Aircraft hangar scheduled inspection and maintenance scheduling method, device, equipment and medium

Abstract

The present application relates to a method, device, equipment and medium for scheduling scheduled inspection and maintenance of an aircraft hangar. The method includes obtaining aircraft maintenance and support data; determining the start time of each process based on the priority number coding method according to the maintenance and support data, and generating a process time matrix without considering resource constraints by left-right iteration; generating a bit sequence list of processes based on the regret value deviation of the latest completion time matrix, and performing a decoding operation; performing an overall index evaluation on the bit sequence list; based on the overall index evaluation, calling an improved teaching and learning algorithm to perform a left-hand operation to perform population evolution processing; performing population evolution processing based on the left-hand operation, and performing processing in the reinforcement learning stage; comparing the target evaluation times obtained by the processing in the reinforcement learning stage with a limit value, and if the target evaluation times are greater than the limit value, outputting the final generated left scheduling plan; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar. A high scheduling efficiency is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of scheduled inspection and maintenance scheduling of hangars, and in particular to a method, device, equipment and medium for scheduling scheduled inspection and maintenance of aircraft hangars. Background Art

[0002] After the aircraft fleet is recovered from the mission, it may face four situations: First, during the mission, the aircraft systems have malfunctioned and need to be transferred to the hangar for maintenance; second, it has reached the scheduled maintenance cycle and also needs to be transferred to the hangar for scheduled inspection and maintenance; third, it continues to perform pre-flight support operations before being dispatched again; fourth, it remains in the hangar without subsequent tasks. The first two situations require coordinated scheduling and planning of maintenance tasks in the hangar, and the maintenance personnel, maintenance workshops / equipment, and space in the hangar are limited and reasonable scheduling and planning are required; in addition, due to the limitation of the number of maintenance and support personnel, it is impossible to flexibly supplement according to the tasks, which requires maintenance personnel to have strong multi-aircraft support and cross-professional support capabilities, and higher skill level requirements, and the efficiency of maintenance operations will directly affect the availability of the aircraft fleet for subsequent missions.

[0003] Effective time planning and resource allocation for aircraft fleet maintenance tasks and related maintenance personnel and facilities and equipment in a hangar operation environment, so that the aircraft can be restored to a usable state as soon as possible, is critical for the fleet to continue to perform tasks. However, in the process of realizing the present invention, the inventor found that the traditional aircraft hangar scheduled inspection and maintenance scheduling method still has the technical problem of low efficiency in aircraft hangar scheduled inspection and maintenance scheduling. Summary of the invention

[0004] Based on this, it is necessary to provide an aircraft hangar scheduled inspection and maintenance scheduling method with high aircraft hangar scheduled inspection and maintenance scheduling efficiency, an aircraft hangar scheduled inspection and maintenance scheduling device, a computer device and a computer-readable storage medium to address the above technical problems.

[0005] In order to achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0006] On the one hand, an embodiment of the present invention provides an aircraft hangar scheduled inspection and maintenance scheduling method, comprising the steps of:

[0007] Obtain aircraft maintenance and support data; maintenance and support data includes the number of aircraft, time of entry, single-aircraft maintenance process, maintenance personnel and skill data;

[0008] According to the maintenance support data, the start time of each process is determined by using the priority number coding method, and the process time matrix without considering resource constraints is generated by left and right iteration; the process time matrix includes the earliest start time matrix, the latest start time matrix, the earliest completion time matrix and the latest completion time matrix;

[0009] Generate a bit sequence list of the process based on the regret value deviation of the latest completion time matrix and perform a decoding operation;

[0010] Conduct overall index evaluation on the position sequence table;

[0011] Based on the overall indicator evaluation, the improved teaching and learning algorithm is called to perform left-hand operation to execute population evolution processing;

[0012] Based on the left operation, the population evolution process is performed, and the process of the reinforcement learning phase is performed;

[0013] The target evaluation times obtained in the reinforcement learning phase are compared with the limit value. If the target evaluation times are greater than the limit value, the final generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of aircraft hangars.

[0014] On the other hand, a device for scheduling scheduled inspection and maintenance of an aircraft hangar is also provided, comprising:

[0015] The data acquisition module is used to obtain aircraft maintenance and support data; the maintenance and support data includes the number of aircraft, time of entry, single-aircraft maintenance process, maintenance personnel and skill data;

[0016] The time matrix module is used to determine the start time of each process based on the maintenance support data by using the priority number coding method, and to iterate left and right to generate the process time matrix without considering resource constraints; the process time matrix includes the earliest start time matrix, the latest start time matrix, the earliest completion time matrix and the latest completion time matrix;

[0017] A list decoding module is used to generate a bit sequence list of the process based on the regret value deviation of the latest completion time matrix and perform a decoding operation;

[0018] Index evaluation module, used to evaluate the overall index of the position sequence table;

[0019] The population evolution module is used to call the improved teaching and learning algorithm to perform left-hand operations and execute population evolution processing based on the overall indicator evaluation;

[0020] A reinforcement learning module, used to perform population evolution processing based on left-hand operations and to perform processing in the reinforcement learning phase;

[0021] The scheduling output module is used to compare the target evaluation times obtained in the reinforcement learning stage with the limit value. If the target evaluation times are greater than the limit value, the final generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

[0022] On the other hand, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method when executing the computer program.

[0023] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, any step of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method is implemented.

[0024] One of the above technical solutions has the following advantages and beneficial effects:

[0025] The above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method, device, equipment and medium, by adopting an improved teaching and learning algorithm, optimizes the calculation and processing of the scheduling plan based on real maintenance support data. Compared with the traditional scheduling method at this stage, it includes the optimization research goal of the hangar maintenance support task of the fleet, focusing on minimizing the maintenance completion time, and considering the availability requirements of the fleet wave dispatch, considering the multi-mode mixing of fault repair and scheduled inspection maintenance, and comprehensively considering the actual constraints such as the maintenance support scope, maintenance parallel capacity, and maintenance station space. It can automatically and efficiently output reliable aircraft hangar scheduled inspection and maintenance scheduling plans to meet the needs of refined maintenance tasks and achieve the purpose of significantly improving the efficiency of aircraft hangar scheduled inspection and maintenance scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flowchart of a method for scheduling scheduled inspection and maintenance of an aircraft hangar in one embodiment;

[0027] Figure 2 is a schematic diagram of a processing flow of an improved teaching and learning algorithm in one embodiment;

[0028] Figure 3 It is a multi-machine guarantee flow chart and a process-skill-personnel correspondence diagram in one embodiment;

[0029] Figure 4 A schematic diagram of a serial scheduling process in an embodiment;

[0030] Figure 5 A Gantt chart for scheduling maintenance personnel in one embodiment;

[0031] Figure 6 The figure is a schematic diagram of the module structure of an aircraft hangar scheduled inspection and maintenance scheduling device in one embodiment. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art to which the present application belongs. The terms used herein in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0034] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0035] In practice, the inventors have found that the currently commonly used complex aircraft hangar inspection and maintenance scheduling methods have the following shortcomings: (1) Calling commercial finite element software or high-precision finite element analysis programs to predict the bearing performance of the structure. Such methods require calling relatively time-consuming high-precision finite element programs to simulate different aircraft structures. A configuration may take several hours or even days to obtain results, and the computational cost increases significantly with the improvement of the accuracy of the results. This is difficult to bear in actual optimization design projects.

[0036] (2) Constructing a complex aircraft hangar inspection and maintenance scheduling model. Although it can quickly obtain the performance parameters of the new structural form compared to the previous method, the constructed prediction model usually has poor prediction accuracy, and it is even more difficult to obtain prediction results that meet the design requirements for the new configuration.

[0037] This application aims at the technical problem of low efficiency of the above-mentioned traditional complex aircraft hangar scheduled inspection and maintenance scheduling method, and provides an efficient aircraft hangar scheduled inspection and maintenance scheduling method based on an improved teaching and learning algorithm. In the hangar scheduled inspection and maintenance mode, the start time of each process is first determined based on the priority number coding method, and the earliest start / latest start / earliest completion / latest completion time is generated by left and right iteration without considering resource constraints. Then, the process list is determined based on regret deviation sampling, and then the population target value is evaluated. The fitness of the population is calculated according to the scheduling process, and then the improved teaching and learning algorithm program is called to perform left-hand operation population evolution, and finally the results are analyzed and the scheduling plan is output. It can be applied to the field of multi-mode hybrid and maintenance guarantee range hangar scheduled inspection and maintenance scheduling, especially involving fault repair and scheduled inspection maintenance with availability requirements for the dispatch of aircraft groups.

[0038] The fleet maintenance in the aircraft hangar plays a vital role in ensuring the safety and reliability of the aircraft. The aircraft needs to be fault-checked before and after the flight mission. If a major fault is detected, it will be recovered to the hangar, moored in place and queued for maintenance. In addition to the unplanned random fault maintenance work, it is also necessary to complete the planned preventive regular inspection and maintenance, which usually includes 25h, 50h, and 100h scheduled maintenance. The hangar needs to complete a wide range of maintenance tasks and inspections to maintain a high level of fleet availability. The fleet availability also depends on the throughput efficiency of the hangar's scheduled maintenance. Therefore, scientific timing planning and resource allocation of relevant maintenance personnel and facilities and equipment are the key to scheduling aircraft fleet hangar maintenance tasks.

[0039] See also Figure 1 and Figure 2 In one embodiment, the present invention provides an aircraft hangar scheduled inspection and maintenance scheduling method, comprising the following steps S0 to S7:

[0040] S0, obtain the aircraft maintenance and support data; the maintenance and support data includes the number of aircraft, entry time, single-aircraft maintenance process, maintenance personnel and skill data.

[0041] Optionally, let the set of aircraft to be maintained be I = {1, 2, …, Nm}. The maintenance process set of the i-th aircraft is J i ={1,2,…,|J i |}, the total maintenance process set of the fleet is J = {(i,j)|i∈I,j∈J i}. Process O ij The set of immediate predecessor processes is Ps ij , process O ij The maintenance work hours are d ij Maintenance process O ij The guaranteed start time is Sm ij , Maintenance process O ij The guarantee ends at Em ij Scheduling constraints include: the timing constraint of starting maintenance of a single aircraft, the maintenance process of each aircraft must be carried out in sequence according to the established maintenance process, the priority constraint of maintenance personnel conflicts, the priority constraint of maintenance equipment / workshop conflicts, the demand for skills in any process should match the number of maintenance personnel assigned to the process and who choose this skill, the demand for various types of maintenance equipment / workshops in any process should match the resources allocated to the process, each maintenance personnel can use at most one skill to perform maintenance work on any maintenance process, fixed support equipment such as power stations are limited by the length of the supply pipeline, workstation space resource constraints, and maintenance workshops can operate in parallel.

[0042] In some embodiments, the above method may further include step S1: initializing the generalized stand-alone maintenance process parameters and initializing the algorithm parameters of the improved teaching and learning algorithm.

[0043] Specifically, initialize the generalized stand-alone maintenance process parameters: ij represents the jth process of the i-th aircraft in the fleet to be repaired. Any maintenance process O ij The demand for the kth (k∈Kc) skill is rc ijk , then the matching relationship between the maintenance process and skills of single machine i can be expressed as the process-skill matrix Similarly, the flexibility of maintenance personnel to skills can be expressed by the personnel-skill matrix Indicates that the maintenance personnel set is Lp = {1,2,…,|Lp|}, the maintenance personnel skill configuration Kc, and the maintenance resource workshop equipment coverage index. Initialization algorithm parameters: class student size is Np, elite teacher group size is Nt, crossover probability is cr, and mutation probability is F. Figure 3 Shown is the multi-machine support flow chart and the process-skill-personnel correspondence diagram.

[0044] S2, according to the maintenance support data, adopts the priority number coding method to determine the start time of each process, and generates the process time matrix without considering resource constraints by left and right iteration; the process time matrix includes the earliest start time matrix, the latest start time matrix, the earliest completion time matrix and the latest completion time matrix.

[0045] It can be understood that the following processing steps may be specifically included:

[0046] S2.1, initialize time parameters, input process dependency matrix adm, corresponding maintenance mode working time matrix d, maintenance aircraft quantity NI and aircraft entry time matrix startime;

[0047] S2.2, based on the maintenance support data, forward scheduling and reverse scheduling are used to analyze the start time and completion time during the project network time analysis.

[0048] Specifically, in step S2.2, the forward scheduling process includes:

[0049] S2.2.1, execute the loop, set i=1;

[0050] S2.2.2, define the to-be-scheduled set matrix Pag, the scheduled set matrix Sg and the schedulable set matrix Dg;

[0051] S2.2.3, perform a cyclic search for all processes in the matrix Pag. If a current process in the current search satisfies the pre-order constraint, the current process is added to the matrix Dg and deleted from the matrix Pag; otherwise, the search for the next process is started;

[0052] S2.2.4, based on the serial scheduling iteration loop, find the earliest start time and loop until all maintenance tasks have been assigned start and end times; e.g. Figure 4 The figure shows a schematic diagram of the serial scheduling process.

[0053] S2.2.5, let i=i+1, jump to step S2.2.2, perform a loop search for the next aircraft, solve the earliest start time and earliest completion time of each aircraft respectively, and exit the loop search when i=NI;

[0054] S2.2.6, record the earliest start time matrix ES and the earliest completion time matrix EF.

[0055] In one embodiment, specifically, in step S2.2, the reverse scheduling process includes:

[0056] S2.3.1, execute the loop, set i = 1;

[0057] S2.3.2, after the parameters are reinitialized, define the matrix of the set to be scheduled Pag, the matrix of the set scheduled Sg and the matrix of the set schedulable Dg;

[0058] S2.3.3, perform a cyclic search for all the processes in the matrix Pag. If a current process in the current search satisfies the post-order constraint, the current process is added to the matrix Dg and deleted from the matrix Pag; otherwise, the search for the next process is started;

[0059] S2.3.4, based on the serial scheduling iterative loop, find the earliest start time and loop until all maintenance tasks have been assigned start and end times;

[0060] S2.3.5, let i=i+1, jump to step S2.3.2, perform a loop search for the next aircraft, solve the latest start time and the latest completion time of each aircraft respectively, and jump out of the loop search when i=NI;

[0061] S2.3.6, shift the process time to the far left, and record the latest start time matrix LS and the latest completion time matrix LF.

[0062] S3, generating a bit sequence list of the process based on the regret value deviation of the latest completion time matrix, and performing a decoding operation.

[0063] It can be understood that in step S3, the method used for generating the bit sequence table is based on the priority number encoding. This priority number represents the priority value of the start time of each process The sequence is the complete process of the aircraft, and the generation method is random.

[0064] In some implementations, the aforementioned operation process may specifically include the following processing steps:

[0065] S3.1, define the set of completed scheduling processes S g and the set of schedulable processes D g ; It can be understood that the decoding scheduling process is from D g Select the process according to the process priority, arrange it at the earliest possible start time, allocate personnel and equipment resources, and add it to the set S g This cyclic iterative process.

[0066] S3.2, let S g =∪ i∈I (i,1), initialize various resource status parameters, Sm i1 =Em i1 ; Sm i1 Indicates maintenance process O ij The guaranteed start time of Em i1 Indicates maintenance process O ij The end time of the coverage;

[0067] S3.3, Calculation Select process based on coding priority And calculate the process The earliest possible start time that satisfies the constraints of the predecessor process

[0068] S3.4, judging whether the idle maintenance personnel corresponding to the required skills, available maintenance equipment / workshop constraints and workstation space constraints are met at time t;

[0069] S3.5, if satisfied, record the maintenance sequence and Otherwise, set t=t+1 and return to execute step S3.4.

[0070] In some embodiments, in step S3.5, the matching judgment and allocation of maintenance personnel and tasks can be performed in the following steps:

[0071] S3.5.1, query the set of idle maintenance personnel Ap at the current time t, if This indicates that the available maintenance personnel do not meet the demand, so let t=t+1 and repeat step S3.5; otherwise, go to step S3.5.2.

[0072] S3.5.2, sort the skill set required for the process according to the shortage of maintenance personnel, that is, the fewer the number of available maintenance personnel with skill c, the higher the priority, so as to obtain the sorting C, and initialize n=1.

[0073] S3.5.3, corresponding skill is k * =C n , according to the matching matrix M P Find remaining available repair personnel with this skill like Then, according to the least skill first and least load first stratification rules, pre-allocate maintenance personnel, update the remaining available maintenance personnel, set n=n+1, and repeat step S3.5.3; otherwise, all previous pre-allocations are invalid, set t=t+1, and repeat step S3.5.

[0074] S3.5.4, if all skills and other constraints are met, update the personnel allocation

[0075] S3.6, if |S g |≥|J|, then the maintenance timing plan and the allocation plan of maintenance personnel and maintenance equipment / workshops are output, and the calculation efficiency index is the f value; where J represents the total maintenance process set of the fleet.

[0076] S4, conduct overall index evaluation on the position sequence table.

[0077] It can be understood that in step S4, the following processing steps may be specifically included:

[0078] S4.1, calculate the scheduling process time according to the scheduling process.

[0079] Specifically, the processing steps are as follows:

[0080] S4.1.1, resource initialization: such as the integration of all personnel, the support process arrangement set of the support group (the jth individual in the i-th profession) and the resource station, the integrated support personnel set-time matrix initialization and the aircraft cabin space-time set.

[0081] S4.1.2, calculate the cumulative remaining process time, loop through the resource station types to initialize the remaining cumulative process time matrix, and loop through the feasible sites of the resource stations.

[0082] S4.2, initialize the dispatch wave weight vector.

[0083] S4.3, based on the principle that the smaller the priority number, the higher the priority, find the support personnel required by the process and allocate and update the support personnel and space based on the time advancement of support personnel needs, the time advancement of space needs, and the time advancement of equipment needs.

[0084] Specifically, the following processes may be included:

[0085] S4.3.1, considering the timing constraints of a single aircraft starting maintenance, the aircraft needs to be transferred to the designated hangar parking position and tied up before maintenance support operations can be carried out, that is,

[0086] S4.3.2, determine that the maintenance procedures for each aircraft must be carried out in the order of the established maintenance process,

[0087] S4.3.3, considering the limited number of maintenance personnel and maintenance equipment / workshops, when different processes have requirements for the same resources, it is necessary to distinguish the order of maintenance according to priority. The priority constraint of maintenance personnel conflict is Among them, Sm ij Indicates maintenance process O ij The guaranteed start time, d ij Indicates maintenance process O ij The maintenance work hours, BM represents a sufficiently large real number, Yp ijeg Represents a Boolean variable. When the maintenance process O ij Maintenance process eg Assigned to the same maintenance personnel and process O ij Priority over process O eg The value is 1 when guaranteed, otherwise it is 0. Considering that there may be parallel operations in the workshop, the maintenance process O eg Not necessarily until process O ij The operation can only be started after the completion of the maintenance equipment / workshop conflict. The priority constraint is

[0088] S4.3.4, resource allocation constraints include the following three levels of constraints:

[0089] S4.3.4.1, the demand for skills in any process should match the number of maintenance personnel assigned to that process and who choose this skill, that is,

[0090] S4.3.4.2, the demand for various types of maintenance equipment / workshops in any process is matched with the resources allocated to the process, that is,

[0091] S4.3.4.3, each maintenance worker uses at most one of these skills to perform maintenance work on any maintenance process, i.e.

[0092] S4.3.5, fixed support equipment such as power supply stations can only provide support for aircraft within the length of the pipeline due to the length limitation of the supply pipeline, while mobile equipment and maintenance workshops can be regarded as fixed equipment with a coverage area covering the entire deck area, that is,

[0093] S4.3.6, with respect to workstation space resources, for any aircraft i, each type of workstation space can only accommodate a number of maintenance personnel to carry out corresponding maintenance operations at any time. Therefore

[0094] S4.3.7, any workshop can accommodate a certain number of maintenance tasks of corresponding professions for parallel maintenance. Therefore, the maintenance equipment can be regarded as a special maintenance workshop with a parallel operation capacity of 1, that is,

[0095] S4.4, calculate and return the efficiency index, maximize the fleet wave availability and minimize the maintenance personnel load variance; the efficiency index is f:

[0096] f=(1-WA)+αLBM

[0097] Among them, W represents the dispatch wave set, WA represents the fleet wave availability, α represents the weight coefficient of the maintenance personnel load variance, the maintenance dispatch commander can adjust the weight according to the support requirements, and LBM represents the maintenance personnel load variance;

[0098] Maximizing the fleet wave availability is maxWA:

[0099]

[0100] Among them, SW w represents the start time of preparation for the dispatch of wave w, v w Indicates the preference weight for wave availability. The earlier the wave, the more important it is to the battle situation, and the greater the weight it takes; ET i represents the maintenance completion time of aircraft i, Nm represents the number of sorties in this wave, and the function sgn(x) represents a step function, sgn(x) = 1 if and only if x ≥ 0, otherwise it is zero. It can be understood that WA defines the average availability of the subsequent sortie waves after the start of the fleet maintenance, and the goal of WA is to maximize the sum of the weighted availability in the sortie wave set.

[0101] Minimizing the maintenance personnel load variance is minLBM:

[0102]

[0103] Among them, TB l represents the total maintenance man-hours of the lth (l∈Lp) maintenance personnel, represents the mean maintenance time of maintenance personnel, Lp represents the set of maintenance personnel. LBM defines the variance of maintenance time of maintenance personnel.

[0104] S5, based on the overall indicator evaluation, calls the improved teaching and learning algorithm to perform left-hand operations to execute population evolution processing.

[0105] It can be understood that the teacher stage and the student stage are cycled, and the left operation performs the population evolution operation.

[0106] Specifically, the following processing steps may be included:

[0107] S5.1, optimal solution record initialized according to performance index;

[0108] S5.2, execute the teacher stage; wherein the teacher stage includes:

[0109] S5.2.1, select the individual with the best fitness as the teacher individual, and implement the DE / target-to-best / 1 teaching variation mechanism for the elite teacher group, that is, x i,G and v i,G are the i-th student individual and mutant individual in the G-th generation class, is a randomly selected individual from the elite teacher group. and are two individuals randomly selected from the student population, and F i is the scaling factor for the i-th student.

[0110] S5.2.2, perform the binomial crossover operation of the standard differential evolution algorithm, and use the crossover probability as an inducement to make the teacher parent population gene string cross with the population to generate a new population gene string. It can be understood that after the mutation operation, the same binomial crossover operation and greedy selection as the standard DE algorithm are performed. i and the crossover probability Cr i In the setting, an adaptive adjustment mechanism can be used. For any individual i in any iterative algebra G, the parameters are generated by sampling using the following distribution:

[0111]

[0112] Among them, Cauchy (M F ,0.1) and Gaussian(M Cr ,0.1) are respectively based on the mean value of M F and M Cr , random sampling from Cauchy and Gaussian distributions with variance 0.1. M F and M CrThe initialization values ​​are all 0.5. The parameter sets that have the function of improving fitness are defined as S F and S Cr , then M F and M Cr The parameter update mechanism is:

[0113]

[0114] Among them, ω F and ω Cr are all weight coefficients that obey the uniform distribution U(0.2,0.8), and the set S F and S Cr The mean value of can be calculated by the following formula:

[0115]

[0116] in, And the fitness improvement value is ΔRf m =|Rf m,old -Rf m,new |.

[0117] S5.2.3 performs a greedy selection operation, evaluates the fitness of the population individuals and sorts them in order. If the optimal fitness is updated, the optimal individual is replaced as the new teacher individual. Specifically, the fitness evaluation of the population individuals is f = (1-WA) + αLBM.

[0118] S5.3, executing the student stage; wherein the student stage includes the following processing:

[0119] S5.3.1, based on the binary tournament, the teacher individual parent is selected and the excellent individuals in the population are subjected to single-point crossover to form two offspring generations.

[0120] S5.3.2, perform independent mutation operations on two offspring and perform peak crossover, calculate offspring fitness and select the best one to be inherited to the next generation; output the historical optimal solution to replace the initialized optimal solution record;

[0121] Specifically, the resource utilization rate (RUR) is used as the guiding factor. Given a maintenance plan S, the instantaneous integrated resource utilization rate (IRUR) at time t is defined as:

[0122]

[0123] in, and They represent the weights of the resource type k of maintenance personnel, maintenance equipment / workshop, and workstation space, respectively, and are calculated based on the ratio of the number of processes required for resources / skills to the total number of processes. For example:

[0124]

[0125] Given a maintenance plan S, starting at time t and with an interval length of l, the interval integrated resource utilization is defined as:

[0126]

[0127] The interval with the highest comprehensive resource utilization (i.e., peak interval) can be expressed as [tp1(S), tp2(S)], and the starting time of the peak interval with a length of l is given by For any student individual x i , select the individual x to be crossed j (j≠i), two individuals take turns as the question-answering student x m and ask student role x l , and its scheduling solutions are expressed as S m and S l , the questioning student learns from the answering student. The individuals generated after crossover are represented as x s , the crossover process is: from the interval [(1 / 4)C max (x m ),(3 / 4)C max (x m )] Random sampling length l, calculate the number of students answering questions x m The peak starting time tp1.

[0128] The intersection of two points can be expressed as: if x l,i ∈<tp1(S m ), then let x s,i :=x l,i -BM; if tp1(S m )≤x l,i ≤tp2(S m ), then let x s,i :=x m,i ; if x l,i >tp2(S m ), then let x s,i :=x l,i +BM. Select two crossover offspring with better fitness to replace the original student individual x i . Calculate the fitness of offspring and select the best ones to be inherited to the next generation. Output the historical optimal solution to replace the initial optimal solution.

[0129] S5.4, record the number of evaluations and return to step S5.1.

[0130] S6, performs population evolution processing based on the left operation and performs the processing of the reinforcement learning stage.

[0131] Specifically, the following processing may be included:

[0132] S6.1, perform exchange neighborhood processing; the interactive neighborhood is to randomly select two maintenance processes without tight preceding or following constraints and perform coding priority number exchange;

[0133] S6.2, performing a fault repair process front insertion neighborhood processing; the fault repair process front insertion neighborhood processing includes: randomly selecting a fault repair process, querying the end time of the selected fault repair process and the immediately preceding process as the starting value of the insertable interval, the start time of the selected fault repair process as the end time of the interval, and inserting the selected fault repair process into any position in the interval;

[0134] S6.3, perform local disaster neighborhood processing of scheduled inspection process. Specifically, select any aircraft i in the fleet to be scheduled for inspection, and then randomly select Na (Na∈Z[4,9]) maintenance processes from it, and perform the catastrophic neighborhood processing of any selected process O ij , calculate its insertable time window, where the start time is process O ij The latest end time te of the predecessor process set ij The end time of the time window is the earliest start time ts of the next process set ij , and randomly insert the process into any position in the time window. This neighborhood is mainly used for medium-scale disturbances in scheduled inspection and maintenance processes.

[0135] S7, compare the target evaluation times obtained in the reinforcement learning stage with the limit value. If the target evaluation times are greater than the limit value, output the final generated left scheduling plan; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

[0136] It can be understood that the final evaluation number Q is compared. If Q < the limit value, then go to step S5. If Q > the limit value, then jump out of the loop and output the final generated left scheduling plan. The limit value can be set based on the experience of the convergence iteration number of the metaheuristic algorithm and the comprehensive consideration of the actual computing power of the computing device. In this application, the limit value can be taken as 10,000 times. Figure 5 As shown, a Gantt chart for maintenance personnel scheduling is obtained.

[0137] The above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method adopts an improved teaching and learning algorithm to optimize the calculation and processing of the scheduling plan based on real maintenance support data. Compared with the traditional scheduling method at this stage, it includes the optimization research goal of the hangar maintenance support task of the fleet, focusing on minimizing the maintenance completion time, and considering the availability requirements of the fleet wave dispatch, considering the multi-mode mixing of fault repair and scheduled inspection maintenance, and comprehensively considering the actual constraints such as the maintenance support scope, maintenance parallel capacity, and maintenance station space. It can automatically and efficiently output reliable aircraft hangar scheduled inspection and maintenance scheduling plans to meet the needs of refined maintenance tasks and achieve the purpose of significantly improving the efficiency of aircraft hangar scheduled inspection and maintenance scheduling.

[0138] It should be understood that although Figure 1 , Figure 2 and Figure 4 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Figure 1 , Figure 2 and Figure 4 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequentially, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0139] See also Figure 6, and also provides an aircraft hangar scheduled inspection and maintenance scheduling device 100, including a data acquisition module 11, a time matrix module 12, a list decoding module 13, an index evaluation module 14, a population evolution module 15, a reinforcement learning module 16 and a scheduling output module 17. Among them, the data acquisition module 11 is used to obtain the aircraft maintenance and support data; the maintenance and support data includes the number of aircraft, the time of entry, the single-machine maintenance process, the maintenance personnel and the skill data. The time matrix module 12 is used to determine the start time of each process based on the maintenance and support data by using a priority number coding method, and iterate left and right to generate a process time matrix that does not consider resource constraints; the process time matrix includes the earliest start time matrix, the latest start time matrix, the earliest completion time matrix and the latest completion time matrix. The list decoding module 13 is used to generate a bit sequence list of the process based on the regret value deviation of the latest completion time matrix, and perform a decoding operation. The index evaluation module 14 is used to perform an overall index evaluation on the bit sequence list. The population evolution module 15 is used to call the improved teaching and learning algorithm to perform a left operation and perform population evolution processing based on the overall index evaluation. The reinforcement learning module 16 is used to perform population evolution processing based on the left operation and perform the processing of the reinforcement learning stage. The scheduling output module 17 is used to compare the target evaluation times obtained by the processing of the reinforcement learning stage with the limit value. If the target evaluation times are greater than the limit value, the final generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

[0140] The above-mentioned aircraft hangar scheduled inspection and maintenance scheduling device 100, through the collaboration of various modules, adopts an improved teaching and learning algorithm, and performs optimization calculation and processing of the scheduling plan based on real maintenance and support data. Compared with the traditional scheduling methods at this stage, it includes the optimization research goal of the hangar maintenance and support tasks of the fleet, focusing on minimizing the maintenance completion time, and considering the availability requirements of the fleet wave dispatch, considering the multi-mode mixing of fault repair and scheduled inspection and maintenance, and comprehensively considering the actual constraints such as the maintenance support scope, maintenance parallel capacity, and maintenance station space. It can automatically and efficiently output reliable aircraft hangar scheduled inspection and maintenance scheduling plans to meet the needs of refined maintenance tasks and achieve the purpose of significantly improving the efficiency of aircraft hangar scheduled inspection and maintenance scheduling.

[0141] In one embodiment, each module of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling device 100 can also be used to implement the corresponding sub-steps in each embodiment of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method.

[0142] Regarding the specific limitations of the aircraft hangar scheduled inspection and maintenance scheduling device 100, please refer to the corresponding limitations of the aircraft hangar scheduled inspection and maintenance scheduling method above, which will not be repeated here. The various modules in the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling device 100 can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of a device with a specific data processing function in the form of hardware, or can be stored in the memory of the aforementioned device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. The aforementioned device can be, but is not limited to, a computer device or a computing system used for aircraft hangar scheduling management.

[0143] On the other hand, a computer device is also provided, including a memory and a processor, the memory stores a computer program, and the processor can implement the following steps when executing the computer program: obtaining aircraft maintenance and support data; the maintenance and support data include the number of aircraft, entry time, single-machine maintenance process, maintenance personnel and skill data; according to the maintenance and support data, the start time of each process is determined by a priority number coding method, and the process time matrix that does not consider resource constraints is generated by left and right iteration; the process time matrix includes the earliest start time matrix, the latest start time matrix, the earliest completion time matrix and the latest completion time matrix; based on the regret value deviation of the latest completion time matrix, a bit sequence list of the process is generated, and a decoding operation is performed; an overall indicator evaluation is performed on the bit sequence list; based on the overall indicator evaluation, an improved teaching and learning algorithm is called to perform a left-hand operation to perform population evolution processing; based on the left-hand operation, the population evolution processing is performed and the processing of the reinforcement learning stage is performed; the target evaluation times obtained by the processing of the reinforcement learning stage is compared with the limit value, and if the target evaluation times is greater than the limit value, the final generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

[0144] In one embodiment, when the processor executes the computer program, it can also implement the steps or sub-steps added in each embodiment of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method.

[0145] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: obtaining aircraft maintenance and support data; the maintenance and support data include the number of aircraft, entry time, single-aircraft maintenance process, maintenance personnel and skill data; according to the maintenance and support data, the start time of each process is determined by a priority number coding method, and a process time matrix that does not consider resource constraints is generated by left and right iteration; the process time matrix includes an earliest start time matrix, a latest start time matrix, an earliest completion time matrix and a latest completion time matrix; based on the regret value deviation of the latest completion time matrix, a bit sequence list of the process is generated, and a decoding operation is performed; an overall indicator evaluation is performed on the bit sequence list; based on the overall indicator evaluation, an improved teaching and learning algorithm is called to perform a left-hand operation to perform population evolution processing; based on the left-hand operation, the population evolution processing is performed and the processing of the reinforcement learning stage is performed; the target evaluation times obtained by the processing of the reinforcement learning stage is compared with a limit value, and if the target evaluation times is greater than the limit value, the final generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

[0146] In one embodiment, when the computer program is executed by a processor, it can also implement the steps or sub-steps added in each embodiment of the above-mentioned aircraft hangar scheduled inspection and maintenance scheduling method.

[0147] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus dynamic random access memory (Rambus DRAM, referred to as RDRAM) and interface dynamic random access memory (DRDRAM).

[0148] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0149] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and all of them belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for scheduling scheduled inspection and maintenance of an aircraft hangar, characterized in that: Includes steps: Obtaining aircraft maintenance and support data; the maintenance and support data includes the number of aircraft, time of entry, single-aircraft maintenance process, maintenance personnel and skill data; According to the maintenance support data, the start time of each process is determined by using a priority number coding method, and a process time matrix without considering resource constraints is generated by left and right iteration; the process time matrix includes an earliest start time matrix, a latest start time matrix, an earliest completion time matrix and a latest completion time matrix; Generate a bit sequence list of processes based on the regret value deviation of the latest completion time matrix, and perform a decoding operation; Performing overall index evaluation on the bit sequence table; Based on the overall indicator evaluation, calling the improved teaching and learning algorithm to perform left-hand operation to execute population evolution processing; Performing population evolution processing based on the left-hand operation and performing processing in the reinforcement learning phase; The target evaluation times obtained by the reinforcement learning stage are compared with the limit value. If the target evaluation times are greater than the limit value, the finally generated left scheduling plan is output; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar; wherein the step of performing overall index evaluation on the bit sequence table includes: Calculate the scheduling process time according to the scheduling process; Initialize the dispatch wave weight vector; Based on the principle that the smaller the priority number, the higher the priority, find the support personnel required by the process and allocate and update the support personnel and update the space based on the time advancement of the support personnel requirements, the time advancement based on the space requirements, and the time advancement based on the equipment requirements; Calculate and return performance indicators to maximize fleet wave availability and minimize maintenance personnel load variance; The performance index is f: f=(1-WA)+αLBM Among them, W represents the dispatch wave set, WA represents the fleet wave availability, α represents the weight coefficient of the maintenance personnel load variance, and LBM represents the maintenance personnel load variance; The maximum fleet wave availability is maxWA: Wherein, SWw represents the start time of sortie preparation of wave w, vw represents the preference weight for wave availability, ETi represents the maintenance completion time of aircraft i, Nm represents the number of sorties in this wave, and function sgn(x) represents a step function, sgn(x) = 1 if and only if x ≥ 0, otherwise it is zero; The minimum maintenance personnel load variance is minLBM: Among them, TB l represents the total maintenance man-hours of the lth (l∈Lp) maintenance personnel, represents the mean maintenance man-hours of the maintenance personnel, and Lp represents the set of maintenance personnel.

2. The aircraft hangar scheduled inspection and maintenance scheduling method according to claim 1, characterized in that: Also includes: Initialize the generalized single-machine maintenance process parameters and initialize the algorithm parameters of the improved teaching and learning algorithm.

3. The aircraft hangar scheduled inspection and maintenance scheduling method according to claim 1 or 2, characterized in that: The steps of determining the start time of each process according to the maintenance support data by using a priority number coding method, and iterating left and right to generate a process time matrix without considering resource constraints include: Initialize time parameters, input process dependency matrix adm, corresponding maintenance mode working time matrix d, maintenance aircraft quantity NI and aircraft entry time matrix startime; According to the maintenance guarantee data, forward scheduling and reverse scheduling are used to analyze the start time and completion time in the process of project network time analysis; The forward scheduling process includes: Execute the loop, set i=1; Define the to-be-scheduled set matrix Pag, the scheduled set matrix Sg and the schedulable set matrix Dg; Perform a cyclic search for all the processes in the matrix Pag. If a current process currently being searched satisfies the pre-order constraint, the current process is added to the matrix Dg and deleted from the matrix Pag; otherwise, the search for the next process is performed; Based on the serial scheduling iterative loop, find the earliest start time and loop until all maintenance tasks have been assigned start and end times; Let i=i+1, jump to the step of defining the to-be-scheduled set matrix Pag, the scheduled set matrix Sg and the schedulable set matrix Dg, perform a cyclic search for the next aircraft, and respectively solve the earliest start time and the earliest completion time of each aircraft until the cyclic search is exited when i=NI; Recording the earliest start time matrix and the earliest completion time matrix; The reverse scheduling process includes: Execute the loop, set i=1; After the parameters are reinitialized, the matrix of the set to be scheduled Pag, the matrix of the set scheduled Sg and the matrix of the set schedulable Dg are defined; Perform a cyclic search for all the processes in the matrix Pag. If a current process in the current search satisfies the post-order constraint, the current process is added to the matrix Dg and deleted from the matrix Pag; otherwise, the search for the next process is performed; Based on the serial scheduling iterative loop, find the earliest start time and loop until all maintenance tasks have been assigned start and end times; Let i=i+1, jump to the step of reinitializing the parameters, define the to-be-scheduled set matrix Pag, the scheduled set matrix Sg and the schedulable set matrix Dg, perform a cyclic search for the next aircraft, and respectively solve the latest start time and the latest completion time of each aircraft until i=NI and exit the cyclic search; Shift the process time to the leftmost end, and record the latest start time matrix and the latest completion time matrix.

4. The aircraft hangar scheduled inspection and maintenance scheduling method according to claim 3, characterized in that: The step of generating a bit sequence list of a process based on the regret value deviation of the latest completion time matrix and performing a decoding operation comprises: Define the set of completed scheduling operations S g and the set of schedulable processes D g ; Let S g =∪ i∈I (i,1), initialize various resource status parameters, Sm i1 =Em i1 ; Sm i1 Indicates maintenance process O ij The guarantee start time, Emi1 represents the maintenance process O ij The end time of the coverage; calculate Select process based on coding priority And calculate the process The earliest possible start time that satisfies the constraints of the predecessor process Determine whether the idle maintenance personnel corresponding to the required skills, available maintenance equipment / workshop constraints and workstation space constraints are met at the time t; If satisfied, record the maintenance sequence and Otherwise, let t=t+1 and return to the step of determining whether the idle maintenance personnel corresponding to the required skills, the available maintenance equipment / workshop constraints, and the workstation space constraints are met at the time t; If |S g |≥|J|, then the maintenance timing plan and the allocation plan of maintenance personnel and maintenance equipment / workshops are output, and the calculation efficiency index is the f value; where J represents the total maintenance process set of the fleet.

5. The aircraft hangar scheduled inspection and maintenance scheduling method according to claim 1, characterized in that: Based on the overall indicator evaluation, the step of calling the improved teaching and learning algorithm to perform left-hand operation and execute population evolution processing includes: An optimal solution record initialized according to the performance indicator; Executive teacher stage; The teacher stage includes: The individual with the best fitness is selected as the teacher individual, and the DE / target-to-best / 1 teaching variation mechanism for the elite teacher group is implemented; Execute the binomial crossover operation of the standard differential evolution algorithm, use the crossover probability as an inducement, and make the teacher parent population gene string cross with the population to generate a new population gene string; Execute the greedy selection operation, evaluate the fitness of the individuals in the population and sort them in sequence. If the optimal fitness is updated, replace the optimal individual with the new teacher individual. Execute the student phase; The student stage includes: Based on the binary tournament, the teacher individual parent is selected and the excellent individuals in the population are cross-pollinated to form two offspring; Perform independent mutation operations on two offspring and perform peak crossover, calculate offspring fitness and select the best one to be inherited to the next generation; output the historical optimal solution to replace the initialized optimal solution record; The step of recording the number of evaluations and returning the optimal solution record initialized according to the performance indicator.

6. The aircraft hangar scheduled inspection and maintenance scheduling method according to claim 5, characterized in that: The step of performing population evolution processing based on the left operation and performing processing in the reinforcement learning phase includes: Performing an exchange neighborhood process; the exchange neighborhood is to randomly select two maintenance processes without tight preceding or following constraints and exchange the coding priority numbers; Performing a fault repair process front insertion neighborhood processing; the fault repair process front insertion neighborhood processing includes: randomly selecting a fault repair process, querying the end time of the selected fault repair process and the immediately preceding process as the starting value of the insertable interval, the start time of the selected fault repair process is the end of the interval, and inserting the selected fault repair process into any position in the interval; Carry out local disaster neighborhood processing during regular inspection procedures.

7. An aircraft hangar scheduled inspection and maintenance scheduling device, characterized in that: include: A data acquisition module is used to acquire aircraft maintenance and support data; the maintenance and support data includes the number of aircraft, time of entry, single-aircraft maintenance process, maintenance personnel and skill data; A time matrix module is used to determine the start time of each process according to the maintenance support data by using a priority number coding method, and to iterate left and right to generate a process time matrix that does not consider resource constraints; the process time matrix includes an earliest start time matrix, a latest start time matrix, an earliest completion time matrix, and a latest completion time matrix; A list decoding module, used to generate a bit sequence list of processes based on the regret value deviation of the latest completion time matrix, and perform a decoding operation; The index evaluation module is used to perform an overall index evaluation on the bit sequence table; wherein, it includes calculating the scheduling process time according to the scheduling process, initializing the dispatch wave weight vector, finding the support personnel required by the process based on the principle that the smaller the priority number, the greater the priority, and allocating and updating the support personnel and updating the space based on the time advancement of the support personnel demand, the time advancement based on the space demand, and the time advancement based on the equipment demand, calculating and returning the efficiency index, maximizing the fleet wave availability and minimizing the maintenance personnel load variance; the efficiency index is f: f=(1-WA)+αLBM Among them, W represents the dispatch wave set, WA represents the fleet wave availability, α represents the weight coefficient of the maintenance personnel load variance, and LBM represents the maintenance personnel load variance; The maximum fleet wave availability is maxWA: Wherein, SWw represents the start time of sortie preparation of wave w, vw represents the preference weight for wave availability, ETi represents the maintenance completion time of aircraft i, Nm represents the number of sorties in this wave, and function sgn(x) represents a step function, sgn(x) = 1 if and only if x ≥ 0, otherwise it is zero; The minimum maintenance personnel load variance is minLBM: Among them, TB l represents the total maintenance man-hours of the lth (l∈Lp) maintenance personnel, represents the mean maintenance man-hours of maintenance personnel, Lp represents the set of maintenance personnel; A population evolution module, used for invoking an improved teaching and learning algorithm to perform a left-hand operation and execute population evolution processing based on the overall indicator evaluation; A reinforcement learning module, used to perform population evolution processing based on the left-hand operation and to perform processing in the reinforcement learning phase; The scheduling output module is used to compare the target evaluation times obtained by the reinforcement learning stage with the limit value, and if the target evaluation times are greater than the limit value, output the final generated left scheduling plan; the left scheduling plan is used to indicate the scheduling plan for scheduled inspection and maintenance of the aircraft hangar.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the aircraft hangar scheduled inspection and maintenance scheduling method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the aircraft hangar scheduled inspection and maintenance scheduling method described in any one of claims 1 to 6 are implemented.

Images