Maintenance and scheduling method and device for armored equipment

Abstract

The present disclosure provides a maintenance scheduling method and apparatus for armored equipment, relating to the technical field of maintenance scheduling. The method comprises: determining all current maintenance tasks after the armored equipment is damaged; assigning corresponding weights to each current maintenance task using a hypergraph generative adversarial network; determining the maintenance time corresponding to each current maintenance task based on each current maintenance task and the corresponding weight, and determining the weighted completion time of all current maintenance tasks; and employing a flood optimization algorithm to optimize the minimization of the weighted completion time of all current maintenance tasks as an optimization objective, thereby determining an optimal maintenance sequence for all current maintenance tasks after optimization. By determining the optimal maintenance sequence, the present disclosure minimizes the weighted completion time required for maintenance, thereby improving the efficiency of armored equipment maintenance scheduling and efficiently implementing post-maintenance scheduling for armored equipment, thereby resolving the problem in related technologies of being unable to implement efficient maintenance scheduling for armored equipment.

Description

Technical Field

[0001] The present disclosure relates to the technical field of maintenance scheduling, and in particular to a maintenance scheduling method and device for armored equipment. Background Art

[0002] In the context of economic globalization, armored equipment plays a vital role in the rapid flow of materials. With the increase in armored equipment throughput and the proliferation of mechanized operations, armored equipment is often operating in an overloaded state. Furthermore, mechanized operations conducted outdoors are susceptible to inclement weather, leading to armored equipment failures. Therefore, effective maintenance and scheduling of armored equipment is crucial to maintaining its efficient operation. However, existing technologies have been unable to achieve efficient maintenance and scheduling for armored equipment.

[0003] With regard to the problem that related technologies cannot achieve efficient maintenance and scheduling of armored equipment, no effective technical solution has been proposed so far. Summary of the Invention

[0004] The main purpose of the present disclosure is to provide a maintenance scheduling method and device for armored equipment to solve the problem in related technologies that armored equipment cannot be efficiently maintained and scheduled.

[0005] To achieve the above objectives, the first aspect of the present disclosure provides a maintenance scheduling method for armored equipment, comprising:

[0006] Identify all current repair tasks following loss of armored equipment;

[0007] Use hypergraph generative adversarial networks to assign corresponding weights to each current maintenance task;

[0008] Based on each current maintenance task and the corresponding weight, determine the maintenance time corresponding to each current maintenance task, and determine the weighted completion time of all current maintenance tasks;

[0009] The flood optimization algorithm is used to minimize the weighted completion time of all current maintenance tasks as the optimization goal, and the optimal maintenance sequence of all current maintenance tasks after optimization is determined.

[0010] Optionally, a hypergraph generative adversarial network is used to assign corresponding weights to each current maintenance task, including:

[0011] Initialize the hypergraph generative adversarial network, use the steepest descent method with momentum to preliminarily train the hypergraph generative adversarial network in a parallel training mode, and obtain the hypergraph generative adversarial network after the preliminary training. The hypergraph generative adversarial network uses the sigmoid function as the transfer function;

[0012] The maintenance workload ratio and maintenance cycle of each historical maintenance task are input into the hypergraph generative adversarial network after preliminary training, and the results are output. The error is determined based on the results of two consecutive outputs. The hidden layer weights and thresholds of the hypergraph generative adversarial network after preliminary training are corrected based on the error, and the training is repeated.

[0013] When the training is repeated until the sum of squares of the errors is less than or equal to the preset error tolerance, or when the number of repeated training reaches the preset maximum number of iterations, the training is stopped to obtain a hypergraph generative adversarial network after repeated training;

[0014] By analyzing and comparing the maintenance workload ratio and maintenance cycle of each historical maintenance task, the accuracy of the hypergraph generative adversarial network after repeated training is determined. The hypergraph generative adversarial network after repeated training is tested to obtain the weight corresponding to each current maintenance task of the armored equipment.

[0015] According to the normal maintenance state or emergency maintenance state of the armored equipment, the current maintenance task of the armored equipment and the change of the corresponding weight of the current maintenance task are determined.

[0016] Furthermore, the maintenance workload ratio and maintenance cycle of each historical maintenance task are input into the hypergraph generative adversarial network after preliminary training, including:

[0017] Normalize the maintenance workload ratio and maintenance cycle of each historical maintenance task to obtain normalized historical maintenance data. Input the normalized historical maintenance data into the hypergraph generative adversarial network after preliminary training.

[0018] After obtaining the hypergraph generative adversarial network after repeated training, the method further includes:

[0019] The normalized historical maintenance data is denormalized to obtain the maintenance workload ratio and maintenance cycle of each historical maintenance task.

[0020] Optionally, the hypergraph generative adversarial network is an interactive hyperedge neuron module or MRL-AHF.

[0021] Furthermore, when the hypergraph generative adversarial network is an interactive hyperedge neuron module, the interactive hyperedge neuron module is used as a generator to capture the complex relationship between data, and the discriminator is an MLP;

[0022] Determine the characteristics of the nodes in the l+1 layer according to the following formula and the characteristics of the hyperedge at layer l+1 :

[0023] ,

[0024] in, is the activation function, is the hypergraph incidence matrix, yes The transpose of is the feature of the hyperedge at layer l, is the feature of the node in layer l, is the weight matrix of the hyperedge at layer l, is the weight matrix of the nodes in layer l, is a hyperparameter;

[0025] The nodes in the last layer are used as the weights corresponding to the current maintenance task.

[0026] Furthermore, when the hypergraph generative adversarial network is MRL-AHF, the potential representation is fed into the encoders EA and EB respectively to obtain the representation and , and the two generated hypergraphs use the incidence matrix and Indicates that hypergraph fusion is achieved through adversarial training strategy to obtain vertex features :

[0027] ,

[0028] in, and They are and The vertex degree matrix of , , and They are and The hyperedge degree matrix of yes The transpose of is the weight matrix;

[0029] The features of the vertices are used as the weights corresponding to the current maintenance task.

[0030] Optionally, determining a current maintenance task of the armored equipment and a change in a weight corresponding to the current maintenance task according to the normal maintenance state or the emergency maintenance state of the armored equipment includes:

[0031] When the armored equipment is in normal maintenance status, the first objective function is established according to the following formula:

[0032] ,

[0033] The first constraint is:

[0034] ,

[0035] Among them, the first objective function is the weighted completion time of all current maintenance tasks under normal maintenance status, n is the total number of current maintenance tasks, is the weight of the i-th maintenance task in the maintenance sequence under normal maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence under normal maintenance conditions, is the maintenance time of the i-th maintenance task in the maintenance sequence under normal maintenance conditions; Indicates whether the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is not to repair the j-th maintenance equipment; is the time to repair the jth maintenance equipment, is the weight of maintaining the jth maintenance equipment, is the maintenance time of the mth maintenance task in the maintenance sequence;

[0036] The first constraint conditions include: only one maintenance device can be repaired at a time, and a maintenance task can only have one position in the maintenance sequence.

[0037] Furthermore, when the armored equipment is in an emergency maintenance state, the second objective function is established according to the following formula:

[0038] ,

[0039] The second constraint is:

[0040] ,

[0041] Among them, the second objective function is the weighted completion time of all current maintenance tasks in the emergency maintenance state, is the weight of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence when the maintenance is completed under emergency maintenance status, is the maintenance time of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the maintenance time of the mth maintenance task in the maintenance sequence under emergency maintenance status; Indicates whether the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is not k, k = 0 represents normal maintenance, k = 1 represents overhaul, and k = 2 represents minor repair; and The major repair time and minor repair time are respectively and are the fluctuation rates of maintenance task weights caused by overhaul and minor repair, respectively; a and b are the minimum and maximum loss levels of maintenance equipment for overhaul, respectively; e and f are the minimum and maximum loss levels of maintenance equipment for minor repair, respectively. is the loss level of the i-th maintenance task in the maintenance sequence under emergency maintenance status;

[0042] The second constraint conditions include: only one piece of maintenance equipment can be repaired at a time, a maintenance task can only have one position in the maintenance sequence, the maintenance tasks in the maintenance sequence are either normal maintenance, overhaul or minor repair, the degree of wear and tear of the maintenance equipment for overhaul of any maintenance task in the maintenance sequence under emergency maintenance conditions is in the range of [a, b], and the degree of wear and tear of the maintenance equipment for minor repair of any maintenance task in the maintenance sequence under emergency maintenance conditions is in the range of [e, f].

[0043] Optionally, a flood optimization algorithm is used to optimize the minimization of the weighted completion time of all current maintenance tasks as the optimization goal, and the optimal maintenance sequence of all current maintenance tasks after optimization is determined, including:

[0044] The number, time and weight of all current maintenance tasks are input into the flood optimization algorithm, and the optimal maintenance sequence of all current maintenance tasks and the maintenance time and weight corresponding to each current maintenance task in the optimal maintenance sequence are output.

[0045] A second aspect of the present disclosure provides a maintenance scheduling device for armored equipment, comprising:

[0046] A task determination unit, used to determine all current repair tasks after the loss of armored equipment;

[0047] A weight assignment unit is used to assign a corresponding weight to each current maintenance task using a hypergraph generative adversarial network;

[0048] a time determination unit, configured to determine the maintenance time corresponding to each current maintenance task based on each current maintenance task and the corresponding weight, and to determine the weighted completion time of all current maintenance tasks;

[0049] The optimization unit is used to adopt a flood optimization algorithm, optimize the minimization of the weighted completion time of all current maintenance tasks as the optimization goal, and determine the optimal maintenance sequence of all current maintenance tasks after optimization.

[0050] A third aspect of the present disclosure provides a computer-readable storage medium storing computer instructions, which are used to enable a computer to execute the armored equipment maintenance scheduling method provided by any one of the first aspects.

[0051] The fourth aspect of the present disclosure provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor executes the maintenance scheduling method for armored equipment provided in any one of the first aspects.

[0052] In the armored equipment maintenance scheduling method provided by the present disclosure, all current maintenance tasks after the armored equipment is damaged are determined; a corresponding weight is assigned to each current maintenance task using a hypergraph generative adversarial network; and the hypergraph generative adversarial network, with its strong nonlinear fitting capabilities, simple learning rules, and ease of computer implementation, is utilized to simulate a human's comprehensive consideration of the value of the maintenance equipment and assign a reasonable weight to each current maintenance task.

[0053] Based on each current maintenance task and its corresponding weight, the maintenance time corresponding to each current maintenance task is determined, and the weighted completion time of all current maintenance tasks is determined. A flood optimization algorithm is used to minimize the weighted completion time of all current maintenance tasks as the optimization objective, and the optimal maintenance sequence of all current maintenance tasks is determined after optimization. The flood optimization algorithm is used to determine the minimized weighted completion time of all current maintenance tasks, determine the optimal maintenance sequence of all current maintenance tasks, and find the optimal scheduling solution that minimizes the weighted completion time required for maintenance. This improves the efficiency of armored equipment maintenance scheduling and effectively implements post-maintenance scheduling for armored equipment, solving the problem of related technologies that cannot achieve efficient maintenance scheduling for armored equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the related technologies, the following briefly introduces the drawings required for use in the specific embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0055] Figure 1 A flowchart of a maintenance scheduling method for armored equipment provided in an embodiment of the present disclosure;

[0056] Figure 2 A block diagram of a maintenance and scheduling device for armored equipment provided in an embodiment of the present disclosure;

[0057] Figure 3 A block diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the embodiments described are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present disclosure.

[0059] It should be noted that the terms "first," "second," and the like in the specification and claims of the present disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate for the embodiments of the present disclosure described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or inherent to these processes, methods, products, or apparatus.

[0060] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure can be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0061] In the context of economic globalization, armored equipment plays a vital role in the rapid flow of materials. With the increase in armored equipment throughput and the proliferation of mechanized operations, armored equipment is often operating in an overloaded state. Furthermore, mechanized operations conducted outdoors are susceptible to inclement weather, leading to armored equipment failures. Therefore, effective maintenance and scheduling of armored equipment is crucial to maintaining its efficient operation. However, existing technologies have been unable to achieve efficient maintenance and scheduling for armored equipment.

[0062] In order to solve the above problems, the embodiment of the present disclosure provides a maintenance scheduling method for armored equipment, such as Figure 1 As shown, the method includes the following steps S11 to S14:

[0063] Step S11: determining all current maintenance tasks after the armored equipment is damaged;

[0064] Step S12: Using a hypergraph generative adversarial network to assign a corresponding weight to each current maintenance task; using the hypergraph generative adversarial network to quantify the weight of the current maintenance task, and taking advantage of the hypergraph generative adversarial network's strong nonlinear fitting ability, simple learning rules, and ease of computer implementation, to simulate manual comprehensive consideration of the value of maintenance equipment, and assign a reasonable weight to each current maintenance task or each maintenance equipment;

[0065] Step S13: Based on each current maintenance task and the corresponding weight, determine the maintenance time corresponding to each current maintenance task, and determine the weighted completion time of all current maintenance tasks;

[0066] Step S14: Using the flood optimization algorithm, the optimization objective is to minimize the weighted completion time of all current maintenance tasks, and determine the optimal maintenance sequence for all current maintenance tasks. The optimal maintenance sequence is the optimal maintenance scheduling solution. The order of each maintenance task in the optimal maintenance sequence can be used to determine the position of each maintenance task in the optimal maintenance sequence. When the maintenance conditions are met, the weighted completion time of each maintenance task can be reduced by rationally arranging the maintenance sequence of each current maintenance task.

[0067] The present invention utilizes a hypergraph generative adversarial network to assign a weight to each current maintenance task, considers the minimized weighted completion time of all current maintenance tasks, and uses a flood optimization algorithm (FLA) for optimization to find the optimal maintenance scheduling plan so that the weighted completion time required for maintenance is minimized, thereby improving the efficiency of armored equipment maintenance scheduling and efficiently realizing post-maintenance scheduling of armored equipment, thus solving the problem in related technologies that it is impossible to realize efficient maintenance scheduling of armored equipment.

[0068] In an optional embodiment of the present disclosure, step S12 includes:

[0069] Initializing a hypergraph generative adversarial network, using a steepest descent method with momentum to preliminarily train the hypergraph generative adversarial network in a parallel training manner, and obtaining a hypergraph generative adversarial network after the preliminary training, wherein the hypergraph generative adversarial network uses a sigmoid function as a transfer function and includes one or two hidden layers;

[0070] The maintenance workload ratio and maintenance cycle of each historical maintenance task are input into the hypergraph generative adversarial network after preliminary training, and the result is output. The error is determined based on the results of two consecutive outputs. The hidden layer weights and thresholds of the hypergraph generative adversarial network after preliminary training are corrected based on the error, and the training is repeated. Among them, the maintenance workload ratio refers to the ratio of the time for repairing a certain maintenance equipment to the total time for repairing equipment of the same category. The use of the maintenance workload ratio can ensure the rationality of the weight of the maintenance of a type of equipment. For example, due to the variety of armored equipment, the same type of armored equipment can also be subdivided in many aspects. Armored cantilever gantry cranes are divided into double cantilevers and single cantilever cranes. The workload of repairing the two types of equipment is originally different, which will affect the determination of the equipment maintenance weight. The embodiment of the present disclosure uses the maintenance workload ratio instead of the maintenance workload as the basis for determining the weight, so as to ensure the rationality of the subsequent determination of the maintenance task weight.

[0071] When the training is repeated until the sum of the squares of the errors is less than or equal to the preset error tolerance, or when the number of repeated training reaches the preset maximum number of iterations, the training is stopped to obtain the hypergraph generative adversarial network after repeated training. When the training is repeated until the sum of the squares of the errors is less than or equal to the preset error tolerance, the hypergraph generative adversarial network converges, and the preset maximum number of iterations can be 1000 times.

[0072] By analyzing and comparing the maintenance workload ratio and maintenance cycle of each historical maintenance task, the accuracy of the hypergraph generative adversarial network after repeated training is determined. The hypergraph generative adversarial network after repeated training is tested to obtain the weight corresponding to each current maintenance task of the armored equipment. Based on the maintenance workload ratio and maintenance cycle of each historical maintenance task, a hypergraph generative adversarial network is trained to more accurately assign weights to maintenance tasks.

[0073] According to the normal maintenance state or emergency maintenance state of the armored equipment, the current maintenance task of the armored equipment and the change of the corresponding weight of the current maintenance task are determined.

[0074] The basic idea of ​​the hypergraph generative adversarial network's learning algorithm is to train the hypergraph generative adversarial network to complete a task by providing it with training examples (such as various historical maintenance tasks). During this process, the weight of each unit is adjusted to reduce the error between the expected output and the actual output. The input of the hypergraph generative adversarial network is the maintenance cycle and the proportion of maintenance workload, and the output is the weight of the maintenance task.

[0075] In a preferred embodiment of the present disclosure, the maintenance workload ratio and maintenance cycle of each historical maintenance task are input into the hypergraph generative adversarial network after preliminary training, including:

[0076] Normalize the maintenance workload ratio and maintenance cycle of each historical maintenance task to obtain normalized historical maintenance data. Input the normalized historical maintenance data into the hypergraph generative adversarial network after preliminary training.

[0077] After obtaining the hypergraph generative adversarial network after repeated training, the method further includes:

[0078] The normalized historical maintenance data is denormalized to obtain the maintenance workload ratio and maintenance cycle of each historical maintenance task.

[0079] By normalizing the maintenance workload ratio and maintenance cycle before input and then performing denormalization after training, the convergence speed and algorithm accuracy of the hypergraph generative adversarial network can be improved.

[0080] In an optional embodiment of the present disclosure, the hypergraph generative adversarial network is an interactive hyperedge neuron module or MRL-AHF. HGGAN proposes using an interactive hyperedge neuron module (IHEN) as a generator to capture complex relationships between data; MRL-AHF utilizes the complementarity between multiple modalities and the interaction within multiple modalities to improve representation learning capabilities and multimodal fusion performance.

[0081] In a preferred embodiment of the present disclosure, when the hypergraph generative adversarial network is an interactive hyperedge neuron module, the interactive hyperedge neuron module is used as a generator to capture the complex relationship between data, and the discriminator is an MLP;

[0082] Determine the characteristics of the nodes in the l+1 layer according to the following formula and the characteristics of the hyperedge at layer l+1 :

[0083] ,

[0084] in, is the activation function, is the hypergraph incidence matrix, yes The transpose of is the feature of the hyperedge at layer l, is the feature of the node in layer l, is the weight matrix of the hyperedge at layer l, is the weight matrix of the nodes in layer l, is a hyperparameter;

[0085] The nodes in the last layer are used as the weights corresponding to the current maintenance task.

[0086] In a preferred embodiment of the present disclosure, when the hypergraph generative adversarial network is MRL-AHF, the potential representation is fed into the encoders EA and EB respectively to obtain the representation and , and the two generated hypergraphs use the incidence matrix and Indicates that hypergraph fusion is achieved through adversarial training strategy to obtain vertex features :

[0087] ,

[0088] in, and They are and The vertex degree matrix of , , and They are and The hyperedge degree matrix of yes The transpose of is the weight matrix;

[0089] The features of the vertices are used as the weights corresponding to the current maintenance task.

[0090] In an optional embodiment of the present disclosure, determining a current maintenance task of the armored equipment and a change in a weight corresponding to the current maintenance task according to whether the armored equipment is in a normal maintenance state or an emergency maintenance state includes:

[0091] When the armored equipment is in normal maintenance status, the first objective function is established according to the following formula:

[0092] ,

[0093] The first constraint is:

[0094] ,

[0095] Among them, the first objective function is the weighted completion time of all current maintenance tasks under normal maintenance status, n is the total number of current maintenance tasks, is the weight of the i-th maintenance task in the maintenance sequence under normal maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence under normal maintenance conditions, is the maintenance time of the i-th maintenance task in the maintenance sequence under normal maintenance conditions; Indicates whether the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is not to repair the j-th maintenance equipment; is the time to repair the jth maintenance equipment, is the weight of maintaining the jth maintenance equipment, is the maintenance time of the mth maintenance task in the maintenance sequence;

[0096] The first constraint is that only one piece of equipment can be repaired at a time, and each maintenance task can only have one position in the maintenance sequence. For example, when a ship enters the port to unload cargo, if an armored quay crane or gantry crane fails, a maintenance team can only repair the armored quay crane first to ensure unloading, and then repair the gantry crane to ensure the cargo is loaded into the yard.

[0097] Furthermore, because armored equipment is subject to sudden damage or loss, the present disclosure considers both normal maintenance conditions and emergency maintenance scheduling arrangements. Emergency maintenance means that during a scheduled maintenance schedule, some equipment to be repaired experiences a change in maintenance status, making the original maintenance method incapable of completing the repair. For example, an armored crane originally scheduled for repair may only have a loose conveyor belt that only requires re-fixing. However, during use, the conveyor belt becomes stuck in a conveyor bearing, causing the bearing to become misaligned due to forced start-up. In this case, the simple re-fixing repair is replaced by a bearing reset, which not only increases the maintenance workload but, if not addressed promptly, may even affect the crane's overall lifespan and increase safety hazards.

[0098] In an emergency maintenance state, the original maintenance task weight changes, and the maintenance time and maintenance method will also change. The disclosed embodiment visualizes this change as a change in the degree of wear and tear of the maintenance equipment. The degree of wear and tear of the equipment refers to the ratio of the equipment's service life to its expected service life. A lower degree of wear and tear of the maintenance equipment indicates that the frequency of use is lower than normal, and there is an unreasonable cost problem. For example, the automatic guided vehicle to be repaired only has a dent in the outer shell and can work, but because the maintenance record has been placed, the working efficiency of the guided vehicle is reduced. The degree of wear and tear of the maintenance equipment increases, and the equipment has a greater hidden danger of failure, which is the focus of maintenance. The change in the degree of wear and tear of the maintenance equipment will change the basic steps required for maintenance accordingly, and the corresponding maintenance tasks, maintenance methods and maintenance time will change. This change will be directly reflected in the total time required to complete the maintenance. These changes must be taken into account to ensure the rationality and economy of the maintenance sequence. Therefore, in an emergency maintenance state, the embodiment of the present disclosure defines equipment maintenance with a high degree of equipment loss as a major repair, and equipment maintenance with a low degree of equipment loss as a minor repair; the embodiment of the present disclosure reasonably arranges the order of maintenance tasks in maintenance scheduling to avoid some important maintenance tasks from being arranged in the wrong maintenance location, thereby reducing maintenance delays, time losses and economic losses.

[0099] In an optional embodiment of the present disclosure, when the armored equipment is in an emergency maintenance state, the second objective function is established according to the following formula:

[0100] ,

[0101] The second constraint is:

[0102] ,

[0103] Among them, the second objective function is the weighted completion time of all current maintenance tasks in the emergency maintenance state, is the weight of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence under emergency maintenance status, is the maintenance time of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the maintenance time of the mth maintenance task in the maintenance sequence under emergency maintenance status; Indicates whether the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is not k, k = 0 represents normal maintenance, k = 1 represents overhaul, and k = 2 represents minor repair; and The major repair time and minor repair time are respectively and are the fluctuation rates of maintenance task weights caused by overhaul and minor repair, respectively; a and b are the minimum and maximum loss levels of maintenance equipment for overhaul, respectively; e and f are the minimum and maximum loss levels of maintenance equipment for minor repair, respectively. is the loss level of the i-th maintenance task in the maintenance sequence under emergency maintenance status;

[0104] The second constraint conditions include: only one piece of maintenance equipment can be repaired at a time; a maintenance task can only have one position in the maintenance sequence; maintenance tasks in the maintenance sequence can be either normal overhaul or minor repair; the wear and tear of the equipment for any overhaul task in the maintenance sequence under emergency maintenance conditions must be in the range [a, b]; and the wear and tear of the equipment for any minor repair task in the maintenance sequence under emergency maintenance conditions must be in the range [e, f]. If the wear and tear is within [a, b], a major repair is performed; if the wear and tear is within [e, f], a minor repair is performed. The maintenance time for each maintenance task is determined based on actual conditions, and the thresholds for overhaul and minor repair are not fixed values ​​but are related to the specific maintenance equipment.

[0105] In an optional embodiment of the present disclosure, step S14 includes:

[0106] The number, time and weight of all current maintenance tasks are input into the flood optimization algorithm, and the optimal maintenance sequence of all current maintenance tasks and the maintenance time and weight corresponding to each current maintenance task in the optimal maintenance sequence are output.

[0107] The disclosed embodiments address the post-maintenance scheduling problem through a flood optimization algorithm. The flood optimization algorithm introduces a new meta-heuristic optimization algorithm that draws inspiration from the complex movement and flow patterns of water masses in watershed flood events. The flood optimization algorithm mathematically models key phenomena, such as the movement of water uphill, flow velocity over time, the effects of soil permeability, and the cyclical increase and decrease of water levels due to precipitation and losses. Using these models, the algorithm guides the population of potential solutions to move and evolve towards increasing optimality.

[0108] In the embodiment of the present disclosure, the input of the flood optimization algorithm is the number, time and corresponding weight of all current maintenance tasks, and the output is the optimal maintenance sequence of all current maintenance tasks, as well as the maintenance time and weight corresponding to each current maintenance task in the optimal maintenance sequence.

[0109] The present disclosure addresses the maintenance scheduling problem of armored equipment after damage, that is, the scheduling problem of post-repair. By analyzing the post-repair scheduling arrangement of armored equipment, a hypergraph generative adversarial network algorithm is used to quantify the weights of armored equipment to be repaired, and a flood optimization algorithm is used to minimize the weighted completion time of maintenance tasks, thereby obtaining an optimized maintenance scheduling sequence and a corresponding maintenance time schedule. Through a maintenance example of armored equipment, the application of the optimized scheduling model in port machinery and equipment is demonstrated, the maintenance sequence of port machinery and equipment is clarified, and maintenance time is saved while ensuring the completion of maintenance tasks, providing a reference for armored equipment maintenance plans.

[0110] From the above description, it can be seen that the present disclosure achieves the following technical effects:

[0111] The present invention uses a hypergraph generative adversarial network to assign a weight to each current maintenance task, considers the minimized weighted completion time of all current maintenance tasks, and uses a flood optimization algorithm for optimization to find the optimal maintenance scheduling solution that minimizes the weighted completion time required for maintenance. This improves the efficiency of armored equipment maintenance scheduling and efficiently implements post-maintenance scheduling for armored equipment, solving the problem of related technologies that cannot achieve efficient maintenance scheduling for armored equipment.

[0112] The position of each maintenance task in the optimal maintenance sequence can be determined by the sequence of each maintenance task in the optimal maintenance sequence. When the maintenance conditions are met, the weighted completion time considering the weight of the maintenance task can be reduced by reasonably arranging the maintenance sequence of each current maintenance task.

[0113] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0114] The present disclosure also provides a maintenance scheduling device for armored equipment for implementing the above method embodiment. Figure 2 As shown, the maintenance scheduling device 20 includes:

[0115] A task determination unit 21 is used to determine all current maintenance tasks after the armored equipment is damaged;

[0116] A weight assigning unit 22 is used to assign a corresponding weight to each current maintenance task using a hypergraph generative adversarial network;

[0117] A time determination unit 23 is configured to determine the maintenance time corresponding to each current maintenance task based on each current maintenance task and the corresponding weight, and to determine the weighted completion time of all current maintenance tasks;

[0118] The optimization unit 24 is configured to use a flood optimization algorithm to optimize the minimization of the weighted completion time of all current maintenance tasks as an optimization goal, and determine the optimal maintenance sequence of all current maintenance tasks after optimization.

[0119] The specific manner in which each unit in the above device embodiment performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0120] The present disclosure also provides an electronic device, such as Figure 3 As shown, the electronic device includes one or more processors 31 and a memory 32. Figure 3 A processor 31 is taken as an example.

[0121] The controller may further include an input device 33 and an output device 34 .

[0122] The processor 31, memory 32, input device 33 and output device 34 may be connected via a bus or other means. Figure 3 The bus connection is taken as an example.

[0123] The processor 31 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips. The general-purpose processor may be a microprocessor or any conventional processor.

[0124] Memory 32, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as the program instructions / modules corresponding to the control method in the embodiments of the present disclosure. Processor 31 executes the non-transitory software programs, instructions, and modules stored in memory 32 to execute various server functional applications and data processing, thereby implementing the armored equipment maintenance scheduling method of the aforementioned method embodiment.

[0125] The memory 32 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the processing device operated by the server, etc. In addition, the memory 32 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 32 may optionally include a memory remotely located relative to the processor 31, and these remote memories may be connected to a network connection device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0126] The input device 33 can receive input digital or character information and generate key signal input related to user settings and function control of the processing device of the server. The output device 34 can include a display device such as a display screen.

[0127] One or more modules are stored in the memory 32 and when executed by one or more processors 31, perform the following operations: Figure 1 The method shown.

[0128] Those skilled in the art will appreciate that all or part of the processes in the above method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes in the above method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FM), a hard disk drive (HDD), or a solid-state drive (SSD). The storage medium can also include a combination of the above types of memory.

[0129] Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A maintenance scheduling method for armored equipment, characterized in that: include: Identify all current repair tasks following loss of armored equipment; Use hypergraph generative adversarial networks to assign corresponding weights to each current maintenance task; Based on each current maintenance task and the corresponding weight, determine the maintenance time corresponding to each current maintenance task, and determine the weighted completion time of all current maintenance tasks; Using the flood optimization algorithm, the optimization objective is to minimize the weighted completion time of all current maintenance tasks and determine the optimal maintenance sequence of all current maintenance tasks after optimization. The optimal maintenance sequence is the optimal maintenance scheduling solution. The position of each current maintenance task in the optimal maintenance sequence is determined by the order of each current maintenance task in the optimal maintenance sequence. The method of using a hypergraph generative adversarial network to assign a corresponding weight to each current maintenance task includes: Initializing a hypergraph generative adversarial network, and preliminarily training the hypergraph generative adversarial network in a parallel training manner using a steepest descent method including momentum to obtain a hypergraph generative adversarial network after the preliminary training, wherein the hypergraph generative adversarial network uses a sigmoid function as a transfer function; Input the maintenance workload ratio and maintenance cycle of each historical maintenance task into the hypergraph generative adversarial network after the initial training, output the result, and determine the error based on the results of two consecutive outputs. Based on the error, the hidden layer weights and thresholds of the hypergraph generative adversarial network after the initial training are corrected, and the training is repeated; wherein the maintenance workload ratio is the ratio of the time to repair a piece of maintenance equipment to the total time to repair equipment of the same category; When the training is repeated until the sum of squares of the errors is less than or equal to the preset error tolerance, or when the number of repeated training reaches the preset maximum number of iterations, the training is stopped to obtain a hypergraph generative adversarial network after repeated training; By analyzing and comparing the maintenance workload ratio and maintenance cycle of each historical maintenance task, the accuracy of the hypergraph generative adversarial network after repeated training is determined, and the hypergraph generative adversarial network after repeated training is tested to obtain the weight corresponding to each current maintenance task of the armored equipment; Determining a current maintenance task of the armored equipment and a change in a weight corresponding to the current maintenance task according to the normal maintenance state or the emergency maintenance state of the armored equipment; Wherein, the hypergraph generative adversarial network is an interactive hyperedge neuron module or MRL-AHF; When the hypergraph generative adversarial network is an interactive hyperedge neuron module, the interactive hyperedge neuron module is used as a generator to capture the complex relationship between data, the discriminator is an MLP, and the nodes in the last layer are used as the weights corresponding to the current maintenance task; When the hypergraph generative adversarial network is MRL-AHF, the potential representation is fed into encoders EA and EB respectively to obtain the representation and , and the two generated hypergraphs use the incidence matrix and Indicates that the features of the vertices are used as the weights corresponding to the current maintenance task.

2. The method according to claim 1, characterized in that in, Inputting the maintenance workload ratio and maintenance cycle of each historical maintenance task into the hypergraph generative adversarial network after the preliminary training is completed includes: Normalizing the maintenance workload ratio and maintenance cycle of each historical maintenance task to obtain normalized historical maintenance data, and inputting the normalized historical maintenance data into the hypergraph generative adversarial network after the preliminary training; After obtaining the hypergraph generative adversarial network after repeated training, the method further includes: The normalized historical maintenance data is subjected to denormalization to obtain the maintenance workload ratio and maintenance cycle of each historical maintenance task.

3. The method according to claim 1, characterized in that When the hypergraph generative adversarial network is an interactive hyperedge neuron module, the characteristics of the nodes in the l+1 layer are determined according to the following formulas: and the characteristics of the hyperedge at layer l+1 : in, is the activation function, is the hypergraph incidence matrix, yes The transpose of is the feature of the hyperedge at layer l, is the feature of the node in layer l, is the weight matrix of the hyperedge at layer l, is the weight matrix of the nodes in layer l, is a hyperparameter.

4. The method according to claim 1, wherein When the hypergraph generation adversarial network is MRL-AHF, hypergraph fusion is achieved through adversarial training strategy to obtain vertex features. : in, and They are and The vertex degree matrix of , , and They are and The hyperedge degree matrix of yes The transpose of is the weight matrix.

5. The method according to claim 1, wherein The determining, based on the normal maintenance state or the emergency maintenance state of the armored equipment, of the current maintenance task of the armored equipment and the change of the weight corresponding to the current maintenance task includes: When the armored equipment is in a normal maintenance state, the first objective function is established according to the following formula: The first constraint is: Among them, the first objective function is the weighted completion time of all current maintenance tasks under normal maintenance status, n is the total number of current maintenance tasks, is the weight of the i-th maintenance task in the maintenance sequence under normal maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence under normal maintenance conditions, is the maintenance time of the i-th maintenance task in the maintenance sequence under normal maintenance conditions; Indicates whether the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is to repair the j-th maintenance equipment, Indicates that the i-th maintenance task is not to repair the j-th maintenance equipment; is the time to repair the jth maintenance equipment, is the weight of maintaining the jth maintenance equipment, is the maintenance time of the mth maintenance task in the maintenance sequence; The first constraint conditions include: only one maintenance device can be repaired at a time, and a maintenance task can only have one position in the maintenance sequence.

6. The method according to claim 5, characterized in that When the armored equipment is in an emergency maintenance state, the second objective function is established according to the following formula: The second constraint is: Among them, the second objective function is the weighted completion time of all current maintenance tasks in the emergency maintenance state, is the weight of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the cumulative completion time of the first i maintenance tasks in the maintenance sequence when the maintenance is completed under emergency maintenance status, is the maintenance time of the i-th maintenance task in the maintenance sequence under emergency maintenance status, is the maintenance time of the mth maintenance task in the maintenance sequence under emergency maintenance status; Indicates whether the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is k, Indicates that the maintenance method of the i-th maintenance task is not k, k = 0 represents normal maintenance, k = 1 represents overhaul, and k = 2 represents minor repair; and The major repair time and minor repair time are respectively and are the fluctuation rates of maintenance task weights caused by overhaul and minor repair, respectively; a and b are the minimum and maximum loss levels of maintenance equipment for overhaul, respectively; e and f are the minimum and maximum loss levels of maintenance equipment for minor repair, respectively. is the loss level of the i-th maintenance task in the maintenance sequence under emergency maintenance status; The second constraint conditions include: only one piece of maintenance equipment can be repaired at a time, a maintenance task can only have one position in the maintenance sequence, the maintenance task in the maintenance sequence is one of normal maintenance, overhaul and minor repair, the degree of wear and tear of the maintenance equipment for overhaul of any maintenance task in the maintenance sequence under emergency maintenance conditions is in the range of [a, b], and the degree of wear and tear of the maintenance equipment for minor repair of any maintenance task in the maintenance sequence under emergency maintenance conditions is in the range of [e, f].

7. The method according to claim 1, characterized in that The flood optimization algorithm is used to optimize the minimization of the weighted completion time of all current maintenance tasks as the optimization goal, and the optimal maintenance sequence of all current maintenance tasks after optimization is determined, including: The number, time and weight of all current maintenance tasks are input into the flood optimization algorithm, and the optimal maintenance sequence of all current maintenance tasks and the maintenance time and weight corresponding to each current maintenance task in the optimal maintenance sequence are output.

8. A maintenance and scheduling device for armored equipment, characterized in that: include: A task determination unit, used to determine all current repair tasks after the loss of armored equipment; A weight assignment unit is used to assign a corresponding weight to each current maintenance task using a hypergraph generative adversarial network; a time determination unit, configured to determine, based on each current maintenance task and the corresponding weight, a maintenance time corresponding to each current maintenance task, and determine a weighted completion time of all current maintenance tasks; An optimization unit is used to optimize the optimal maintenance sequence of all current maintenance tasks by using a flood optimization algorithm, with minimization of the weighted completion time of all current maintenance tasks as the optimization objective, wherein the optimal maintenance sequence is the optimal maintenance scheduling plan, and the position of each current maintenance task in the optimal maintenance sequence is determined by the order of each current maintenance task in the optimal maintenance sequence; The method of using a hypergraph generative adversarial network to assign a corresponding weight to each current maintenance task includes: Initializing a hypergraph generative adversarial network, and preliminarily training the hypergraph generative adversarial network in a parallel training manner using a steepest descent method including momentum to obtain a hypergraph generative adversarial network after the preliminary training, wherein the hypergraph generative adversarial network uses a sigmoid function as a transfer function; Input the maintenance workload ratio and maintenance cycle of each historical maintenance task into the hypergraph generative adversarial network after the initial training, output the result, and determine the error based on the results of two consecutive outputs. Based on the error, the hidden layer weights and thresholds of the hypergraph generative adversarial network after the initial training are corrected, and the training is repeated; wherein the maintenance workload ratio is the ratio of the time to repair a piece of maintenance equipment to the total time to repair equipment of the same category; When the training is repeated until the sum of squares of the errors is less than or equal to the preset error tolerance, or when the number of repeated training reaches the preset maximum number of iterations, the training is stopped to obtain a hypergraph generative adversarial network after repeated training; By analyzing and comparing the maintenance workload ratio and maintenance cycle of each historical maintenance task, the accuracy of the hypergraph generative adversarial network after repeated training is determined, and the hypergraph generative adversarial network after repeated training is tested to obtain the weight corresponding to each current maintenance task of the armored equipment; Determining a current maintenance task of the armored equipment and a change in a weight corresponding to the current maintenance task according to the normal maintenance state or the emergency maintenance state of the armored equipment; Wherein, the hypergraph generative adversarial network is an interactive hyperedge neuron module or MRL-AHF; When the hypergraph generative adversarial network is an interactive hyperedge neuron module, the interactive hyperedge neuron module is used as a generator to capture the complex relationship between data, the discriminator is an MLP, and the nodes in the last layer are used as the weights corresponding to the current maintenance task; When the hypergraph generative adversarial network is MRL-AHF, the potential representation is fed into encoders EA and EB respectively to obtain the representation and , and the two generated hypergraphs use the incidence matrix and Indicates that the features of the vertices are used as the weights corresponding to the current maintenance task.

Images