Transportation and reassembly scheduling method for large equipment

Abstract

The application discloses a kind of large-scale equipment transportation and reassembly scheduling method, can effectively solve the coordination and overall arrangement problem between transportation, part disassembly reassembly and part packing in large-scale equipment transportation process.The application first uses the maximum minimum ant colony algorithm to find the way, proposes new pheromone initialization method, simultaneously using pheromone local update strategy makes its convergence speed faster, set pheromone upper and lower limit prevents premature local optimum, finally using neighborhood search optimization ensures that the planning of transportation path is better, provides optimal predicted transportation time for overall process arrangement;The application uses critical path algorithm to formulate disassembly, transportation in assembly process sequence and timetable to ensure that overall process time consumption is shortest, and provides time schedule for part packing and train arrangement, finally three-dimensional packing is carried out to part according to time schedule, determines specific packing scheme and train arrangement, and vehicle loading efficiency is higher.

Description

Technical Field

[0001] The present invention belongs to the technical field of large-scale equipment disassembly and assembly scheduling, and in particular relates to a transportation and reassembly scheduling method for large-scale equipment. Background Art

[0002] my country is a major construction powerhouse, with numerous projects underway in recent years. However, many of these projects require large, specialized equipment. The weight and volume of this equipment exceed the carrying capacity of standard trucks, as well as the weight and height limits of bridges and tunnels, making it difficult to transport directly by vehicle. This necessitates disassembling this equipment before transporting it to its intended location and then reassembling it on-site. This disassembly-transport-and-assembly model significantly increases the complexity and difficulty of transport scheduling, and also increases the difficulty of optimizing the overall process.

[0003] Currently, there are mature solutions for the path planning sub-problems and the packing problem in the disassembly, transportation, and reassembly of large equipment. Chinese patent application publication number CN114610045A provides a robot path planning method and system based on an improved ant colony algorithm. Its improved pheromone initialization method and pheromone update strategy accelerate convergence while improving path planning performance. Reference [Zhang Defu, Peng Yu, Zhu Wenxing, et al. Hybrid simulated annealing algorithm for solving the three-dimensional packing problem [J]. Chinese Journal of Computers, 2009, 032(011): 2147-2156. DOI: 10.3724 / SP.J.1016.2009.02147.] proposes an efficient hybrid simulated annealing algorithm for solving the three-dimensional packing problem and also provides a new composite block generation algorithm to improve packing efficiency.

[0004] However, these existing solutions lack effective coordination with the equipment disassembly and reassembly phases. They fail to consider the constraints imposed by the equipment disassembly sequence on the order in which parts are loaded onto trucks and the order in which the equipment installation team's vehicles depart. Consequently, the integration of these processes relies on manual experience and trial-and-error methods, lacking effective planning and scheduling mechanisms. This leads to inefficiency and instability in the transportation and reassembly processes, impacting project progress and cost control. Therefore, an innovative method for scheduling the transportation and reassembly of large equipment is needed to improve efficiency, reduce risk, and optimize the entire process. Summary of the Invention

[0005] In view of the above, the present invention provides a method for transporting and reassembling scheduling large equipment, which solves the coordination and planning problems among transportation, parts disassembly and reassembly, and parts packing during the transportation of large equipment. The transport route planning is faster, the vehicle loading efficiency is higher, and the overall process takes less time.

[0006] A method for transporting and reassembling large equipment comprises the following steps:

[0007] (1) Planning the vehicle transportation route, using the maximum minimum ant colony algorithm to find the path, and then using the neighborhood search algorithm to optimize the obtained path, determining the optimal transportation path and estimating the transportation time;

[0008] (2) Using the critical path algorithm to obtain the time schedule of disassembly and assembly, so as to determine the transportation order of the parts;

[0009] (3) Packing the parts according to the time schedule, determining the packing scheme of the parts and the arrangement of the loading vehicles.

[0010] Further, the specific process of the maximum minimum ant colony algorithm in step (1) is as follows:

[0011] 1.1 Initialize pheromone according to the starting point, end point and node connectivity matrix;

[0012] 1.2 Initialize the taboo list;

[0013] 1.3 Initialize the current ant;

[0014] 1.4 Create a new candidate node list according to the taboo list and node connectivity matrix;

[0015] 1.5 Calculate the path weight according to the pheromone concentration by the following formula:

[0016]

[0017] Wherein: P ij is the weight of the path (i,j), T ij is the pheromone concentration on the path (i,j), (i,j) represents the path from node i to node j, D ij is the direct distance between node i and node j, alpha is the parameter representing the importance of pheromone, beta is the parameter representing the importance of heuristic factor, and N is the node set;

[0018] 1.6 Calculate the probability of the ant moving from the current node to each candidate node according to the path weight by the following formula, and select the next node from the candidate list according to the probability:

[0019]

[0020] Wherein: s ij is the probability of the ant moving from node i to node j, and n is the number of nodes in the node set;

[0021] 1.7 Update the taboo list, current path and path length information;

[0022] 1.8 If the next node is not the end point, return to step 1.4;

[0023] 1.9 If the current ant number is less than the set ant number, return to step 1.3;

[0024] 1.10 Find the optimal path from the paths walked by the ants;

[0025] 1.11 Record the optimal path and its length information;

[0026] 1.12 Update the optimal path pheromone by the following formula:

[0027]

[0028] wherein T ij is the pheromone concentration on the path (i, j) after updating, δ is the pheromone release coefficient, and L is the length of the path (i, j);

[0029] 1.13 Calculate the pheromone evaporation;

[0030] 1.14 Adjust the pheromone of each connected path to be within the set upper and lower limits;

[0031] 1.15 If the current iteration number is less than the maximum iteration number, return to step 1.2;

[0032] 1.16 Select the optimal path and record the relevant information by synthesizing the results of each iteration.

[0033] Further, the specific process of initializing the pheromone in step 1.1 is as follows:

[0034] 1.1.1 Initialize the pheromone matrix, i.e., set the pheromone concentration of all non-connected paths to 0 and the pheromone concentration of connected paths to 1 according to the node connectivity matrix;

[0035] 1.1.2 Determine the vector a direction from the starting point to the ending point and calculate the determination of each path;

[0036] 1.1.3 Traverse the pheromone matrix, i.e., update the pheromone of each path according to the following formula:

[0037] T ij ′ = T ij + cos(θ),

[0038] wherein T ij is 1 if the path (i, j) is connected, T ij is 0 if the path (i, j) is not connected, and θ is the angle between the path (i, j) and the vector a, wherein θ takes 90° if the path (i, j) is not connected;

[0039] 1.1.4 Adjust pheromone of each communication path to be within the set upper and lower limit interval.

[0040] Further, the upper and lower limit interval of the pheromone is [min, max], min = 0.1, where D is the direct distance between the start point and the end point, and ε is the convergence coefficient.

[0041] Further, the neighborhood search algorithm in step (1) is specifically as follows:

[0042] Step 1: initialize each parameter including the current optimal solution, the number of iterations, and the maximum number of iterations;

[0043] Step 2: initialize the current node;

[0044] Step 3: traverse the initial solution and try to exchange the node with the current node;

[0045] Step 4: if the result after the exchange is better than the current optimal solution, update the current optimal solution, and then return to Step 2, and the number of iterations + 1;

[0046] Step 5: update the current node;

[0047] Step 6: if the current node is the end point and no node is exchanged, it means that the local optimal solution has been reached, and the optimization is directly exited;

[0048] Step 7: if the number of iterations reaches the maximum number of iterations, the optimization is exited.

[0049] Further, the critical path algorithm in step (2) is specifically as follows:

[0050] 2.1 Initialize each data, construct the constraint matrix of each step including transportation and the matrix of step time consumption, so as to draw the AOE (Activity on Edge Network) network;

[0051] 2.2 Initialize the earliest start time ve1 and the latest start time vl1 of the start node, both of which are set to 0;

[0052] 2.3 Traverse the AOE network from the start node in the forward direction, and calculate the earliest start time of each node according to the following formula:

[0053] ve i = Max(ve j +w ij ),

[0054] where ve i is the earliest start time of node i, w ijLet T(i,j) be the time spent on the activity from node i to node j, <i,j> be the arc from node i to node j, and T be the set of arcs pointing to node i.

[0055] 2.4 After the traversal, the length of the critical path vl is obtained as follows, and the latest start time is initialized as vl.

[0056] vl = Max(ve i ),

[0057] 2.5 The AOE network is traversed in reverse from the end node, and the latest start time of each node is calculated according to the following formula:

[0058] vl j = Min(vl i -w ij ),

[0059] wherein vl i is the latest start time of node i.

[0060] 2.6 The earliest start time and the latest start time of each activity step are calculated according to the following formula by using ve i and vl i of each node:

[0061] e vi = ve Evi

[0062] l vi = vl Lvi -w EviLvi

[0063] wherein e vi is the earliest start time of activity v i , l vi is the latest start time of activity v i , ve Evi is the earliest start time of the predecessor node of activity v i , vl Lvi is the latest start time of the successor node of activity v i , and w EviLvi represents the activity time of activity v i .

[0064] 2.7 Finally, the schedule is obtained according to the earliest start time and the latest start time of each activity step.

[0065] Further, the step (3) determines the loading train schedule of the parts according to the time schedule, and the specific implementation process is as follows:

[0066] A1. Traverse the transport time column of the time schedule to find the parts that are not transported at present and add them to the remaining untransported parts list List, and find the part with the earliest transport time in List and add it to the current untransported parts list cur_List, and record the time as el;

[0067] A2. Traverse the parts in cur_List in descending order of quality, and three-dimensionally pack the three-dimensional tree, and select the deepest leaf from the leaves of the three-dimensional tree as the current vehicle loading scheme;

[0068] A3. Take the current vehicle loading scheme as the tree root, traverse the parts in List with transport time earlier than el in descending order of quality, and three-dimensionally pack the three-dimensional tree, and select the deepest leaf from the leaves of the three-dimensional tree as the final loading scheme of the vehicle, at this time the vehicle is considered to be full, the number of cars car_num is increased by 1 and step A2 is returned;

[0069] A4. If cur_List is empty, take the current vehicle loading scheme as the tree root, traverse the parts in List with transport time earlier than el in descending order of quality, and three-dimensionally pack the three-dimensional tree, and select the deepest leaf from the leaves of the three-dimensional tree as the final loading scheme of the vehicle, at this time the vehicle is considered to be full, the number of cars car_num is increased by 1 and step A1 is returned;

[0070] A5. If List is empty, output the new time schedule and the specific loading car arrangement of the parts.

[0071] Further, the step (3) determines the packing scheme of the parts according to the transport order through three-dimensional packing, and the specific implementation process is as follows:

[0072] B1. The parts are regarded as these cuboids, and the cuboids have three different faces, which are A face, B face and C face;

[0073] B2. The left rear lower vertex of the cuboid box is taken as the coordinate origin from the perspective of the observer, and the length, width and height of the box are taken as the X, Y and Z coordinate axes of the three-dimensional space to establish a coordinate system;

[0074] B3. The coordinate origin is taken as the root node, and the X, Y and Z coordinate axes are taken as the three branch nodes of the root node to construct a packing scheme ternary tree, and the root node of the ternary tree corresponds to the initial state of the empty box. In addition to the three branch nodes and the root node, each node on the ternary tree corresponds to a packing method;

[0075] B4. Initialize the packing scheme, and divide the packing method into the following three packing schemes:

[0076] ① the flat-laying scheme is composed of two kinds of packing modes in which the Z axis is perpendicular to the A face, and the flat-laying scheme corresponds to the left sub-tree of the packing scheme ternary tree;

[0077] ② the side-laying scheme is composed of two kinds of packing modes in which the Z axis is perpendicular to the B face, and the side-laying scheme corresponds to the middle sub-tree of the packing scheme ternary tree;

[0078] ③ the vertical-laying scheme is composed of two kinds of packing modes in which the Z axis is perpendicular to the C face, and the vertical-laying scheme corresponds to the right sub-tree of the packing scheme ternary tree;

[0079] B5. all the packing schemes are traversed according to the two kinds of packing modes of each packing scheme until all the leaf nodes of the packing scheme ternary tree are generated; after the packing modes corresponding to the leaf nodes are executed, no packing mode can be executed any more;

[0080] B6. the process node numbers between each leaf node and the root node are calculated, and the packing scheme formed by the leaf node branch with the largest process node number is the optimal packing scheme.

[0081] The scheduling method effectively solves the coordination and overall arrangement problem among transportation, part disassembly and reassembly and part packing in the large equipment transportation process; the scheduling method firstly uses the max-min ant colony algorithm to search a path, proposes a new pheromone initialization method, simultaneously adopts a pheromone local updating strategy to make the convergence speed faster, sets the upper and lower bounds of the pheromone to prevent falling into a local optimum too early, finally uses a neighborhood search optimization to ensure that the transportation path planning is optimal, and provides an optimal predicted transportation time for overall process arrangement; the scheduling method uses a critical path algorithm to determine the sequence and time table of disassembly, transportation and assembly process to ensure that the overall process time is the shortest, and provides a time table for part packing and train arrangement, finally packs the parts according to the time table, determines the specific packing scheme and train arrangement, and the vehicle loading efficiency is higher. BRIEF DESCRIPTION OF DRAWINGS

[0082] Figure 1 It is a flow relationship schematic diagram of the scheduling method and the system thereof.

[0083] Figure 2 It is a neighborhood search schematic diagram.

[0084] Figure 3 It is a transportation road network schematic diagram, in which the star-shaped points are transportation starting points and ending points, the cross points are transportation nodes, and the lines are optional transportation paths, and the length of the line represents the length of the equivalent transportation time.

[0085] Figure 4The AOE network diagram for disassembling and transporting assembly activities, wherein, except for the screw machine and the connecting bridge, hoisting and disassembling refer to disassembling and hoisting away, the disassembling and hoisting away of the screw machine and the connecting bridge are separated instead of being continuously performed; except for the screw machine, hoisting refers to lowering into the well and assembling, and the lowering into the well and assembling of the screw machine are separated.

[0086] Figure 5 The figure is a schematic diagram of a transportation route, and the red line in the figure is a transportation path. DETAILED DESCRIPTION

[0087] In order to more specifically describe the present application, the technical solutions of the present application are described in detail below in combination with the drawings and specific embodiments.

[0088] The large equipment transportation and reassembly scheduling method proposed in the present application is implemented as shown in Figure 1 The scheduling system designs three modules to respectively solve the three main problems of transportation route planning, hoisting and transporting time planning and part containerizing train arrangement, the three modules are independent in work but interconnected in data, and mutual coordination between the modules and processes is realized.

[0089] The vehicle transportation route planning module uses the max-min ant colony algorithm to search for a path, the obtained path is optimized using neighborhood search, the optimal transportation path is determined, and the transportation time is estimated, and the specific process is as follows:

[0090] Step 1-1 uses the max-min ant colony algorithm to search for a path;

[0091] Step 1-1-1: initialize pheromone according to the start point and the end point and the node connectivity matrix, assign 0 to all non-connected paths and 1 to connected paths, and then determine the direction of vector a from the start point to the end point; traverse the pheromone matrix, and update the pheromone of each connected path as follows: T ij = T ij + cos(θ), θ is the included angle between the connected path and the vector a from the start point to the end point, if the path is not connected, θ is 90°, finally adjust the pheromone of each connected path to be not less than min and not greater than max, min = 0.1,

[0092] Step 1-1-2: initialize the taboo list;

[0093] Step 1-1-3: initialize the current ant;

[0094] Step 1-1-4: create a new candidate node list according to the taboo list and the node connectivity matrix;

[0095] Step 1-1-5: calculate the path weight according to the path pheromone concentration

[0096] Step 1-1-6: Calculate the probability of transferring from the current node to each candidate node according to the path weight, and select the next node from the candidate list according to the probability

[0097] Step 1-1-7: Update the taboo list, current path, path length, and other information;

[0098] Step 1-1-8: If the node is not the end point, return to Step 1-1-4;

[0099] Step 1-1-9: If the number of ants is less than the set number of ants, return to Step 1-1-3;

[0100] Step 1-1-10: Find the optimal path from the m paths walked by the m ants;

[0101] Step 1-1-11: Record the optimal path and its length information;

[0102] Step 1-1-12: Update the pheromone of the optimal path;

[0103] Using the max-min ant colony algorithm, only update the pheromone brought by the optimal path among all ants in this generation to speed up the convergence rate, the update formula is:

[0104] Step 1-1-13: Calculate the evaporation of pheromone, the calculation formula is: ij T ij = γ * T

[0105] Step 1-1-14: Adjust the pheromone of each connected path to be not less than min and not greater than max, min = 0.1,

[0106] Step 1-1-15: If the current iteration number is less than the maximum iteration number, return to Step 1-1-2;

[0107] Step 1-1-16: Combine the results of each iteration to select the optimal path and record its information.

[0108] Step 1-2 neighborhood search algorithm, as shown in Figure 2

[0109] Step 1-2-1: Initialize the current optimal solution, iteration number, maximum iteration number, and other parameters;

[0110] Step 1-2-2: Initialize the current node;​

[0111] Step1-2-3: traverse the preliminary solution, try to exchange the node with the current node;

[0112] Step1-2-4: if the result after exchange is better than the current optimal solution, update the current optimal solution, then return to Step2, iteration number +1;

[0113] Step1-2-5: update the current node;

[0114] Step1-2-6: if the current node is the end point, no node is exchanged, it means that the local optimal solution has been reached, then directly exit the optimization;

[0115] Step1-2-7: if the iteration number reaches the maximum iteration number, exit the optimization.

[0116] The parts loading and unloading arrangement module uses the critical path algorithm to obtain the disassembly and assembly time arrangement, so as to determine the transportation order of the parts, the specific process is as follows:

[0117] Step2-1: initialize each data, construct the constraint matrix of each step including transportation, and the time consumption matrix of each step, so as to draw the AOE network;

[0118] Step2-2: initialize ve1 and vl1 of the starting node, both are 0;

[0119] Step2-3: traverse the AOE network from the starting node, calculate the earliest start time of each node: ve i =Max(ve j +w ij ),

[0120] Step2-4: after traversing, the length of the critical path is obtained: vl=Max(ve i ), and initialize the latest start time of the end node as vl;

[0121] Step2-5: traverse the AOE network from the end node, calculate the latest start time of each node: vl j =Min(vl i -w ij ),

[0122] Step2-6: calculate e i and l i of each activity process from ve i and vl i of each node: e vi =veEvi , l vi = vl Lvi -w EviLvi ;

[0123] Step2-7: Finally, according to the earliest and latest occurrence time of each activity step, the time table is obtained.

[0124] The part packing module packs the parts according to the time table, determines the specific packing scheme and train arrangement, and the specific process is as follows:

[0125] Step3-1: Traverse the transportation time column of the part time table, find out the parts that have not been transported and add them to the List, then find out the part with the latest transportation time and the earliest part, add it to the cur_List, and record the time as el;

[0126] Step3-2: Traverse the parts in the cur_List in descending order of part quality, and pack them in the three-dimensional packing of the ternary tree, select the deepest leaf from the leaves of the ternary tree as the current vehicle loading scheme;

[0127] Step3-2-1: Obtain the size of the part, and put the part into the smallest cuboid that can hold the part, and consider the part as these cuboids; the cuboid has three different faces, denoted as A face, B face and C face;

[0128] Step3-2-2: From the perspective of the observer, the lower left corner of the box with the shape of the cuboid is set as the coordinate origin, and the length, width and height of the box are set as the X, Y and Z coordinate axes of the three-dimensional space, and the coordinate system is established;

[0129] Step3-2-3: Construct a packing scheme ternary tree with the coordinate origin as the root node and the X, Y and Z coordinate axes as the three branch nodes of the root node; the root node of the packing scheme ternary tree corresponds to the initial state of the empty box, and each node on the packing scheme ternary tree corresponds to a packing method except for the three branch nodes corresponding to the X, Y and Z coordinate axes and the root node;

[0130] Step3-2-4: Initialize the packing scheme, and divide the packing method into three packing schemes:

[0131] (1) The two schemes that make the Z axis perpendicular to the A face constitute the flat laying scheme, and the flat laying scheme corresponds to the left subtree of the packing scheme ternary tree;

[0132] (2) The two schemes that make the Z axis perpendicular to the B face constitute the side laying scheme, and the side laying scheme corresponds to the middle subtree of the packing scheme ternary tree;

[0133] (3) The two schemes that make the Z axis perpendicular to the C face constitute the vertical laying scheme, and the vertical laying scheme corresponds to the right subtree of the packing scheme ternary tree;

[0134] Step3-2-5: According to two kinds of packing ways of each packing scheme, all packing schemes are traversed until all leaf nodes of the packing scheme ternary tree are generated; after the packing way corresponding to the leaf node is executed, no other packing way can be executed;

[0135] Step3-2-6: Calculate the number of process nodes between each leaf node and the root node, and the leaf node branch with the most process nodes constitutes the best packing scheme.

[0136] Step3-3: The current vehicle loading scheme is the tree root, and the List is traversed in descending order of the mass of the part with the earliest transportation time earlier than el, and the ternary tree three-dimensional packing is selected from the leaves of the ternary tree to select the deepest leaf as the final loading scheme of the vehicle. At this time, the vehicle is considered to be full, car_num = car_num + 1 and return to Step3-2 (if there is a vehicle quantity limit and it is exceeded, the transportation schedule of the remaining parts needs to be updated according to the vehicle round trip time. We did not consider the vehicle quantity limit when making the schedule, but considered that all transportation can be carried out at the same time);

[0137] Step3-4: If the cur_List list is empty, try the current vehicle loading scheme as the tree root, and traverse the List in descending order of the mass of the part with the earliest transportation time earlier than el, and the ternary tree three-dimensional packing is selected from the leaves of the ternary tree to select the deepest leaf as the final loading scheme of the vehicle. At this time, the vehicle is considered to be full, car_num = car_num + 1 and return to Step3-1;

[0138] Step3-5: If the List list is empty, output a new schedule and the specific part loading arrangement.

[0139] Next, we take the disassembly and transportation assembly process of a shield machine as an example. The transportation network is shown in Figure 3 , which is relatively sparse and not necessarily connected between nodes, so it cannot be directly transported from the starting point to the destination. The overall process AOE network is shown in Figure 4 , and the overall workflow mainly includes disassembly, transportation and assembly. The disassembly and assembly processes are relatively linear, while the transportation part is more flexible due to the downhole and installation sequence. The overall volume and mass of the shield machine are large, and it cannot be directly transported, but it can be transported after being disassembled. However, even after disassembly, the individual parts are still very heavy and need to be transported by heavy trucks. The specifications of the main components are shown in Table 1:

[0140] Table 1

[0141]

[0142]

[0143] The vehicle transportation route planning module determines the transportation path and expected time according to the transportation starting point and ending point and the road network, uses the max-min ant colony algorithm for path finding, and obtains a preliminary solution: first, generate initial pheromone, set the pheromone concentration of each connected path to 1 first, and set the non-connected path to 0, then add the cosine of the angle between the vector of the path and the vector from the current node to the destination, which is positive when the two vectors are in the same hemisphere, and negative when they are not in the same hemisphere (at this time, the pheromone concentration of the same path back and forth is different), finally check the pheromone boundary, obtain the initial pheromone concentration as shown in Table 2:

[0144] Table 2

[0145] 0 0 0 1.735851 0 0 0 0 1.393558 0 1.423986 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.765619 0 0 0 0 0 1.890658 0 0 0 0 0 0 0 0 0 0 1.668783 0 0 0 0 0.966383 1.558506 0 0 0 0.475923 0 0 0 0 0 0.343795 0 0 0 0 0 0 0.268909 0 0 0 0 1.868691 0 0 0 0 0.307042 0 0 0 0 0.530385 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.669848 0 0 0 0 0 0 0 1.995827 0 1.958587 0 0 0 0 0 0 0 0 1.94531 0 0 0 0 0 0 0 0.724605 0 0 0 0 0 0 0 0 0 0 0 0 0.975233 0 0 0 0 0 0 0 0 1404857 0 0 0 0 0 0 0 1576612 0 0 0 0 0 0 0 0 2 1.239123 0 0.629201 0 0 0 0 0 0 0.981339 0 1.146687 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.5023 0 0 0 0 0 0 0 0 0 0 0 0 1.790912 0 0 0 1.261166 0 0 0 0 0 0 0 0 0 0 0 0 0 1.990314 0 0 0 0 0.660397 0 1.289414 0 0 0 0 0 0 0 0 0 0 0 0 0 1.739987 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.954735 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.934245 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0211765 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

[0146] According to the initialized pheromone, the max-min ant colony algorithm is used to obtain a preliminary solution, the parameters are set to Alpha = 1, Beta = 1, Gamma = 0.75, the maximum number of iterations iter_max = 25, Delta = 10, the number of ant colonies m = 20, and the output result after iteration is shown in Figure 5 1->11->10->20.

[0147] The neighborhood search is used for optimization, and the solution obtained by the max-min ant colony algorithm is a local optimal solution and does not need to be further modified through the neighborhood search algorithm verification, so the optimal transportation path is determined: 1->11->10->20, and the expected transportation time is 9.7554, which is rounded up to 10.

[0148] Then call the part loading and unloading scheduling module, input the AOE network, and obtain the expected transportation time of the parts from the vehicle transportation route planning module, and update the AOE network.

[0149] The critical path algorithm is used to obtain the time schedule of disassembly and assembly, and the earliest and latest time schedule of each node ve, vl is calculated, as shown in Table 3:

[0150] Table 3

[0151]

[0152]

[0153] The earliest and latest occurrence time schedule of each activity e, l is calculated, as shown in Table 4:

[0154] Table 4

[0155]

[0156]

[0157] We found that due to the disassembly and assembly steps are more linear, the disassembly and assembly of each part is the key path, and in the transportation of parts, only the transportation of No. 5 trolley is the key path, and the rest of the transportation activities have a certain time arrangement space.

[0158] The part packing module first reads the part loading and unloading schedule obtained by the part loading and unloading scheduling module, and focuses on extracting the earliest start time e and the latest look time l of the part transportation time arrangement to perform three-dimensional packing:

[0159] First of all, we will normalize all the parts into cuboids, and the disc-shaped cutter, front shield, middle shield, and shield tail will be converted into a cuboid with diameter x diameter x height. The width limit for heavy goods transported on the first-class highway is 3.5m to 4.5m, the length limit is 14m to 20m, the height limit is greater than or equal to 3m and less than 3.8m, and the weight limit is greater than or equal to 20t and less than 100t. Based on the above, the vehicle volume limit is set to 20m in length, 4.5m in width, and 3.8m in height, and the load limit is 100t.

[0160] Then constantly find the latest transportation time of the parts that have not been transported, and load the parts from the earliest transportation time. Choose parts from the parts with the smallest transportation time to fill the vehicle as much as possible. Each loading of parts will form three sub-trees of vertical, horizontal and vertical placement. Each sub-tree will form 1 to 2 different placement schemes due to its specific shape. Pruning is needed based on whether it exceeds the vehicle loading constraints. If it is not due to the combination of parts placement but due to the size and weight of a single part exceeding the truck standard, only the single part that exceeds the standard is transported and a minimum placement scheme is found. Then continue to place other parts on each unpruned scheme to generate new schemes.

[0161] Since the use of the shield tunneling machine is usually in a narrow area on the ground, the transportation order is arranged according to the latest departure time to minimize the space occupation burden on the ground at the destination. The specific train arrangement is shown in Table 5:

[0162] Table 5

[0163]

[0164] The above description of the embodiments is for the purpose of enabling one of ordinary skill in the art to make and use the application and is not intended to limit the application as construed in the broadest scope possible. Inasmuch as modifications to the above described embodiments can readily be made by persons of ordinary skill in the art, it is intended that the application not be limited to the embodiments described above but should be construed in the broadest scope possible.

Claims

1. A method for transporting and reassembling large equipment, comprising the following steps: (1) Plan the vehicle transportation route, use the maximum-minimum ant colony algorithm to find the path, and then use the neighborhood search algorithm to optimize the obtained path to determine the optimal transportation route and estimate the transportation time; (2) Use the critical path algorithm to obtain a time schedule for disassembly and assembly to determine the order in which the parts are transported; (3) Carry out three-dimensional packing of parts according to the time schedule, and determine the loading schedule of parts according to the time schedule. The specific implementation process is as follows: A1. Search the transportation time column of the schedule table, find the parts that have not yet been transported, and add them to the remaining untransported parts list. Then, find the part with the latest transportation time and the earliest transportation time in the list, add it to the current list of parts to be transported, cur_List, and record the time as el. A2. Traverse the parts in cur_List in descending order of quality, pack them into a three-dimensional ternary tree, and select the deepest leaf from the ternary tree as the current vehicle loading solution. A3. Using the current vehicle loading plan as the root, traverse the list of parts with the earliest transport time earlier than el in descending order of quality, pack them into a ternary tree, and select the deepest leaf from the leaves of the ternary tree as the final loading plan for the vehicle. At this point, the vehicle is considered fully loaded, the number of cars, car_num, is incremented by 1, and the process returns to step A2. A4. If cur_List is empty, try the current vehicle loading plan as the root. Traverse the list of parts with an earlier shipping time than el in descending order of quality, pack them into a ternary tree, and select the deepest leaf from the leaves of the ternary tree as the final loading plan for the vehicle. At this point, the vehicle is considered fully loaded, the number of cars car_num is incremented by 1, and the process returns to step A1. A5. If the List is empty, output the new schedule and the specific loading schedule for the parts. The packing plan for parts is determined based on the transportation sequence through 3D packing. The specific implementation process is as follows: B1. Think of the parts as rectangular parallelepipeds with three different faces, labeled A, B, and C. B2. From the observer's perspective, set the lower left vertex of the rectangular box as the coordinate origin. Set the length, width, and height of the box as the X, Y, and Z coordinate axes of three-dimensional space to establish a coordinate system. B3. Construct a ternary tree of packing solutions with the origin as the root node and three branch nodes rooted at the X, Y, and Z axes. The root node of the ternary tree corresponds to the initial empty box state. Except for the three branch nodes and the root node, each node in the ternary tree corresponds to a packing solution. B4. Initialize the packing scheme and classify the packing methods into the following three types: ① The two packing methods that make the Z axis perpendicular to the A surface constitute a flat placement solution, which corresponds to the left subtree of the packing solution ternary tree; ② The side-laying scheme is formed by two packing methods that make the Z axis perpendicular to the B surface. This side-laying scheme corresponds to the neutron tree of the packing scheme ternary tree; ③ The two packing methods that make the Z axis perpendicular to the C surface form a vertical placement scheme, which corresponds to the right subtree of the packing scheme ternary tree; B5. Traverse all packing solutions according to the two packing methods for each packing solution until all leaf nodes of the packing solution ternary tree are generated. After the packing method corresponding to a leaf node is executed, no further packing method can be executed. B6. Calculate the number of process nodes between each leaf node and the root node. The packing solution consisting of the leaf node branch with the largest number of process nodes is the optimal packing solution.

2. The method for transport and reassembly scheduling according to claim 1, characterized in that: The specific process of the maximum-minimum ant colony algorithm in step (1) is as follows: 1.1 Initialize pheromone according to the starting point, end point and node connectivity matrix; 1.2 Initialize the taboo table; 1.3 Initialize the current ant; 1.4 Create a new list of candidate nodes based on the taboo table and node connectivity matrix; 1.5 Calculate the path weight according to the pheromone concentration using the following formula; in: P ij is the weight of path (i, j), T ij is the pheromone concentration on the path (i, j), (i, j) represents the path from node i to node j, D ij is the direct distance between node i and node j, α is the parameter representing the importance of pheromone, β is the parameter representing the importance of heuristic factor, and N is the node set; 1.6 Calculate the probability of an ant moving from the current node to each candidate node using the following formula based on the path weight, and select the next node from the candidate list based on the probability; Where: s ij is the probability of an ant transferring from node i to node j, and n is the number of nodes in the node set; 1.7 Update the taboo table, current path, and path length information; 1.8 If the next node is not the end point, return to step 1.4; 1.9 If the current ant number is less than the set number of ants, return to step 1.3; 1.10 Find the best path from multiple paths taken by multiple ants; 1.11 Record the optimal path and its length information; 1.12 Update the optimal path pheromone using the following formula; Where: T ij ′ is the pheromone concentration on the updated path (i, j), δ is the pheromone release coefficient, and L is the length of the path (i, j); 1.13 Calculate pheromone evaporation; 1.14 Adjust the pheromone content of each connected path so that it is within the set upper and lower limits; 1.15 If the current number of iterations is less than the maximum number of iterations, return to step 1.2; 1.16 Based on the results of each iteration, select the optimal path and record relevant information.

3. The method for transport and reassembly scheduling according to claim 2, wherein: The specific process of initializing pheromones in step 1.1 is as follows: 1.1.1 Initialize the pheromone matrix, that is, according to the node connectivity matrix, set the pheromone concentration of all disconnected paths to 0 and the pheromone concentration of connected paths to 1; 1.1.2 Determine the direction of vector a from the starting point to the end point, and calculate and determine each path; 1.1.3 Traverse the pheromone matrix, that is, update the pheromone of each path according to the following formula: Where: If the path (i, j) is connected, then T ij is 1, if the path (i, j) is not connected, then T ij is 0, θ is the angle between path (i, j) and vector a. If path (i, j) is not connected, θ is 90°; 1.1.4 Adjust the pheromone content of each connected path so that it is within the set upper and lower limits.

4. The method for transport and reassembly scheduling according to claim 2, wherein: The upper and lower limits of the pheromone are [min, max], min = 0.1, Where D is the direct distance between the starting point and the end point, and ε is the convergence coefficient.

5. The method for transport and reassembly scheduling according to claim 1, wherein: The specific process of the neighborhood search algorithm in step (1) is as follows: Step 1: Initialize various parameters including the current optimal solution, number of iterations, and maximum number of iterations; Step 2: Initialize the current node; Step 3: Traverse the preliminary solution and try to exchange the node with the current node; Step 4: If the result after the exchange is better than the current optimal solution, update the current optimal solution and then return to Step 2 with the number of iterations + 1; Step 5: Update the current node; Step 6: If the current node is the end point and no node is exchanged, it means that the local optimal solution has been reached and the optimization is exited directly; Step 7: If the number of iterations reaches the maximum number of iterations, exit the optimization.

6. The method for transport and reassembly scheduling according to claim 1, characterized in that: The specific process of the critical path algorithm in step (2) is as follows: 2.1 Initialize various data, construct the constraint matrix of each step including transportation and the matrix of step time consumption, and draw the AOE network; 2.2 Initialize the earliest start time ve1 and the latest start time vl1 of the starting node, both set to 0; 2.3 Start from the starting node and traverse the AOE network forward. Calculate the earliest start time of each node according to the following formula: in: ve i is the earliest start time of node i, w ij is the time taken for the activity from node i to node j,<i,j> represents the arc from node i to node j, and T is the set of arcs pointing to node i; 2.4 After the traversal is completed, the length vl of the critical path is obtained as follows, and the latest start time is initialized to vl; Where: N is the node set; 2.5 Start from the end node and traverse the AOE network in reverse order. Calculate the latest start time of each node according to the following formula: Among them: vl i is the latest start time of node i, vl j is the latest start time of node j; 2.6 The ve of each node i and vl i The earliest and latest start times of each activity process are calculated using the following formula: to vi =and Evi l vi =vl Lvi -w EviLvi Among them: e vi For activity v i The earliest start time, l vi For activity v i The latest start time, ve Evi For activity v i The earliest start time of the prerequisite node, vl Lvi For activity v i The latest start time of the successor node, w EviLvi Indicates activity v i activities that consume time; 2.7 Finally, a schedule is drawn up based on the earliest and latest start times of each activity step.

Images