the distance matrix DX , and the calculation formula is as follows:

[0194] (37)

[0195] where is the average value of the DX th row of the matrix , is the average value of the DX th column of the matrix , is the average value of all elements of the matrix DX .

[0196] ③ Calculate the distance covariance and the distance variance , and the calculation formula is as follows:

[0197] (38)

[0198] (39)

[0199] respectively represent the corresponding values after double centering the distance matrix X, Y of the target DX , n is the number of configuration schemes, is the distance covariance of the targets X and Y , is the distance variance of the target X .

[0200] ④The distance correlation between two targets X and Y is calculated as follows:

[0201] (40)

[0202] 4) Calculate the weight of each target , and the calculation formula is as follows:

[0203] (41)

[0204] where is the standard deviation of target , is the distance correlation between target and target , and the same applies to .

[0205] In this embodiment, quantifying the weight of the first target based on the distance correlation between each target can effectively ensure the objectivity of the weight of the first target.

[0206] After determining the optimization model of the transport capacity structure, since the optimization model of the transport capacity structure needs to perform multi-objective optimization, therefore, in order to determine the optimal configuration plan, it is necessary to combine the whale optimization algorithm to iteratively search for the feasible solution set, as Figure 5 shown Figure 5 is the flow chart of an embodiment for determining the optimal configuration plan provided by the present invention, including:

[0207] S501: Initialize the parameters of the feasible solution set, and set the population size and the maximum number of iterations;

[0208] S502: Set the fitness function based on the maximum chamber utilization rate and the minimum number of lock discharges;

[0209] S503: Calculate the fitness value of each whale individual in the population, and iteratively update it to the maximum number of iterations based on the whale optimization algorithm, and determine the optimization configuration plan corresponding to the maximum fitness value as the optimal configuration plan.

[0210] In this embodiment, after obtaining the feasible solution set, in order to select the most suitable optimal plan from it, it is also necessary to construct a sub-model of the ship lock passing capacity to overall optimize the transit time efficiency of the ships passing the dam after the transport capacity adjustment. First, determine the ship priority, and then establish an optimization model for arranging the lock chambers with the highest overall priority of the ships as the objective function, and use the optimization algorithm to make the model continuously converge to the maximum chamber utilization rate.

[0211] A suitable lock passing arrangement plan should comprehensively consider two aspects: the utilization rate of the lock chamber area and the ship priority. 1) Determine the objective function of the model

[0212] This model takes the overall highest priority of the ships passing through the lock as the objective function, and uses an optimization algorithm to make the model continuously converge towards the maximum utilization rate of the lock chamber. The constraint conditions that the objective function obeys are the lock constraints and the passing-through-lock rules constraints in the ship capacity structure optimization model for the previous ships. The comprehensive maximization function of ship priority and lock chamber utilization rate in the model is described as follows:

[0213] (42)

[0214] In the formula, represents the queue of ships waiting for the lock, represents whether the th ship is selected, represents the ship i 's priority.

[0215] Before arranging the lock chamber, it is necessary to determine the priority of a single ship first.

[0216] Finally, based on the whale optimization algorithm, calculate the optimal solution of the lock passing capacity sub-model:

[0217] ① Initialize parameters

[0218] When calculating the ship arrangement combination of the lock, it is necessary to initialize the parameters, set the population size N and the maximum number of iterations T.

[0219] ② Initialize the population

[0220] Randomly generate n initial solutions as the initial population. Each solution represents a lock chamber arrangement strategy. Each whale individual in the population represents a potential lock arrangement plan (feasible solution). Each solution contains information such as the position of each ship and the lock number it belongs to.

[0221] ③ Design the fitness function

[0222] Define the fitness function to evaluate the quality of the solution. The optimization objective of the model is to minimize the number of lockings and improve the average area utilization rate of the lock chamber. Based on these two objectives, design the fitness function. The fewer the number of lockings, the higher the fitness. The average area utilization rate takes a negative value, and the larger the value, the better the fitness. The specific expression is as follows:

[0223]

[0224] is the number of lockings for a certain ship capacity structure, is the area occupied by a certain ship in a certain lock chamber, is the area of a certain lock, is the maximum number of lockings within the specified time.

[0225] ④ Iteratively search for the optimal solution

[0226] After the initial solution is generated, calculate the fitness value of each whale individual in the initial population, find the individual with the best fitness value as the optimal position, and then use the whale's strategy of surrounding prey, hunting behavior strategy (bubble net attack strategy), and searching for prey strategy to update the position of the whale. The steps are as follows:

[0227] a. Strategy of surrounding prey

[0228] Whale individuals can identify the position of the prey and surround it in a circle. However, since the optimal position in the search space is unknown, the algorithm assumes that the optimal position of the current population is the target prey position, and other whales will try to approach and surround the target prey position. In this process, other whale individuals use Equation (43) to update their positions to approach the optimal individual:

[0229] (43)

[0230] (44)

[0231] a (45)

[0232] (46)

[0233] a (47)

[0234] In the formula, is the position of the whale individual that obtains the global optimal solution in the population at the th generation, is the position of the individual at the th iteration, is the position of the individual at the th iteration, represents the surrounding step size, and are coefficient vectors, is the maximum number of iterations, is a random number between, and a is a number that gradually decreases from 2 to 0 during the iteration process.

[0235] b. Hunting behavior strategy (bubble net attack strategy)

[0236] b1. Shrinking encirclement mechanism

[0237] During the iteration, the shrinking encirclement is achieved by reducing the value of from 2 to 0. Since the fluctuation range of is [- , when When it decreases, the fluctuation range will also shrink. Update the position of the whale individual using Equation (43).

[0238] b2. Spiral position update mechanism

[0239] To simulate the hunting method of whales advancing in a spiral, update the position of the whale individual using Equation (48). In the formula is the distance between the current whale and the optimal prey, is a constant controlling the logarithmic spiral shape, is a random number on [-1, 1], and cos represents the cosine function. The formula is as follows:

[0240] (48)

[0241] (49)

[0242] c. Prey search strategy

[0243] In the prey search stage, the whale individual no longer selects the optimal individual as the target for position update, but randomly selects an individual from the current population as the target for position update, thereby increasing the search range and maintaining population diversity while searching for the optimal solution. The calculation formula is as follows:

[0244] (50)

[0245] (51)

[0246] In the formula, is the position of a whale individual randomly selected from the current individuals, is the distance between the current whale and the random whale.

[0247] To simulate the behavior of whales swimming in a spiral, assume that there is a 50% probability of choosing the shrinking encircling mechanism or the spiral model to update the position of the whale during the optimization process, that is, there is a which is a random number between [0, 1]. When , use Equation (48) for position update. When , determine the update method according to the value of . When , use Equation (43) for position update. When | , use Equation (50) for position update. Perform boundary constraint processing on the updated individuals, recalculate the fitness values of the individuals, and then update the best individual according to the greedy selection mechanism , repeat the iterative optimization until the maximum number of iterations is reached, and output the optimal gate arrangement plan.

[0248] To clearly describe the data processing process of the whale optimization algorithm, as Figure 6 shown, Figure 6 is a schematic flowchart of an embodiment of the whale optimization algorithm provided by the present invention. WOA (Whale Optimization Algorithm) refers to the whale optimization algorithm.

[0249] After parameter initialization, by calculating the coefficient vector and continuously iteratively updating the whale position, the optimal result is obtained by combining simulation technology. In summary, in this embodiment, by iteratively processing the configuration scheme set one by one and maximizing the deviation of the fitness function according to the whale optimization algorithm, multi-objective optimization is achieved, effectively ensuring the reliability of the optimal configuration scheme.

[0250] Furthermore, after determining the optimal configuration scheme, in order to better monitor and control the water transportation situation, it is also necessary to evaluate the ship congestion status corresponding to the optimal configuration scheme, as Figure 7 shown, Figure 7 is a schematic flowchart of an embodiment of the evaluation of ship congestion status provided by the present invention, including:

[0251] S701: Select the index measurement factors for ship congestion;

[0252] S702: Obtain the factor set and evaluation set of the optimal configuration scheme based on the index measurement factors;

[0253] S703: Perform a composite operation on the factor set and the evaluation set through the interval average method to obtain a comprehensive index.

[0254] As a preferred embodiment, in S701, it is necessary to select the index measurement factors for ship congestion based on the time characteristics and spatial characteristics of ship congestion, starting from the perspectives of the duration, intensity, and scope of the congestion problem, and taking factors directly or indirectly affecting the ship congestion status such as the ship lock passing capacity, ship traffic flow, lockage scheduling, and anchorage as the basis.

[0255] As a preferred embodiment, in S702, four congestion evaluation indicators are selected as the evaluation set, namely the number of ships waiting for the lock, the average waiting time of ships for the lock, the longest waiting time, and the anchorage usage status.

[0256] Among them, the factor set is the set composed of each factor of the ship congestion index. Based on the selected congestion indicators, then the factor set {maximum number of congested ships, average waiting time, longest waiting time, anchorage utilization rate}.

[0257] As a preferred embodiment, in S703, the congestion evaluation algorithm adopts the fuzzy comprehensive evaluation method. According to the fuzzy comprehensive idea, the steps of congestion evaluation modeling are as follows: First, determine the factor set of the object to be evaluated according to the congestion indicators and the evaluation set . Then, determine the weight and membership degree vector of each factor respectively. After fuzzy transformation, a fuzzy evaluation matrix is obtained. Finally, perform fuzzy operation on the fuzzy judgment matrix and the weight vector of the factors to obtain the fuzzy comprehensive evaluation set

[0258] To better identify the congestion level, the traffic congestion state is divided into 5 levels, namely severe congestion, obvious congestion, slight congestion, relatively smooth, and smooth. Therefore, the determined evaluation set of index measurement factors is {smooth, relatively smooth, slight congestion, obvious congestion, severe congestion}. For the calculation of the evaluation value that comprehensively considers all congestion indicators, according to the principle of fuzzy transformation, the interval average method is adopted to perform composite operation on the evaluation result matrix and the interval average matrix, and the evaluation matrix is transformed into a comprehensive index , also known as the ship congestion index

[0259] Based on the fuzzy mathematics theory, assume the ship congestion index as a numerical value in the interval. When , it indicates that the waterway hub is in an extremely smooth state; when , it indicates that the ship congestion problem at the waterway hub is extremely serious The numerical values between reflect the different states of ship congestion at the waterway hub. According to the previously defined congestion levels, determine the index value range of the five-level ship congestion state, and the value range is 0~0.2, 0.2~0.4, 0.4~0.6, 0.6~0.8, 0.8~1, corresponding to the congestion levels of v1~v5 respectively

[0260] To clearly show the process of optimizing the configuration scheme of the ship transport capacity structure, as Figure 8 shown Figure 8 is the flow chart of an embodiment for obtaining the optimal configuration of the transport capacity structure and evaluation provided by the present invention. First, construct the transport capacity demand of the ship according to the operation demand of the ship and the existing ship operation data; then, perform preliminary optimization on the generated set of feasible solutions according to the Gaussian algorithm, screen out some infeasible solutions, and set constraints and optimization objective functions according to the transport capacity structure optimization model combined with the transport capacity demand. Then, perform simulation verification on the preliminarily screened solutions according to the whale optimization algorithm to obtain the optimal configuration scheme of the ship transport capacity structure; finally, in order to better evaluate the reliability of the optimal configuration scheme, a congestion model is specially set up to evaluate the congestion effect of the optimal configuration scheme, so as to better assist the water transportation management and dispatching work

[0261] In the above - mentioned manner, by constructing an optimized model of the transport capacity structure to screen the set of feasible configuration solutions for the ship's transport capacity structure, the required configuration solutions are obtained, ensuring the rationality of the transport capacity structure. Then, through the whale optimization algorithm, iterative search is carried out on the configuration solutions. Guided by the maximum utilization rate of the lock chamber, the best is selected through simulation, achieving the optimal configuration solution of the transport capacity structure and greatly reducing resource waste.

[0262] To solve the above problems, the present invention also provides a ship transport capacity structure optimization device, as Figure 9 shown Figure 9 is a structural block diagram of an embodiment of the ship transport capacity structure optimization device provided by the present invention. The ship transport capacity structure optimization device 900 includes:

[0263] A feasible solution set acquisition module 901, a feasible configuration solution set acquisition module, is used to obtain the set of feasible configuration solutions for the ship's transport capacity structure;

[0264] A transport capacity structure optimization model construction module 902, which is used to construct a transport capacity structure optimization model. The transport capacity structure optimization model includes an optimization objective, constraint conditions, and a multi - objective optimization function;

[0265] An optimal configuration solution determination module 903, which is used to perform multi - objective optimization on the feasible configuration solutions based on the transport capacity structure optimization model and perform iterative search in combination with the whale optimization algorithm to determine the optimal configuration solution;

[0266] Among them, the optimization objectives include the minimum transport capacity adjustment degree, the maximum lock chamber utilization rate, and the minimum number of lock discharges.

[0267] As Figure 10 shown, the present invention also correspondingly provides an electronic device 1000. The electronic device 1000 includes a processor 1001, a memory 1002, and a display 1003. Figure 10 Only some components of the electronic device 1000 are shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively.

[0268] The processor 1001 can be a central processing unit (CPU), a microprocessor, or other data - processing chips in some embodiments, and is used to run the program code stored in the memory 1002 or process data, such as the ship transport capacity structure optimization method in the present invention.

[0269] In some embodiments, the processor 1001 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 1001 may be local or remote. In some embodiments, the processor 1001 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0270] The memory 1002 may be an internal storage unit of the electronic device 1000 in some embodiments, such as the hard disk or memory of the electronic device 1000. The memory 1002 may also be an external storage device of the electronic device 1000 in other embodiments, such as a plug-in hard disk equipped on the electronic device 1000, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc.

[0271] Furthermore, the memory 1002 may include both the internal storage unit of the electronic device 1000 and the external storage device. The memory 1002 is used to store the application software installed on the electronic device 1000 and various types of data.

[0272] The display 1003 may be an LED display, a liquid crystal display, a touch liquid crystal display, and an OLED (Organic Light-Emitting Diode) toucher, etc. in some embodiments. The display 1003 is used to display the information of the electronic device 1000 and to display a visual user interface. The components 1001 - 1003 of the electronic device 1000 communicate with each other through a system bus.

[0273] In one embodiment, when the processor 1001 executes the ship transport capacity structure optimization program in the memory 1002, the following steps may be implemented:

[0274] Obtain magnetic resonance imaging data and construct a brain functional network based on the magnetic resonance imaging data;

[0275] Extract multi-scale topological features of the brain functional network;

[0276] Input the multi-scale topological features into a trained target ship transport capacity structure optimization model to determine the class label of the magnetic resonance imaging data.

[0277] It should be understood that when the processor 1001 executes the ship transport capacity structure optimization program in the memory 1002, in addition to the above functions, other functions may also be implemented. For specific details, reference may be made to the description of the corresponding method embodiments above.

[0278] Furthermore, the embodiments of the present invention do not specifically limit the type of the mentioned electronic device 1000. The electronic device 1000 can be a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop, or other portable electronic devices. Exemplary embodiments of the portable electronic device include, but are not limited to, portable electronic devices equipped with IOS, android, microsoft, or other operating systems. The above-mentioned portable electronic devices can also be other portable electronic devices, such as a laptop with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 1000 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0279] Correspondingly, the embodiments of the present invention also provide a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the programs or instructions are executed by a processor, the steps or functions in the ship transport capacity structure optimization method provided by the above-mentioned various method embodiments can be realized.

[0280] Those skilled in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program. The computer program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.

[0281] The ship transport capacity structure optimization method and device provided by the present invention have been introduced in detail above. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for optimizing ship capacity structure, characterized in that: include: Obtain a set of feasible configuration solutions for ship capacity structure; Constructing a transport capacity structure optimization model, wherein the transport capacity structure optimization model includes optimization objectives, constraint conditions, and a multi-objective optimization function; Based on the capacity structure optimization model, the feasible configuration scheme is optimized by multiple objectives, and combined with the whale optimization algorithm for iterative search to determine the optimal configuration scheme; The optimization objectives include minimizing the capacity adjustment, maximizing the lock chamber utilization, and minimizing the number of lock discharges; The feasible configuration scheme set is clustered according to the Gaussian convolution clustering model to obtain a feasible solution set, including: Calculate density convolution estimation values ​​of all data points of the feasible configuration scheme, and determine the data points whose density convolution estimation values ​​are within a first preset range as candidate center points; The data point corresponding to the maximum value of density convolution estimation is taken as the first center point, the conditional probability ratios between the remaining candidate center points and the first center point are calculated respectively, and the candidate center points whose conditional probability ratios are greater than a preset conditional probability threshold are removed to obtain an updated center point; The data point corresponding to the second maximum value is estimated as the second center point by density convolution of the updated center point, the conditional probability ratios between the remaining updated center points and the second center point are calculated respectively, and the candidate center points whose conditional probability ratios are greater than a preset conditional probability threshold are removed to obtain a secondary updated center point; Repeat the iteration until the minimum value of the density convolution estimation among the candidate center points is used as the center point to obtain the initial pivot center point, optimize the parameters of the Gaussian convolution clustering model, determine the final pivot center point based on the initial pivot center point, and determine the feasible configuration scheme corresponding to the final pivot center point as the feasible solution set; Among them, the density convolution estimate of all data points of the feasible configuration solution is calculated, including: First, the Gaussian kernel function covariance matrix H is estimated, and the eigenvector of each capacity structure configuration scheme is extracted. Each capacity structure configuration scheme is represented as a eigenvector matrix to obtain the scheme set ; Secondly, the calculation solution set Two feature matrices of It is a set of solutions A data range matrix consisting of the ranges in each dimension; Solution Set Perform Gaussian convolution data transformation and convert each feature vector matrix X into a new data set Y represented as: In the formula, , for The data range matrix is yes The inverse matrix of It is a dataset Y and X The compression ratio matrix between This means that each data point is compressed according to its density value. , Indicates n The density value of the data points is estimated, and diag means converting the vector into a diagonal matrix; G For the solution set Gaussian convolution data transformation, that is, , G The weighted covariance matrix of is calculated as follows: In the formula, For the dataset The covariance matrix of for G The data range matrix is , yes The inverse matrix of Gaussian kernel function covariance matrix H The calculation formula is as follows: in, r is the correction factor, for The unit matrix, d Represents the dimension of the data point; According to the above formula, for any feasible configuration scheme, the corresponding density convolution estimate is calculated.

2. The method for optimizing ship capacity structure according to claim 1, characterized in that: The multi-objective optimization of the feasible configuration scheme based on the capacity structure optimization model includes: Constructing a multi-objective optimization function according to the optimization objective and the constraint conditions; The target weights of the optimization targets are respectively set, and the multi-objective optimization function is adjusted according to the target weights to obtain a target capacity structure optimization model; A feasible solution set of the feasible configuration scheme is screened out according to the target capacity structure optimization model.

3. The method for optimizing ship capacity structure according to claim 2, characterized in that: The calculation formula of the multi-objective optimization function is: in, is the objective function of the number of gate discharges, is the maximum number of gate trips within the specified time. is the objective function for optimizing the asymptotic performance of capacity, is the objective function of the lock chamber utilization, These are the corresponding weights for the three different objective functions.

4. The method for optimizing ship capacity structure according to claim 2, characterized in that: The iterative search combined with the whale optimization algorithm to determine the optimal configuration solution includes: Initialize the parameters of the feasible solution set, set the population size and the maximum number of iterations; The fitness function is set based on the maximum utilization of the lock chamber and the minimum number of gate discharges; The fitness value of each whale individual in the population is calculated, and it is iteratively updated to the maximum number of iterations based on the whale optimization algorithm, and the feasible solution corresponding to the maximum fitness value is determined as the optimal configuration solution.

5. The method for optimizing ship capacity structure according to claim 2, characterized in that: The setting of the target weight of the optimization target includes: Calculate the standard deviation of the first objective according to the feasible solution set; Calculating the distance correlation between the first target and the remaining targets respectively; A target weight of the first target is determined based on the standard deviation and the distance correlation.

6. The method for optimizing ship capacity structure according to claim 1, characterized in that: The feasible configuration scheme set for obtaining the ship capacity structure includes: Obtain cargo transportation demand for each route based on historical data; Generating a plurality of ship capacity structure configuration schemes that meet the cargo transportation demand to obtain the feasible configuration scheme set; The ship capacity structure configuration scheme includes a combination of ships of different types, tonnages and numbers.

7. The method for optimizing ship capacity structure according to claim 1, characterized in that: After determining the optimal configuration solution, it also includes: Select the index calculation factor of ship congestion; Obtaining a factor set and an evaluation set of the optimal configuration solution based on the indicator measurement factors; The factor set and the evaluation set are compounded by the interval average method to obtain a comprehensive index.

8. A ship capacity structure optimization device, characterized in that: Used to execute the ship capacity structure optimization method according to any one of claims 1 to 7, comprising: A feasible configuration solution set acquisition module is used to acquire a feasible configuration solution set of a ship capacity structure; A capacity structure optimization model building module is used to build a capacity structure optimization model, wherein the capacity structure optimization model includes optimization objectives, constraints and multi-objective optimization functions; An optimal configuration scheme determination module is used to perform multi-objective optimization on the feasible configuration scheme based on the capacity structure optimization model, and determine the optimal configuration scheme by iterative search in combination with the whale optimization algorithm; Among them, the optimization objectives include minimizing the capacity adjustment, maximizing the lock chamber utilization rate and minimizing the number of lock discharges.

Images