Scheduling method and system for ship segmented lightering vehicles

Abstract

The invention provides a scheduling method and system for ship segmented lightering vehicles. The method comprises the following steps: establishing a basic database; adjusting algorithm parameters of the genetic algorithm, and configuring each algorithm parameter and a corresponding calculation result into a training database; training a generative adversarial network model based on the training database; acquiring a real-time lightering task, and generating real-time algorithm parameters through the generative adversarial network model; and obtaining an optimal lightering task sequence corresponding to the real-time lightering task based on the real-time algorithm parameters and a genetic algorithm, and outputting the optimal lightering task sequence. By adjusting genetic algorithm parameters, the problem that algorithm parameters are difficult to select can be solved. And when the individual fitness is calculated, the distance is converted into time for calculation, so that the calculation accuracy is improved. By adding the priority encouragement value for calculation, the requirement that part of tasks need to be scheduled preferentially can be met. And advantages and disadvantages of different schemes can be quantitatively judged to adapt to changes of actual conditions, and intelligent and efficient management and control of ship segmented lightering vehicles are realized.

Description

Technical Field

[0001] The present application relates to the technical field of shipbuilding, and in particular to a method and system for dispatching vehicles for ship-stage transfer. Background Art

[0002] Ship block construction is an efficient and high-quality production method commonly used in the modern shipbuilding industry. In this process, a ship is designed into multiple pre-planned independent structural units, which are called "ship blocks". Each block is precisely manufactured in the shipyard according to the drawings, including structural parts, outfitting parts and some pre-outfitting work content of the corresponding parts of the hull. Ship block construction technology plays a vital role in improving the overall competitiveness of the shipbuilding industry, reducing production costs, and ensuring construction progress and product quality.

[0003] As modern ship design develops towards larger and more complex structures, block construction has become the mainstream. Therefore, the use of efficient block transfer technology is an indispensable part of supporting the block construction process of ships.

[0004] At present, the scheduling method for ship-stage transfer vehicles relies on algorithms to automatically judge the quality of the scheduling plan. Therefore, the calculation effect of the algorithm directly affects the quality of the final scheduling plan. The existing algorithm has the following problems:

[0005] Single algorithms such as existing genetic algorithms are prone to falling into local optimality: when using a hybrid algorithm of genetic algorithm and taboo search for solution optimization, the solution result of the genetic algorithm is used as the input of the taboo search algorithm, but the taboo search algorithm cannot affect the genetic algorithm, and cannot jump out of the local optimality during the genetic algorithm process. The degree of optimization of the solution result is limited compared to the single genetic algorithm. When serial calculation is used, there is a problem of low computational efficiency. In addition, the parameter setting in the genetic algorithm has an important impact on the performance and convergence speed of the algorithm. The adjustment of parameters requires repeated trials and adjustments. In practical problems, different problems may have different sensitivity to parameters; it cannot meet the complex needs of the site, such as some barge tasks need to be bargeed first. Therefore, it is difficult to determine the best parameter combination.

[0006] In summary, it is necessary to provide an improved technical solution to address the above-mentioned deficiencies in the prior art. Summary of the invention

[0007] The purpose of the embodiments of the present application is to provide a scheduling method and system for ship-stage transfer vehicles, which can solve the problem of difficult parameter selection, and can also quantitatively judge the advantages and disadvantages of different schemes to achieve intelligent and efficient control of ship-stage transfer vehicles.

[0008] In a first aspect, a method for dispatching a ship to transport vehicles in sections is provided, comprising the following steps:

[0009] S1. Obtain historical lightering tasks of ship sections, and establish a basic database based on the historical lightering tasks;

[0010] S2. Use a genetic algorithm to calculate the historical lightering tasks in the basic database, adjust the algorithm parameters of the genetic algorithm, and configure each algorithm parameter and the corresponding calculation result as a set of training data, and input multiple sets of training data into the training database;

[0011] S3, training a generative adversarial network model based on the training database;

[0012] S4, obtaining a real-time lightering task, and generating real-time algorithm parameters through the generative adversarial network model;

[0013] S5, obtaining an optimal lightering task sequence corresponding to the real-time lightering task based on real-time algorithm parameters and genetic algorithm;

[0014] S6. The optimal transfer task sequence is converted into a driving route instruction and output to complete the dispatch of transfer vehicles.

[0015] In an implementable manner, the calculation result at least includes a barge task sequence of daily barge tasks.

[0016] In one practicable manner, the generative adversarial network model includes at least a generative model and a discriminative model; the generative model is used to output corresponding algorithm parameters after inputting the transshipment task; the discriminative model is used to output good and bad labels after inputting the algorithm parameters.

[0017] In one practicable manner, in step S4, at least the following steps are included:

[0018] S41, reading the real-time lightering task data input by the dispatcher;

[0019] S42, inputting the real-time lightering task data into the generation model;

[0020] S43, the generation model generates and outputs corresponding algorithm parameters as real-time algorithm parameters according to the acquired mapping relationship.

[0021] In one practicable manner, in step S5, at least the following steps are included:

[0022] S51, randomly generating multiple lightering task sequences based on the real-time lightering tasks, and using the multiple lightering task sequences as the initial population of the genetic algorithm;

[0023] S52, taking each barge task sequence as an individual, and calculating the individual fitness corresponding to each barge task sequence;

[0024] S53, selecting the barge task sequence corresponding to the maximum individual fitness, and judging whether the individual fitness of the current barge task sequence meets the evolution termination condition; when the evolution termination condition is met, outputting the current barge task sequence as the optimal barge task sequence.

[0025] In one practicable manner, the individual fitness is determined based on the total idle time, total driving time, and total driving time of the flatbed truck of the real-time transfer task.

[0026] In one possible implementation, the individual fitness is calculated as:

[0027]

[0028] H: individual fitness;

[0029] X 1 : Calculation weight of total idle time;

[0030] X 2 : Calculation weight of total driving time;

[0031] X 3 : The calculation weight of the total driving time of the flatbed truck, and X 1 +X 2 +X 3 =1;

[0032] Y: time penalty weight;

[0033] Z: overweight penalty weight;

[0034] Q: Priority encouragement value;

[0035] A: Task sequence value of the priority task;

[0036] N: number of flatbed trucks;

[0037] S i : The empty distance of the i-th flatbed truck;

[0038] v1: The flatbed truck has an empty speed of 3 km / h;

[0039] T i : The load distance of the i-th flatbed truck;

[0040] v2: The flatbed truck has a load-carrying speed of 2 km / h;

[0041] μ: The mean value of the data set formed by the total driving time of each flatbed truck.

[0042] In an practicable manner, in step S53, judging whether the individual fitness of the current lightering task sequence meets the evolution termination condition at least includes the following contents:

[0043] S531. Perform a catastrophic change on the current population to generate a new population;

[0044] S532, repeating steps S52 to S53 to obtain the optimal transshipment task sequence after the disaster;

[0045] S532. Compare the individual fitness of the optimal barge task sequence after the disaster with the individual fitness of the optimal barge task sequence before the disaster, and determine whether the new population meets the disaster termination conditions based on the comparison results. When the new population meets the disaster termination conditions, determine and output the optimal barge task sequence.

[0046] In an practicable manner, in step S532, at least the following contents are also included:

[0047] When the individual fitness of the optimal barge task sequence after the disaster is the same as the individual fitness of the optimal barge task sequence before the disaster, and when the individual fitness remains the same after a predetermined number of disasters, the optimal barge task sequence after the disaster or the optimal barge task sequence before the disaster is output;

[0048] When the individual fitness of the optimal barge task sequence after the disaster is less than that of the optimal barge task sequence before the disaster, the optimal barge task sequence before the disaster is output;

[0049] When the individual fitness of the optimal barge task sequence after the disaster is greater than the individual fitness of the optimal barge task sequence before the disaster, steps S531 to S532 are repeated until the current population meets the disaster termination condition, and the optimal barge task sequence is determined and output.

[0050] According to a second aspect of the present application, a scheduling system for ship-staged transfer of vehicles is also provided, comprising:

[0051] The training and reasoning module is used to train the generative adversarial network model and to infer and optimize the computational parameters of the genetic algorithm;

[0052] Basic data management module, used to maintain and manage historical lightering task data of ship sections;

[0053] The lightering task sequence calculation module is used to calculate and output the optimal lightering task sequence of the ship sections;

[0054] The vehicle dispatching module is used to receive the optimal transfer task sequence and generate and output the driving route instructions.

[0055] In an practicable manner, the dispatching system further includes an execution terminal, which performs corresponding lightering tasks upon receiving the formal route instruction, thereby completing the dispatching of lightering of the ship sections.

[0056] Compared with the prior art, the beneficial effects of this application are:

[0057] In the technical solution of the present application, the algorithm is optimized by adjusting the genetic algorithm parameters, which can solve the problem of difficulty in selecting existing algorithm parameters. When calculating individual fitness, by assigning different weights to the total idle time standard deviation, the total driving time standard deviation, and the total driving time standard deviation of the flatbed truck, and then adding the inverse as the objective function, the distance is converted into time for calculation to improve the accuracy of the calculation. By adding the priority encouragement value calculation, the need for priority scheduling of some tasks can be met. Parallel computing is achieved by adding multi-threading to increase the calculation speed of the algorithm. The present application can not only accurately quantify the advantages and disadvantages of different schemes, but also solve the problems of difficulty in selecting existing algorithm parameters, low calculation efficiency, difficulty in quantifying the advantages and disadvantages of different schemes for scheduling ship segmented transfer vehicles, difficulty in achieving better scheduling schemes for ship segmented transfer vehicles, and inability to adapt to changes in actual conditions, so as to realize intelligent and efficient management and control of ship segmented transfer vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 The present invention is a flowchart of a method for dispatching vehicles by ship in sections according to an embodiment of the present invention.

[0059] Figure 2 Schematic diagram of a dispatching system for ship-stage transfer of vehicles according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. These embodiments are only used to illustrate the present invention, but not to limit the present invention.

[0061] In the description of the present invention, it should be noted that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0062] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0063] Furthermore, in the description of the present invention, unless otherwise specified, “plurality” means two or more.

[0064] According to the first aspect of the present application, see Figure 1 First, a method for dispatching a ship to transport vehicles in sections is provided, comprising the following steps:

[0065] S1. Obtain historical lightering tasks of ship sections, and establish a basic database based on the historical lightering tasks;

[0066] The historical lightering tasks include at least a plurality of daily lightering tasks.

[0067] S2. Use a genetic algorithm to calculate the historical transshipment tasks in the basic database, adjust the algorithm parameters of the genetic algorithm, and configure each algorithm parameter and the corresponding calculation result as a set of training data, and input multiple sets of training data into the training database.

[0068] It should be noted that the calculation result at least includes the transfer task sequence of the daily transfer task.

[0069] It should also be noted that the algorithm parameters include at least setting optimization objectives, fitness functions, priority functions, end conditions, disaster conditions, priority scheduling tasks, initial population size, optimal number of individuals in each generation, crossover probability, mutation probability, fitness parameters, overweight penalty parameters, timeout penalty parameters, and priority encouragement parameters.

[0070] S3. Training a generative adversarial network model based on the training database.

[0071] In an practicable manner, the generative adversarial network model includes at least a generative model and a discriminative model. The generative model is used to input the barge task and then output the corresponding algorithm parameters. The discriminative model is used to input the algorithm parameters and then output the good and bad labels.

[0072] S4. Obtain real-time lightering tasks, and generate real-time algorithm parameters through the generative adversarial network model.

[0073] In one practicable manner, in step S4, at least the following steps are included:

[0074] S41. Read the real-time transfer task data input by the dispatcher.

[0075] S42, inputting the real-time lightering task data into the generation model.

[0076] S43, the generation model generates and outputs corresponding algorithm parameters as real-time algorithm parameters according to the acquired mapping relationship.

[0077] S5. Obtain the optimal barge task sequence corresponding to the real-time barge task based on the real-time algorithm parameters and the genetic algorithm.

[0078] In one practicable manner, in step S5, at least the following steps are included:

[0079] S51. Randomly generate multiple lightering task sequences based on the real-time lightering tasks, and use the multiple lightering task sequences as the initial population of the genetic algorithm.

[0080] S52, taking each barge task sequence as an individual, and calculating the individual fitness corresponding to each barge task sequence.

[0081] It should be noted that the individual fitness is determined based on the total idle time, total driving time, and total driving time of the flatbed truck in the real-time transshipment task.

[0082] Specifically, the total idle time, total driving time and standard deviation of the total driving time of the flatbed trucks in each individual are calculated. The difference in the total driving time of the flatbed trucks is described by the standard deviation. The weights of the three are added together and the inverse is taken as the individual fitness H:

[0083]

[0084] H: individual fitness;

[0085] X 1 : Calculation weight of total dead time;

[0086] X 2 : Calculation weight of total driving time;

[0087] X 3 : The calculation weight of the total driving time of the flatbed truck, and X 1 +X 2 +X 3 =1;

[0088] Y: time penalty weight;

[0089] Z: overweight penalty weight;

[0090] Q: Priority encouragement value;

[0091] A: Task sequence value of the priority task;

[0092] N: number of flatbed trucks;

[0093] S i : The empty distance of the i-th flatbed truck;

[0094] v1: The flatbed truck has an empty speed of 3 km / h;

[0095] T i : The load distance of the i-th flatbed truck;

[0096] v2: The flatbed truck has a load speed of 2km / h;

[0097] μ: The mean value of the data set formed by the total driving time of each flatbed truck.

[0098] The present application converts the distance into time to calculate the individual fitness, taking into account the different speeds of a flatbed truck with cargo and without cargo. When the transportation distance is the same, there is a difference in time. Therefore, the individual fitness calculated in the present application is more accurate.

[0099] S53, selecting the barge task sequence corresponding to the maximum individual fitness, and judging whether the individual fitness of the current barge task sequence meets the evolution termination condition. When the evolution termination condition is met, outputting the current barge task sequence as the optimal barge task sequence.

[0100] In an practicable manner, judging whether the individual fitness of the current barge task sequence meets the evolution termination condition at least includes the following contents:

[0101] S531. Carry out a catastrophic change on the current population to generate a new population.

[0102] S532. Repeat steps S52 to S53 to obtain the optimal transshipment task sequence after the disaster.

[0103] S532. Compare the individual fitness of the optimal barge task sequence after the disaster with the individual fitness of the optimal barge task sequence before the disaster, and determine whether the new population meets the disaster termination conditions based on the comparison results. When the new population meets the disaster termination conditions, determine and output the optimal barge task sequence.

[0104] Specifically, when the individual fitness of the optimal barge task sequence after the disaster is the same as the individual fitness of the optimal barge task sequence before the disaster, and when the individual fitness remains the same after a predetermined number of disasters, the optimal barge task sequence after the disaster or the optimal barge task sequence before the disaster will be output.

[0105] When the individual fitness of the optimal barge task sequence after the disaster is less than that of the optimal barge task sequence before the disaster, the optimal barge task sequence before the disaster is output.

[0106] When the individual fitness of the optimal barge task sequence after the disaster is greater than the individual fitness of the optimal barge task sequence before the disaster, steps S531 to S532 are repeated until the current population meets the disaster termination condition, and the optimal barge task sequence is determined and output.

[0107] S6. The daily transfer task sequence in step S5 is converted into a driving route instruction and output, thereby completing the dispatch of transfer vehicles.

[0108] See also Figure 2 According to the second aspect of the present application, a scheduling system for ship-staged transfer of vehicles is also provided, comprising:

[0109] The training and reasoning module is used to train the generative adversarial network model and to infer and optimize the computational parameters of the genetic algorithm;

[0110] Basic data management module, used to maintain and manage historical lightering task data of ship sections;

[0111] The lightering task sequence calculation module is used to calculate and output the optimal lightering task sequence of the ship sections;

[0112] The vehicle dispatching module is used to receive the optimal transfer task sequence and generate and output the driving route instructions.

[0113] In an practicable manner, the dispatching system further includes an execution terminal, which performs corresponding lightering tasks upon receiving the formal route instruction, thereby completing the dispatching of lightering of the ship sections.

[0114] In summary, the present application optimizes the algorithm by adjusting the genetic algorithm parameters, which can solve the problem of difficulty in selecting existing algorithm parameters. When calculating individual fitness, by assigning different weights to the total idle time standard deviation, the total driving time standard deviation, and the total driving time standard deviation of the flatbed truck, and then adding the inverse as the objective function, the distance is converted into time for calculation to improve the accuracy of the calculation. By adding the priority encouragement value calculation, the need for priority scheduling of some tasks can be met. Parallel computing is achieved by adding multi-threading to increase the calculation speed of the algorithm. The present application can not only accurately quantify the advantages and disadvantages of different schemes, but also solve the problems of difficulty in selecting existing algorithm parameters, low calculation efficiency, difficulty in quantifying the advantages and disadvantages of different schemes for scheduling ship segmented transfer vehicles, difficulty in achieving better scheduling schemes for ship segmented transfer vehicles, and inability to adapt to changes in actual conditions, so as to realize intelligent and efficient management and control of ship segmented transfer vehicles.

[0115] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.

Claims

1. A method for dispatching vehicles for ship-to-ship transport, characterized in that: The following steps are involved: S1. Obtain historical lightering tasks of ship sections, and establish a basic database based on the historical lightering tasks; S2. Use a genetic algorithm to calculate the historical lightering tasks in the basic database, adjust the algorithm parameters of the genetic algorithm, and configure each algorithm parameter and the corresponding calculation result as a set of training data, and input multiple sets of training data into the training database; S3, training a generative adversarial network model based on the training database; S4, obtaining a real-time lightering task, and generating real-time algorithm parameters through the generative adversarial network model; S5, obtaining an optimal lightering task sequence corresponding to the real-time lightering task based on real-time algorithm parameters and genetic algorithm; S6. The optimal transfer task sequence is converted into a driving route instruction and output to complete the dispatch of transfer vehicles.

2. The method for dispatching ship-stage transfer vehicles according to claim 1, characterized in that: The calculation result at least includes the lightening task sequence of the daily lightening task.

3. The method for dispatching ship-stage transfer vehicles according to claim 1, characterized in that: The generative adversarial network model includes at least a generative model and a discriminative model; the generative model is used to input the barge task and output the corresponding algorithm parameters; the discriminative model is used to input the algorithm parameters and output the good and bad labels.

4. The method for dispatching ship-stage transfer vehicles according to claim 1, characterized in that: In step S4, at least the following steps are included: S41, reading the real-time lightering task data input by the dispatcher; S42, inputting the real-time lightering task data into the generation model; S43, the generation model generates and outputs corresponding algorithm parameters as real-time algorithm parameters according to the acquired mapping relationship.

5. The method for dispatching ship-stage transfer vehicles according to claim 1, characterized in that: In step S5, at least the following steps are included: S51, randomly generating multiple lightering task sequences based on the real-time lightering tasks, and using the multiple lightering task sequences as the initial population of the genetic algorithm; S52, taking each barge task sequence as an individual, and calculating the individual fitness corresponding to each barge task sequence; S53, selecting the barge task sequence corresponding to the maximum individual fitness, and judging whether the individual fitness of the current barge task sequence meets the evolution termination condition; when the evolution termination condition is met, outputting the current barge task sequence as the optimal barge task sequence.

6. The method for dispatching ship-stage transfer vehicles according to claim 5, characterized in that: The individual fitness is determined according to the total idle time, total driving time, and total driving time of the flatbed truck of the real-time transshipment task.

7. The method for dispatching ship-stage transfer vehicles according to claim 6, characterized in that: The calculation of individual fitness is: H: individual fitness; X1: Calculation weight of total idle time; X2: calculation weight of total travel time; X3: calculation weight of the total driving time of the flatbed truck, and X1+X2+X3=1; Y: time penalty weight; Z: overweight penalty weight; Q: Priority encouragement value; A: Task sequence value of the priority task; N: number of flatbed trucks; S i : The empty distance of the i-th flatbed truck; v1: The flatbed truck has an empty speed of 3 km / h; T i : The load distance of the i-th flatbed truck; v2: The flatbed truck has a load speed of 2km / h; μ: The mean value of the data set formed by the total driving time of each flatbed truck.

8. The method for dispatching ship-stage transfer vehicles according to claim 1, characterized in that: In step S53, judging whether the individual fitness of the current barge task sequence meets the evolution termination condition includes at least the following contents: S531. Perform a catastrophic change on the current population to generate a new population; S532, repeating steps S52 to S53 to obtain the optimal transshipment task sequence after the disaster; S532. Compare the individual fitness of the optimal barge task sequence after the disaster with the individual fitness of the optimal barge task sequence before the disaster, and determine whether the new population meets the disaster termination conditions based on the comparison results. When the new population meets the disaster termination conditions, determine and output the optimal barge task sequence.

9. The method for dispatching ship-stage transfer vehicles according to claim 8, characterized in that: In step S532, at least the following contents are included: When the individual fitness of the optimal barge task sequence after the disaster is the same as the individual fitness of the optimal barge task sequence before the disaster, and when the individual fitness remains the same after a predetermined number of disasters, the optimal barge task sequence after the disaster or the optimal barge task sequence before the disaster is output; When the individual fitness of the optimal barge task sequence after the disaster is less than that of the optimal barge task sequence before the disaster, the optimal barge task sequence before the disaster is output; When the individual fitness of the optimal barge task sequence after the disaster is greater than the individual fitness of the optimal barge task sequence before the disaster, steps S531 to S532 are repeated until the current population meets the disaster termination condition, and the optimal barge task sequence is determined and output.

10. A dispatching system for ship-stage transfer of vehicles, characterized in that: include: The training and reasoning module is used to train the generative adversarial network model and to infer and optimize the computational parameters of the genetic algorithm; Basic data management module, used to maintain and manage historical lightering task data of ship sections; The lightering task sequence calculation module is used to calculate and output the optimal lightering task sequence of the ship sections; The vehicle dispatching module is used to receive the optimal transfer task sequence and generate and output the driving route instructions.

11. The dispatching system for ship-stage transfer vehicles according to claim 10, characterized in that: The dispatching system also includes an execution terminal, which performs corresponding lightering tasks after receiving the formal route instruction to complete the lightering dispatch of the ship sections.

Images