Railway container station collaborative scheduling optimization method

Abstract

The invention discloses a railway container station collaborative scheduling optimization method, and belongs to the technical field of task scheduling, and the method comprises the steps: obtaining an equipment work form which comprises the operation sequence and the distribution scheme of each equipment; according to the equipment worksheet, obtaining an interference equipment scheduling table through space-time state deduction; dividing potential interference of the equipment into different types; in the interference device dispatch table, detecting and adjusting potential interference of various devices by using an interference detection and adjustment algorithm, and finally obtaining an interference-free device dispatch table; according to the non-interference equipment scheduling table, the state of the storage yard is deduced step by step, the completeness of operation is detected, and the fitness of the corresponding individual is assigned according to the detection result. The method has the effect of improving the rail-mounted crane-reach stacker collaborative operation efficiency.

Description

Technical Field

[0001] The present application relates to the technical field of task scheduling, and in particular to a method for optimizing coordinated scheduling of railway container terminals. Background Art

[0002] Railway container terminals are important nodes for container multimodal transport. With the increasing number of China-Europe trains, improving the operational organization efficiency of container terminals through optimization of loading and unloading equipment scheduling is an important way to improve the quality of railway container transport services and promote the development of multimodal transport.

[0003] Currently, there are few studies on the scheduling of loading and unloading equipment in railway yards. Most of them focus on the scheduling of rail cranes. In container yards, reach stackers, as a mobile machine, often cooperate with rail cranes to load and unload containers. Compared with rail cranes, reach stackers have certain advantages in operating efficiency and flexibility, but their operating capacity is greatly limited by the storage conditions of the yard. In addition, due to the differences in the operating methods of reach stackers and rail cranes, there are obvious differences between the collaborative operation of reach stackers-rail cranes and the collaborative operation of multiple rail cranes in terms of interference rules. Therefore, how to improve the efficiency of rail crane-reach stacker collaborative operation is a more complex optimization scheduling problem. Summary of the invention

[0004] In order to solve the collaborative scheduling problem, the present application provides a railway container terminal collaborative scheduling optimization method.

[0005] The present application provides a railway container terminal coordinated dispatch optimization method using the following technical solutions:

[0006] Obtaining a device work list, wherein the device work list includes a job sequence and an allocation plan for each device;

[0007] According to the equipment work form, obtaining the interfering equipment scheduling table through time-space state deduction;

[0008] The potential interference of equipment is divided into the following types: the rail crane and the reach crane are loading and unloading at the same bay at the same time; when the rail crane is loading and unloading, the reach crane actively passes under the rail crane; when the reach crane is loading and unloading, the rail crane with heavy load actively passes over the reach crane; when the rail crane with heavy load is moving, the reach crane actively passes under the rail crane;

[0009] In the interference equipment scheduling table, an interference detection and adjustment algorithm is used to detect and adjust the potential interference of various types of equipment, and finally a non-interference equipment scheduling table is obtained;

[0010] According to the non-interference equipment scheduling table, the yard status is gradually deduced, the feasibility of the operation is tested, and the fitness of the corresponding individual is assigned according to the test results.

[0011] Furthermore, ignoring the influence of human and natural factors, the starting and ending positions of the task are known, and the container does not need to overturn, the objective function shown in formula (1) is established;

[0012]

[0013] Establish the constraints shown in formula (2) to formula (17):

[0014]

[0015]

[0016] in:

[0017] i,j represent the container task number; N represents the total number of containers in the container task; k represents the loading and unloading equipment number, where equipment 0 is a reach stacker and equipment 1 is a rail stacker; M represents a sufficiently large positive number; represents the time when device k performs task i to grab the container; It indicates the time when device k performs task i to place the container; Indicates the box-picking operation time of device k; S represents the box placement operation time of device k; kt represents the location of device k at time t; Indicates the starting position of task i, representing stack position, shell position, and layer position respectively; represents the end position of task i, which represents the stack position, shell position, and layer position respectively; X, Y, and Z represent the number of stacks, shells, and layers in the yard area respectively; X 1 ,X 2 Indicates a special mark stack position, where X 1 X is the train loading and unloading line closest to the container truck running line. 2 For truck loading and unloading lines; Indicates the moving distance of equipment k in the direction of the storage yard when performing task i; Indicates the moving distance of device k in the direction of the stacking position when executing task i; represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the direction of the yard bay position; represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the stacking direction of the yard; h represents the safe distance that the reach stacker needs to maintain when the rail crane is in a heavy load or loading and unloading state; It represents the average moving speed of equipment k along the direction of the yard bay; represents the average moving speed of equipment k along the stacking direction; u ij,k∈{0,1}: If the container task numbered i, j is assigned to the operating device k and the operating device k completes the container task numbered i and then performs the container task j, then u ij,k =1, otherwise u ij,k = 0. In addition, when tasks i and j are the same task, u ij,k =1 to 0; g xyz,t ∈0,1: If there are containers in slots x, y, z at time t, then g xyz,t =1, otherwise g xyz,t =0;

[0018] Formula (1) is the objective function, which indicates that the completion time of the batch job is the shortest;

[0019] Formula (2) indicates that the starting point of the task assigned to the reach stacker cannot be located at X 1 Below the stack position;

[0020] Formula (3) indicates that the end point of the task assigned to the reach stacker cannot be located at X 1 Below the stack position;

[0021] Formula (4) represents the time relationship between the time when the next task is grabbed by the track crane and the time when the previous task is placed;

[0022] Formula (5) represents the time relationship between the grabbing time and the placing time of the current task of the rail crane;

[0023] Formula (6) represents the time relationship between the grabbing time of the next task and the placing time of the previous task of the front crane;

[0024] Formula (7) represents the time relationship between the grabbing time and the placing time of the current task of the reach stacker;

[0025] Formula (8) indicates that each task is assigned to at most one device;

[0026] Formula (9) indicates that each task has at most one subsequent task;

[0027] Formula (10) indicates that each task has at most one immediate predecessor task;

[0028] Formula (11) indicates that the reach stacker should prioritize tasks whose starting points are close to the container loading and unloading line, and then execute tasks whose starting points are far from the loading and unloading line;

[0029] Formula (12) indicates that the reach stacker should prioritize tasks whose end points are far from the container loading and unloading line, and then execute tasks whose end points are close to the loading and unloading line;

[0030] Formula (13) indicates that the loading and unloading equipment should first process the tasks whose starting points are located at the high level, and then process the tasks whose starting points are located at the low level;

[0031] Formula (14) indicates that the loading and unloading equipment should first process the tasks whose endpoints are located at the lower level, and then process the tasks whose endpoints are located at the higher level;

[0032] Formula (15) indicates that there is no container blocking the travel route of the reach stacker to grab the container;

[0033] Formula (16) indicates that there is no container blocking the route of the reach stacker to place the container;

[0034] Formula (17) indicates that during the loading and unloading operations and the movement of loads on the rail crane, the reach stacker should maintain a certain safety distance from it.

[0035] Furthermore, the adjustment algorithm adopts a dual-stage coding mode, which consists of a three-stage spatiotemporal planning fitness calculation method, an elite retention strategy, a tournament selection operator, a synchronous sequential crossover operator, an adaptive two-point mutation operator and a heuristic population guidance operator.

[0036] Furthermore, the dual-segment coding mode adopts an integer coding mode and uses dual-segment gene expression in chromosome coding design. The genes in the first half of the chromosome are job sequence genes, which are used to express the job sequence scheme; the genes in the second half of the chromosome are task allocation genes, which are used to express the task allocation scheme.

[0037] Furthermore, the synchronous sequential crossover operator is based on the double-segment coding mode of the population individuals. For the operation order layer of the chromosome, an individual is randomly selected from the selected parent individuals, the gene segment between the two randomly selected crossover points is copied to the daughter chromosome, and the index position is recorded;

[0038] Delete the genes in the other parent individual that are repeated in the offspring chromosome, copy the remaining genes to the vacant positions of the offspring individual chromosomes in sequence, and record the index positions; finally, perform the above operations on the chromosome assignment layer according to the index positions to obtain the offspring chromosomes.

[0039] Furthermore, the adaptive two-point mutation operator divides the individuals in the population into high-quality individuals, low-quality individuals and infeasible individuals according to the average fitness of the population and the feasibility of the individual decoding scheme. During the evolution process, the mutation probability of the individuals is dynamically adjusted according to their fitness performance, as shown in Formula 18:

[0040]

[0041] Furthermore, the heuristic population guidance operator has the following steps:

[0042] Fitness evaluation: Perform fitness evaluation on individuals in the population and select individuals with fitness in the bottom N% of the population, where N is a variable parameter;

[0043] Waiting time calculation: Decode the selected individuals, use the three-stage spatiotemporal planning fitness calculation method to obtain the corresponding equipment scheduling table, and use the equipment scheduling table to determine the waiting time of each task in the scheduling plan;

[0044] Identify and rearrange key tasks: Sort all tasks in descending order according to waiting time, take the task with the longest waiting time as the key task, and try to place it at the beginning or end of the task sequence to come up with alternative solutions;

[0045] Fitness re-evaluation: re-evaluate the fitness of the adjusted sequence to verify whether the position adjustment effectively improves the fitness of the individual;

[0046] Scheme comparison and selection: Compare the adjusted scheduling scheme with the original scheme, select the optimal scheme for encoding to generate new individuals, and replace the original individuals in the population.

[0047] In summary, compared with the prior art, the above technical solution has the following beneficial effects:

[0048] The present application discloses a railway container yard coordinated dispatch optimization method, which avoids interference between equipment through precise equipment dispatch, thereby reducing the interruption of operation and waiting time caused by interference. The equipment can work according to the optimal operation sequence and allocation scheme, thereby improving the overall operation efficiency. The equipment dispatch table ensures the reasonable allocation of each equipment in time and space, avoiding the waste and idleness of resources. By gradually deducing the state of the yard, the utilization of resources can be further optimized and the utilization rate of equipment can be improved. The interference detection and adjustment algorithm can timely detect and adjust the potential interference between equipment, thereby avoiding safety hazards caused by accidents such as equipment collision. By detecting the completeability of the operation, it can be ensured that the equipment operates under safe conditions. This process relies on advanced information technology and algorithms to realize the intelligence and automation of equipment dispatch. Decision makers can make more scientific and reasonable decisions based on the non-interference equipment dispatch table and deduction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a typical layout diagram of the railway container terminal in the embodiment of the present application.

[0050] Figure 2 It is a flow chart of the improved genetic algorithm in the embodiment of the present application.

[0051] Figure 3 It is a schematic diagram of chromosome encoding in the embodiments of this application.

[0052] Figure 4 This is a schematic diagram of the first type of potential interference adjustment in the embodiment of the present application.

[0053] Figure 5 This is a schematic diagram of the second type of potential interference adjustment in the embodiment of the present application.

[0054] Figure 6 This is a schematic diagram of the third type of potential interference adjustment in the embodiment of the present application.

[0055] Figure 7 This is a schematic diagram of the fourth type of potential interference adjustment in the embodiment of the present application.

[0056] Figure 8 It is a flow chart of the interference detection and adjustment algorithm in the embodiment of the present application.

[0057] Fig. 9 It is a schematic diagram of synchronous sequential cross-operation based on job sequence and task allocation in an embodiment of the present application.

[0058] Fig.10 It is a schematic diagram of two-point variation operation in an embodiment of the present application.

[0059] Fig.11 It is a schematic diagram of key task identification and rearrangement operations in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The present application is further described in detail below in conjunction with all the accompanying drawings.

[0061] The present application embodiment discloses a railway container terminal coordinated dispatch optimization method, referring to Figure 1 In a typical railway container terminal, assuming that there are N tasks of loading and unloading container trains or container trucks in a waiting batch, and the starting and ending positions of each task are known, consider how to assign N tasks to a rail crane and a reach crane for simultaneous operation, and reasonably plan the operation sequence of each equipment to shorten the completion time of the batch tasks as much as possible. Different from the problem of collaborative operation scheduling of multiple rail cranes, the two types of equipment are allowed to cross each other when the rail crane is unloaded, but there is potential interference between the equipment during the loading and unloading operation or heavy load movement of the rail crane. In addition, the reach crane is affected by the occupancy status of the container storage location and the layout of the train loading and unloading line, and there are regional restrictions where some cargo spaces cannot be operated.

[0062] The technical solution adopted by the present invention to solve the technical problem includes a modeling method for the coordinated scheduling problem of rail crane-reach stacker in railway container terminals and an improved genetic algorithm. The modeling method for the coordinated scheduling problem of multiple types of loading and unloading equipment in railway terminals has the following steps:

[0063] First, define the following parameters:

[0064] i,j represent the container task number;

[0065] N represents the total number of containers for the container mission;

[0066] k represents the loading and unloading equipment number, where equipment 0 is a reach stacker and equipment 1 is a rail stacker;

[0067] M represents a sufficiently large positive number;

[0068] represents the time when device k performs task i to grab the container;

[0069] It indicates the time when device k performs task i to place the container;

[0070] Indicates the box-picking operation time of device k;

[0071] Indicates the box placement operation time of device k;

[0072] S kt represents the location of device k at time t;

[0073] Indicates the starting position of task i, representing stack position, shell position, and layer position respectively;

[0074] Indicates the end position of task i, representing stack position, shell position, and layer position respectively;

[0075] X, Y, and Z represent the number of stacks, shells, and layers in the yard area, respectively;

[0076] X 1 ,X 2 Indicates a special mark stack position. For specific positions, see Figure 1 ;

[0077] Indicates the moving distance of equipment k in the direction of the storage yard when performing task i;

[0078] Indicates the moving distance of device k in the direction of the stacking position when executing task i;

[0079] represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the direction of the yard bay position;

[0080] represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the stacking direction of the yard;

[0081] h represents the safe distance that the reach stacker needs to maintain when the rail crane is under heavy load or loading and unloading state;

[0082] It represents the average moving speed of equipment k along the direction of the yard bay;

[0083] represents the average moving speed of equipment k along the stacking direction of the yard;

[0084] u ij,k ∈{0,1}: If the container task numbered i, j is assigned to the operating device k and the operating device k completes the container task numbered i and then performs the container task j, then u ij,k =1, otherwise u ij,k = 0. In addition, when tasks i and j are the same task, u ij,k =1 to 0;

[0085] g xyz,t ∈0,1: If there are containers in slots x, y, z at time t, then g xyz,t =1, otherwise g xyz,t =0.

[0086] Then, assume the following conditions:

[0087] Condition 1: Ignore the impact of human and natural factors;

[0088] Condition 2: The starting and ending positions of the task are known;

[0089] Condition 3: The container types are the same, all are 20-foot standard containers;

[0090] Condition 4: The travel speed of the rail crane / reach crane, and the time for picking up and unloading containers are all fixed parameters;

[0091] Condition 5: The situation where the container needs to be overturned is not considered.

[0092] Finally, the objective function as shown in formula (1) is established, and the constraints as shown in formulas (2) to (17) are established:

[0093]

[0094]

[0095] in:

[0096] Formula (1) is the objective function, which indicates that the completion time of the batch job is the shortest;

[0097] Formula (2) indicates that the starting point of the task assigned to the reach stacker cannot be located at X 1 Below the stack;

[0098] Formula (3) indicates that the end point of the task assigned to the reach stacker cannot be located at X 1 Below the stack;

[0099] Formula (4) represents the time relationship between the time when the next task is grabbed by the track crane and the time when the previous task is placed;

[0100] Formula (5) represents the time relationship between the grabbing time and the placing time of the current task of the rail crane;

[0101] Formula (6) represents the time relationship between the grabbing time of the next task and the placing time of the previous task of the front crane;

[0102] Formula (7) represents the time relationship between the grabbing time and the placing time of the current task of the reach stacker;

[0103] Formula (8) indicates that each task is assigned to at most one device;

[0104] Formula (9) indicates that each task has at most one subsequent task;

[0105] Formula (10) indicates that each task has at most one immediate predecessor task;

[0106] Formula (11) indicates that the reach stacker should prioritize tasks whose starting points are close to the container loading and unloading line, and then execute tasks whose starting points are far from the loading and unloading line;

[0107] Formula (12) indicates that the reach stacker should prioritize tasks whose end points are far from the container loading and unloading line, and then execute tasks whose end points are close to the loading and unloading line;

[0108] Formula (13) indicates that the loading and unloading equipment should first process the tasks whose starting points are located at the high level, and then process the tasks whose starting points are located at the low level;

[0109] Formula (14) indicates that the loading and unloading equipment should first process the tasks whose endpoints are located at the lower level, and then process the tasks whose endpoints are located at the higher level;

[0110] Formula (15) indicates that there is no container blocking the travel route of the reach stacker to grab the container;

[0111] Formula (16) indicates that there is no container blocking the route of the reach stacker to place the container;

[0112] Formula (17) indicates that during the loading and unloading operations and the movement of loads on the rail crane, the reach stacker should maintain a certain safety distance from it.

[0113] The improved genetic algorithm proposed in the present invention is based on the genetic algorithm, and designs a three-stage spatiotemporal planning fitness calculation method to solve the problems of chromosome fitness calculation and feasibility judgment. It introduces an elite retention strategy, a synchronous sequential crossover operator, an adaptive mutation operator, and a heuristic population guidance operator to improve the fitness convergence speed and stability. The flowchart of the improved genetic algorithm is shown in the figure. Figure 2 shown.

[0114] The improved genetic algorithm proposed in the present invention adopts an integer coding mode and uses double-segment gene expression in chromosome coding design. The genes in the first half of the chromosome are job sequence genes for expressing job sequence schemes; the genes in the second half of the chromosome are task allocation genes for expressing task allocation schemes. Figure 3 For example, the first gene is coded as a task number, and the second gene is coded as the type of equipment. 0 represents a rail crane, and 1 represents a reach crane. The same sequence number in the two gene segments represents the equipment performing the task. In the coding example, tasks 1 and 2 are assigned to reach cranes, and the task sequence is 2, 1; tasks 3, 4, and 5 are assigned to rail cranes, and the task sequence is 4, 3, and 5.

[0115] The improved genetic algorithm proposed in the present invention has the following steps for calculating the fitness of three-stage spatiotemporal planning:

[0116] According to the equipment task allocation and operation sequence obtained after decoding by the genetic algorithm, the equipment interference scheduling table is generated. For the convenience of description, the following symbols are added: represents the initial position of device k, Indicates the initial stack position of device k, represents the position of the grasping position of task i, The stack position representing the grab position of task i, represents the position of task i, The stack position indicating the placement of task i, d b Indicates the length of a container space in the yard, d s Indicates the width of a container space in the yard. represents the time when device k starts to move to the starting point of task i, represents the time when device k arrives at the starting point of task i, represents the time when device k starts picking up the box of task i, represents the time when device k starts to move to the end point of task i, represents the time when device k reaches the end point of task i, Indicates the time when device k starts the box placement operation of task i. In the first stage, the potential interference between the rail crane and the reach crane is not considered. The process of obtaining the scheduling table of the interfering equipment through time-space state deduction is as follows: If device k is a reach crane and task i is the first task of the reach crane, then:

[0117]

[0118]

[0119] If task i is not the first task of the reach stacker and its predecessor task is j, then

[0120]

[0121] The subsequent deduction process is the same as the first task. If the device is a rail crane, the above deduction process needs to be adjusted. If the device k is a rail crane, and if task i is the first task of the rail crane, then:

[0122]

[0123] If task i is not the first task of the rail crane and its predecessor task is j, then:

[0124]

[0125] The subsequent deduction process is the same as the first task.

[0126] In the equipment interference scheduling table, potential equipment interference is detected and adjusted. Potential equipment interference is divided into the following four types: (1) The rail crane and the reach crane are loading and unloading at the same bay at the same time; (2) When the rail crane is loading and unloading, the reach crane actively passes under the rail crane; (3) When the reach crane is loading and unloading, the rail crane with heavy load actively passes above the reach crane; (4) When the rail crane is moving with heavy load, the reach crane actively passes under the rail crane.

[0127] The first type of potential interference detection and adjustment process is Figure 4 Take this as an example. Figure 4 (a) is the time-space state diagram of the equipment before adjustment. At position "5", two equipments are simultaneously loaded and unloaded. The adjustment method is as follows: Figure 4 As shown in (b), by delaying the start time of the rail crane task, the operational interference between the two devices at position "5" is eliminated.

[0128] The second type of potential interference detection and adjustment process is Figure 5 Take this as an example. Figure 5 (a) is the time-space state diagram of the equipment before adjustment. During the grabbing operation at bay "7", a front crane crossing event occurred. The adjustment method is as follows: Figure 5 As shown in (b), by delaying the start time of the reach stacker placing the container task, the reach stacker can be prevented from crossing during the rail crane operation. It is worth noting that the adjusted plan has the reach stacker crossing during the heavy-load movement of the rail crane, which needs further adjustment.

[0129] The third type of potential interference detection and adjustment process is Figure 6 Take this as an example. Figure 6 (a) is the time-space state diagram of the equipment before adjustment. During the grabbing operation at bay "5", a heavy-load crossing event of the rail crane will occur. The adjustment method is as follows: Figure 6As shown in (b), by postponing the task of placing containers by the rail crane, the crossing of the heavy-loaded rail crane during the reach stacker operation was avoided. It is worth noting that the adjusted plan resulted in the crossing of the reach stacker during the heavy-loaded movement of the rail crane, which requires further adjustment.

[0130] The fourth type of potential interference adjustment is to adjust the constraints that the reach crane cannot cross during the heavy-load movement of the rail crane. Due to the different moving speeds of the rail crane and the reach crane, in addition to the interference of moving towards each other, there may also be interference of moving in the same direction. Taking the adjustment of the interference of moving towards each other as an example, the adjustment method is as follows: Figure 7 shown. Figure 7 (a) is the equipment time-space state diagram before adjustment. After the rail crane performs the grabbing operation, it needs to perform a heavy load movement from bay 5 to bay 0, during which a reach crane crossing event will occur. The adjustment method is shown in Figure 7(b). By delaying the start time of the rail crane's task of placing the container, the reach crane is controlled to cross the rail crane within the time period when the rail crane is waiting in place.

[0131] The potential interference of equipment is complex and diverse. During the interference adjustment process, the interference type will change, making it difficult to directly verify the feasibility of the scheduling plan and calculate the completion time of the batch task. To solve the above problems, a set of interference detection and adjustment algorithms is designed. The specific process is as follows: Figure 8 shown.

[0132] Specifically, the pseudo code of the interference detection and adjustment algorithm of this embodiment is as follows:

[0133] The first type of potential interference detection and adjustment time calculation:

[0134] Step 1: Input the device scheduling table Shadule_Table and the task pointer Index, and go to Step 2;

[0135] Step 2: Create the screening result table Result and go to Step 3;

[0136] Step 3: Get the task start time STime(Index), arrival time ATime(Index), operation time PTime(Index), task end time ETime(Index), end position EBay(Index) and task device Device(Index) of the benchmark task, set i=0, and go to Step 4;

[0137] Step 4: Determine whether the task in the i-th row is the last task in the schedule table. If so, go to Step 8, otherwise go to Step 5;

[0138] Step 5: Get the task start time STime(i), arrival time ATime(i), operation time PTime(i), task end time ETime(i), end location EBay(i) and task device Device(i) of the i-th task, and go to Step 6;

[0139] Step 6: If the time period [PTime(Index), ETime(Index)] overlaps with the time period [PTime(i), ETime(i)] and the operation position EBay(Index) == EBay(i) and Device(Index) != Device(i), then add the i-th row of tasks to Result and go to Step 7, otherwise go to Step 7;

[0140] Step 7: Let i=i+1, and go to Step 4;

[0141] Step 8: Determine whether the length of Result is 0. If so, go to Step 10. Otherwise, go to Step 9.

[0142] Step 9: Extract the first row of data in Result, obtain its index value j in Shadule_Table, and go to Step 10;

[0143] Step 10: Let the adjustment device = Device (j), the adjustment position = j, and the adjustment time = ETime (Index) + SafeTime - PTime (j), return the flag bit True, and output the adjustment device, adjustment position and adjustment time, and the program ends;

[0144] Step 11: Return the flag bit False and the program ends;

[0145] Second and third types of potential interference detection and adjustment time calculation:

[0146] Step 1: Input the device scheduling table Shadule_Table and the task pointer Index, and go to Step 2;

[0147] Step 2: Create the screening result table Result and go to Step 3;

[0148] Step 3: Get the operation time PTime(Index) of the benchmark task, the task end time ETime(Index), the end position EBay(Index), set i=0, and go to Step 4;

[0149] Step 4: Determine whether the task in the i-th row is the last task in the schedule table. If so, go to Step 7, otherwise go to Step 5;

[0150] Step 5: Get the task start time STime(i), arrival time ATime(i), start position SBay(i), end position EBay(i), task device Device(i), and task type Type(i) of the task in the i-th row, and go to Step 6;

[0151] Step 6: Let (STime(i), SBay(i)) be point A, (ATime(i), EBay(i)) be point B, (STime(Index), EBay(Index)) be point C, (ETime(Index), EBay(Index)) be point D, and go to Step 7;

[0152] Step 7: Determine whether Device (i) is a rail crane or a reach crane. If it is a rail crane, go to Step 8; otherwise, go to Step 9.

[0153] Step 8: If the line segment With line segment If they intersect and Type(i) is a placement task, add the task in row i to Result and go to Step 10; otherwise, go to Step 10;

[0154] Step 9: If the line segment With line segment If they intersect, add the task in row i to Result and go to Step 10; otherwise, go to Step 10;

[0155] Step 10: Let i = i + 1, and go to Step 4;

[0156] Step 11: Determine whether the length of Result is 0. If so, go to Step 15. Otherwise, go to Step 12.

[0157] Step 12: Extract the first row of data in Result, obtain its index value j in Shadule_Table, and go to Step 13;

[0158] Step 13: Let the adjustment device = Device(j), the adjustment position = j, use the point (STime(j), SBay(j)) and the point (ETime(j), EBay(j)) to construct the straight line equation f(x), and calculate point cross =f -1 (EBay(index)), adjustment time = ETime(index) + SafeTime-point cross , go to Step 14;

[0159] Step 14: Return the flag bit True, and output the adjustment device, adjustment position and adjustment time, and the program ends;

[0160] Step 15: Return the flag bit False and the program ends;

[0161] The fourth type of potential interference detection and adjustment time calculation:

[0162] Step 1: Input the device scheduling table Shadule_Table and the task pointer Index, and go to Step 2;

[0163] Step 2: Create the screening result table Result and go to Step 3;

[0164] Step 3: Get the task start time STime (Index), arrival time ATime (Index), operation time PTime (Index), task end time ETime (Index), start bay SBay (Index), end bay EBay (Index), task device Device (Index) and task type Type (Index) of the benchmark task, and go to Step 4;

[0165] Step 4: If Device (Index) is a rail crane and Type (Index) is a grab task, go to Step 22; otherwise, set i = 0 and go to Step 5;

[0166] Step 5: Determine whether the task in the i-th row is the last task in the device scheduling table. If so, go to Step 12, otherwise go to Step 6;

[0167] Step 6: Get the task start time STime(i), arrival time ATime(i), start position SBay(i), end position EBay(i), task device Device(i), and task type Type(i) of the task in the i-th row, and go to Step 7;

[0168] Step 7: Let (STime(i), SBay(i)) be point A, (ATime(i), EBay(i)) be point B, (STime(Index), SBay(Index)) be point C, (ETime(Index), EBay(Index)) be point D, and go to Step 8;

[0169] Step 8: Determine whether the Device (Index) is a rail crane or a reach crane. If it is a rail crane, go to Step 9; otherwise, go to Step 10.

[0170] Step 9: If Device (i) is a front loader and the line segment With line segment If they intersect, add the task in row i to Result and go to Step 11; otherwise, go to Step 11;

[0171] Step 10: If Device(i) is a rail crane, Type(i) is a placement task and the line segment With line segment If they intersect, add the task in row i to Result and go to Step 11; otherwise, go to Step 11;

[0172] Step 11: Let i=i+1, and go to Step 5;

[0173] Step 12: Determine whether the length of Result is 0. If so, go to Step 22. Otherwise, go to Step 13.

[0174] Step 13: Extract the first row of data in Result, obtain its index value j in Shadule_Table, set dt1 = 0, dt2 = 0, and go to Step 14;

[0175] Step 14: Note that (STime(j), SBay(j)) is point E, (ATime(j), EBay(j)) is point F, and go to Step 15;

[0176] Step 15: Determine the line segment With line segment Whether it crosses, if so, go to Step 16, otherwise go to Step 17;

[0177] Step 16: Let STime(j) = STime(j) + 1, ATime(j) = ATime(j + 1, dt1 = dt1 + 1, and go to Step 14;

[0178] Step 17: Restore STime(j) and ATime(j), and go to Step 18;

[0179] Step 18: Determine the line segment With line segment Whether it is crossed, if so, go to Step 19, otherwise go to Step 20;

[0180] Step19: Let STime(Index)=STime(Index)+1, ATime(Index)=ATime(Index)+1, dt2=dt2+1, go to Step18;

[0181] Step 20: Determine the size relationship between dt1 and dt2. If dt1 < dt2, then set the adjustment device = Device(j), the adjustment position = j, and the adjustment time = dt1; otherwise, set the adjustment device = Device(Index), the adjustment position = Index, and the adjustment time = dt2, and go to Step 21;

[0182] Step 21: Return the flag bit True, and output the adjustment device, adjustment position, and adjustment time, and the program ends;

[0183] Step 22: Return the flag bit False, and the program ends;

[0184] Based on the equipment conflict-free scheduling table and the initial yard status table, gradually deduce the yard status, detect the feasibility of the operation, and assign fitness values to the corresponding individuals according to the detection results. The process for judging the feasibility of the operation sequence is as follows:

[0185] Step 1: Input the equipment scheduling table Shadule_Table and the container position information table;

[0186] Step 2: Sort the scheduling table in ascending order according to the operation time;

[0187] Step 3: Determine whether the task in the i-th row is the last task in the equipment scheduling table. If so, go to Step 7; otherwise, go to Step 9;

[0188] Step 4: Determine whether Device(i) is a rail-mounted crane or a reach stacker. If it is a rail-mounted crane, go to Step 6; otherwise, go to Step 5;

[0189] Step 5: According to the container position information table, determine whether the equipment can reach the container operation position. If so, go to Step 6; otherwise, go to Step 8;

[0190] Step 6: According to the container position information table, determine whether the equipment has the operating conditions. If so, go to Step 7; otherwise, go to Step 8;

[0191] Step 7: Let i = i + 1, and go to Step 3;

[0192] Step 8: Output the fitness of the corresponding individual as the completion time of the last task in Shadule_Table + 100000000;

[0193] Step 9: Output the fitness of the corresponding individual as the completion time of the last task in Shadule_Table

[0194] In order to avoid the problem of loss of excellent genes that may occur in the genetic process, the improved genetic algorithm proposed in the present invention introduces an elite retention strategy based on the binary tournament selection operator, and directly copies some individuals with the best current fitness to the new population without crossover mutation operations, so as to prevent the best individuals in the population from being lost during the evolution process, resulting in the algorithm failing to converge to the global optimal solution.

[0195] The improved genetic algorithm proposed in this invention, in order to ensure the feasibility of the allocation scheme after the crossover of the offspring individuals, designs a synchronous sequential crossover operator based on the job sequence and task allocation based on the double-segment coding mode of the population individuals, such as Fig. 9 As shown in the figure, the specific process is as follows: for the chromosome operation order layer, select an individual from the selected parent individuals at random, copy the gene segment between the two randomly selected crossover points to the daughter chromosome, and record the index position. Delete the genes in the other parent individual that are repeated in the daughter chromosome, copy the remaining genes in sequence to the vacant positions of the daughter individual chromosome, and record the index position. Finally, according to the index position, perform the above operations on the chromosome operation allocation layer to obtain the daughter chromosome.

[0196] The improved genetic algorithm proposed in the present invention uses an adaptive mutation strategy to divide individuals in the population into high-quality individuals, low-quality individuals and infeasible individuals according to the average fitness of the population and the feasibility of the individual decoding plan. During the evolution process, the mutation probability of the individual is dynamically adjusted according to its fitness performance. The probability calculation is shown in Formula 35. Based on this strategy, a two-point exchange mutation operator is implemented in the evolution process of the population. The two-point exchange operation of this operator is only limited to the job sequence layer of the chromosome. The selection of the exchange position adopts a random strategy. When the two selected tasks belong to the same loading and unloading equipment, this operation realizes the disturbance of the job sequence of the loading and unloading equipment. If the two tasks do not belong to the same loading and unloading equipment, this operation realizes the change of the task allocation plan of the loading and unloading equipment. The specific operation is as follows Fig.10 shown.

[0197]

[0198] The improved genetic algorithm proposed in the present invention uses a heuristic population guidance operator to guide the individuals with the worst fitness in the population during the evolution process in order to improve the search efficiency of the population for high-quality solutions in the search process, so as to improve the overall search efficiency. The steps are as follows:

[0199] Fitness evaluation: Perform fitness evaluation on individuals in the population and select individuals with fitness in the bottom N% of the population, where N is a variable parameter.

[0200] Waiting time calculation: Decode the selected individuals, use the three-stage spatiotemporal planning fitness calculation method to obtain the corresponding equipment scheduling table, and use the equipment scheduling table to determine the waiting time of each task in the scheduling plan.

[0201] Identify and rearrange key tasks: Sort all tasks in descending order according to waiting time, take the task with the longest waiting time as the key task, and try to place it at the beginning or end of the task sequence to obtain alternative solutions. The specific process is as follows: Fig.11 shown.

[0202] Fitness re-evaluation: The adjusted sequence is re-evaluated to verify whether the position adjustment effectively improves the fitness of the individual.

[0203] Scheme comparison and selection: Compare the adjusted scheduling scheme with the original scheme, select the optimal scheme for encoding to generate new individuals, and replace the original individuals in the population.

[0204] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. A railway container terminal coordinated dispatch optimization method, characterized in that: The following steps are involved: Obtaining a device work list, wherein the device work list includes a job sequence and an allocation plan for each device; According to the equipment work form, obtaining the interfering equipment scheduling table through time-space state deduction; The potential interference of equipment is divided into the following types: the rail crane and the reach crane are loading and unloading at the same bay at the same time; when the rail crane is loading and unloading, the reach crane actively passes under the rail crane; when the reach crane is loading and unloading, the rail crane with heavy load actively passes over the reach crane; when the rail crane with heavy load is moving, the reach crane actively passes under the rail crane; In the interference equipment scheduling table, an interference detection and adjustment algorithm is used to detect and adjust the potential interference of various types of equipment, and finally a non-interference equipment scheduling table is obtained; According to the non-interference equipment scheduling table, the yard status is gradually deduced, the feasibility of the operation is tested, and the fitness of the corresponding individual is assigned according to the test result.

2. A railway container terminal coordinated dispatch optimization method according to claim 1, characterized in that: Ignoring the influence of human and natural factors, the starting and ending positions of the task are known, and the container does not need to overturn, the objective function shown in formula (1) is established; Establish the constraints shown in formula (2) to formula (17): in: i,j represent the container task number; N represents the total number of containers in the container task; k represents the loading and unloading equipment number, where equipment 0 is a reach stacker and equipment 1 is a rail stacker; M represents a sufficiently large positive number; represents the time when device k performs task i to grab the container; It indicates the time when device k performs task i to place the container; Indicates the box-picking operation time of device k; S represents the box placement operation time of device k; kt represents the location of device k at time t; Indicates the starting position of task i, representing stack position, shell position, and layer position respectively; Indicates the end position of task i, representing stack position, bay position, and layer position respectively; X, Y, and Z represent the number of stacks, bays, and layers in the yard area respectively; X1 and X2 represent special marked stack positions, where X1 is the train loading and unloading line closest to the container truck running line, and X2 is the container truck loading and unloading line; Indicates the moving distance of equipment k in the direction of the storage yard when performing task i; Indicates the moving distance of device k in the direction of the stacking position when executing task i; represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the direction of the yard bay position; represents the moving distance of equipment k from the end point of task i to the starting point of container task j along the stacking direction of the yard; h represents the safe distance that the reach stacker needs to maintain when the rail crane is in a heavy load or loading and unloading state; It represents the average moving speed of equipment k along the direction of the yard bay; represents the average moving speed of equipment k along the stacking direction; u ij,k ∈{0,1}: If the container task numbered i, j is assigned to the operating device k and the operating device k completes the container task numbered i and then performs the container task j, then u ij,k =1, otherwise u ij,k = 0. In addition, when tasks i and j are the same task, u ij,k =1 to 0; g xyz,t ∈0,1: If there are containers in slots x, y, z at time t, then g xyz,t =1, otherwise g xyz,t =0; Formula (1) is the objective function, which indicates that the completion time of the batch job is the shortest; Formula (2) indicates that the starting point of the task assigned to the reach stacker cannot be located below the X1 stack position; Formula (3) indicates that the end point of the task assigned to the reach stacker cannot be located below the X1 stack position; Formula (4) represents the time relationship between the time when the next task is grabbed by the track crane and the time when the previous task is placed; Formula (5) represents the time relationship between the grabbing time and the placing time of the current task of the rail crane; Formula (6) represents the time relationship between the grabbing time of the next task and the placing time of the previous task of the front crane; Formula (7) represents the time relationship between the grabbing time and the placing time of the current task of the reach stacker; Formula (8) indicates that each task is assigned to at most one device; Formula (9) indicates that each task has at most one subsequent task; Formula (10) indicates that each task has at most one immediate predecessor task; Formula (11) indicates that the reach stacker should prioritize tasks whose starting points are close to the container loading and unloading line, and then execute tasks whose starting points are far from the loading and unloading line; Formula (12) indicates that the reach stacker should prioritize tasks whose end points are far from the container loading and unloading line, and then execute tasks whose end points are close to the loading and unloading line; Formula (13) indicates that the loading and unloading equipment should first process the tasks whose starting points are located at the high level, and then process the tasks whose starting points are located at the low level; Formula (14) indicates that the loading and unloading equipment should first process the tasks whose endpoints are located at the lower level, and then process the tasks whose endpoints are located at the higher level; Formula (15) indicates that there is no container blocking the travel route of the reach stacker to grab the container; Formula (16) indicates that there is no container blocking the route of the reach stacker to place the container; Formula (17) indicates that during the loading and unloading operations and the movement of loads on the rail crane, the reach stacker should maintain a certain safety distance from it.

3. The method for optimizing coordinated dispatching of railway container terminals according to claim 1 is characterized in that: The adjustment algorithm adopts a dual-stage coding mode and consists of a three-stage spatiotemporal planning fitness calculation method, an elite retention strategy, a tournament selection operator, a synchronous sequential crossover operator, an adaptive two-point mutation operator and a heuristic population guidance operator.

4. A railway container terminal coordinated dispatch optimization method according to claim 3, characterized in that: The dual-segment coding mode adopts an integer coding mode and uses dual-segment gene expression in chromosome coding design. The genes in the first half of the chromosome are job sequence genes for expressing job sequence schemes; the genes in the second half of the chromosome are task allocation genes for expressing task allocation schemes.

5. The method for optimizing coordinated dispatching of railway container terminals according to claim 3 is characterized in that: The synchronous sequential crossover operator is based on the double-segment coding mode of the population individuals. For the operation order layer of the chromosome, an individual is randomly selected from the selected parent individuals, the gene segment between the two randomly selected crossover points is copied to the daughter chromosome, and the index position is recorded; Delete the genes in the other parent individual that are repeated in the offspring chromosome, copy the remaining genes to the vacant positions of the offspring individual chromosomes in sequence, and record the index positions; finally, perform the above operations on the chromosome assignment layer according to the index positions to obtain the offspring chromosomes.

6. A railway container terminal coordinated dispatch optimization method according to claim 3, characterized in that: The adaptive two-point mutation operator divides individuals in the population into high-quality individuals, low-quality individuals, and infeasible individuals according to the average fitness of the population and the feasibility of the individual decoding scheme. During the evolution process, the mutation probability of the individual is dynamically adjusted according to its fitness performance, as shown in Formula 18:

7. A railway container terminal coordinated dispatch optimization method according to claim 3, characterized in that: The heuristic population guidance operator has the following steps: Fitness evaluation: Perform fitness evaluation on individuals in the population and select individuals with fitness in the bottom N% of the population, where N is a variable parameter; Waiting time calculation: Decode the selected individuals, use the three-stage spatiotemporal planning fitness calculation method to obtain the corresponding equipment scheduling table, and use the equipment scheduling table to determine the waiting time of each task in the scheduling plan; Identify and rearrange key tasks: Sort all tasks in descending order according to waiting time, take the task with the longest waiting time as the key task, and try to place it at the beginning or end of the task sequence to come up with alternative solutions; Fitness re-evaluation: re-evaluate the fitness of the adjusted sequence to verify whether the position adjustment effectively improves the fitness of the individual; Scheme comparison and selection: Compare the adjusted scheduling scheme with the original scheme, select the optimal scheme for encoding to generate new individuals, and replace the original individuals in the population.

Images