Aircraft in-and-out transfer scheduling method, device and equipment based on Grey Wolf algorithm

Abstract

The present application relates to a method, device and equipment for aircraft in-and-out transfer scheduling based on the gray wolf algorithm. The method includes: obtaining the current in-and-out transfer task parameters; calling a pre-built scheduling model; initializing the gray wolf population according to the in-and-out transfer task parameters, selecting α wolf, β wolf or δ wolf in the gray wolf population by roulette to perform neighborhood optimization, obtaining neighborhood gray wolf individuals, and updating the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individuals, α wolf, β wolf and δ wolf; iteratively updating the positions of α wolf, β wolf and δ wolf, and when the pre-selected iteration stop condition is met, stopping the iteration and outputting the current in-and-out transfer scheduling plan; the in-and-out transfer scheduling plan is used for aircraft in-and-out transfer scheduling. The use of this method can improve the update efficiency of the optimal solution, enhance the algorithm's optimization ability, and thus efficiently execute fleet in-and-out transfer scheduling.

Description

Technical Field

[0001] The present application relates to the technical field of aircraft transfer scheduling, and in particular to a method, device and equipment for aircraft in-and-out transfer scheduling based on a grey wolf algorithm. Background Art

[0002] During missions, fleets of aircraft undergo numerous deployment and recovery cycles. Due to limitations in offshore platform design and support capabilities, different support processes for a single aircraft often require different parking positions. This necessitates multiple transfers between platform parking positions and hangar parking positions. Whether preparing for immediate deployment or preparing for re-deployment, the large number of aircraft entering and exiting the hangar and platform transfers places significant pressure on the confined offshore platform. Furthermore, with the continued advancement of unmanned aerial vehicles (UAVs) and tractors, they will become an integral part of the offshore platform operating environment, adding significant complexity to an already highly tense and dangerous environment. The two fundamental elements of development, progress, and safety and stability place greater demands on the automation of the planning and scheduling of aircraft, ground vehicles, and personnel. Summary of the Invention

[0003] Based on this, it is necessary to provide an aircraft in-and-out transfer scheduling method, device and equipment based on the Grey Wolf algorithm to address the above technical problems.

[0004] An aircraft in-and-out transfer scheduling method based on the Grey Wolf algorithm, the method comprising:

[0005] Obtaining current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred;

[0006] Invoking a pre-built scheduling model; the constraints of the scheduling model include equipment allocation exclusivity constraints, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group;

[0007] Initialize the gray wolf population according to the inbound and outbound transfer task parameters, select α wolf, β wolf or δ wolf from the gray wolf population by using a roulette wheel method to perform neighborhood optimization to obtain a neighborhood gray wolf individual, and update the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individual, α wolf, β wolf and δ wolf;

[0008] Iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, stop the iteration and output the inbound and outbound transfer scheduling plan represented by the current α wolf; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0009] An aircraft in-and-out transfer scheduling device based on the Grey Wolf algorithm, the device comprising:

[0010] A parameter acquisition module is used to obtain the current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred;

[0011] A model calling module is used to call a pre-built scheduling model; the constraints of the scheduling model include exclusive constraints on equipment allocation, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group;

[0012] A population update module is used to initialize the gray wolf population according to the inbound and outbound transfer task parameters, select α wolves, β wolves or δ wolves from the gray wolf population by a roulette wheel method to perform neighborhood optimization to obtain neighborhood gray wolf individuals, and update the positions of α wolves, β wolves and δ wolves according to the fitness function values ​​corresponding to the neighborhood gray wolf individuals, α wolves, β wolves and δ wolves;

[0013] The transfer scheduling module is used to iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, the iteration is stopped and the inbound and outbound transfer scheduling plan represented by the current α wolf is output; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0014] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0015] Obtaining current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred;

[0016] Invoking a pre-built scheduling model; the constraints of the scheduling model include equipment allocation exclusivity constraints, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group;

[0017] Initialize the gray wolf population according to the inbound and outbound transfer task parameters, select α wolf, β wolf or δ wolf from the gray wolf population by using a roulette wheel method to perform neighborhood optimization to obtain a neighborhood gray wolf individual, and update the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individual, α wolf, β wolf and δ wolf;

[0018] Iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, stop the iteration and output the inbound and outbound transfer scheduling plan represented by the current α wolf; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0019] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0020] Obtaining current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred;

[0021] Invoking a pre-built scheduling model; the constraints of the scheduling model include equipment allocation exclusivity constraints, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group;

[0022] Initialize the gray wolf population according to the inbound and outbound transfer task parameters, select α wolf, β wolf or δ wolf from the gray wolf population by using a roulette wheel method to perform neighborhood optimization to obtain a neighborhood gray wolf individual, and update the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individual, α wolf, β wolf and δ wolf;

[0023] Iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, stop the iteration and output the inbound and outbound transfer scheduling plan represented by the current α wolf; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0024] The above-mentioned aircraft in-and-out transfer scheduling method, device and equipment based on the gray wolf algorithm, by calling a pre-built scheduling model, coordinates the in-and-out storage and platform transfer between the platform and the hangar, and discretizes the gray wolf hunting process of the gray wolf algorithm with a simple structure to solve the complex scheduling model, specifically performing neighborhood optimization on the α wolf, β wolf or δ wolf in the gray wolf population to obtain the neighborhood gray wolf individuals, and updating the positions of the α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individuals, α wolf, β wolf and δ wolf, iterating this process to output the optimal in-and-out transfer scheduling plan, determine the transfer timing and transfer resource allocation, and improve the transfer efficiency. The embodiment of the present invention can improve the update efficiency of the optimal solution, enhance the optimization ability of the algorithm, and thus efficiently execute the in-and-out transfer scheduling of the fleet. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a process scenario diagram of an aircraft in-and-out transfer scheduling method based on the Grey Wolf algorithm in one embodiment;

[0026] Figure 2 A schematic diagram of a hunting process of a gray wolf pack in one embodiment;

[0027] Figure 3 A schematic diagram of a field re-insertion operation in one embodiment;

[0028] Figure 4 A schematic diagram of a two-point crossover operation in another embodiment;

[0029] Figure 5 A schematic diagram of a field reverse operation in one embodiment;

[0030] Figure 6 1 is a flow chart of an IGWO algorithm according to an embodiment;

[0031] Figure 7 Schematic diagram of the changing trend of optimization indicators for 12-machine dispatch transfer scheduling in another embodiment, wherein (a) is a schematic diagram of the changing trend of transfer completion time, (b) is a schematic diagram of the changing trend of transfer group load variance, and (c) is a schematic diagram of the changing trend of transfer group transfer time;

[0032] Figure 8 This is a Gantt chart for scheduling aircraft fleet in and out of storage and deck transfer under the optimal plan for deploying 12 aircraft in another embodiment;

[0033] Figure 9 A Gantt chart showing the scheduling of transfer equipment under the optimal plan for deploying 12 machines in one embodiment;

[0034] Figure 10 Schematic diagram of the comparison of the convergence curve of the completion time of the 12-machine dispatch and transfer scheduling in one embodiment;

[0035] Figure 11 1 is a structural block diagram of an aircraft in-and-out transfer scheduling device based on the Grey Wolf algorithm in one embodiment;

[0036] Figure 12 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0038] Aircraft platform in-and-out parking and platform transfers primarily involve three phases of the aircraft mission process: direct dispatch preparation, re-dispatch preparation, and aircraft return to storage. Therefore, the Fleet Transfer Scheduling for Carrier Aircraft (FTSCA) problem involves three flows: after completing a mission, the fleet returns to the platform's head temporary parking stand. After recovery, different transfer operations are performed based on the subsequent mission requirements of each aircraft. Platform transfers occur from the platform's head temporary parking stand to the platform stand; in-hangar transfers occur from the platform's head temporary parking stand to the hangar stand; and when an aircraft in the hangar is assigned a mission, out-hangar transfers occur from the hangar stand to the platform stand. Because transfers between the offshore platform and the hangar require the use of elevators and tractors, these three transfer processes interfere with each other in terms of resources and space. Therefore, coordinated planning of aircraft in-and-out parking and platform transfers is necessary to determine the transfer sequence and resource allocation to improve transfer efficiency.

[0039] In one embodiment, Figure 1 As shown, a method for scheduling aircraft inbound and outbound transfers based on the Grey Wolf algorithm is provided, comprising the following steps:

[0040] Step 102: Obtain the current inbound and outbound transfer task parameters.

[0041] Parameters for inbound and outbound transfer tasks include the number of aircraft to be transferred and their aircraft numbers. They also include the number of transfer equipment and transfer groups. Transfer equipment includes platform tractors, hangar tractors, and elevators. Transfer groups are responsible for platform tractor transfer groups, hangar tractor transfer groups, and elevator transfer groups. Each type of transfer group is grouped according to the corresponding number of transfer equipment.

[0042] Step 104: call the pre-built scheduling model.

[0043] To solve the DSPMF problem, the following assumptions are made: (1) The operation process of the transfer equipment at any stage cannot be interrupted, and other tasks can only be accepted after the current transfer task is completed; (2) Whether the aircraft is selected to transfer equipment or the transfer equipment transfers the aircraft, the unique principle must be followed. At any stage, the aircraft can only be transferred by one transfer equipment, and each transfer equipment can only transfer one aircraft; (3) The transfer equipment at a certain stage is a parallel machine, which can meet the transfer needs of any aircraft, but the time cost of the transfer equipment in place must be considered; (4) The transfer path library is a static library, and all parameters in the path are constant values; (5) This scheduling is a static scheduling study and does not involve dynamic interference.

[0044] The constraints of the scheduling model include the exclusivity constraint of equipment allocation, the aircraft transfer timing constraint, the aircraft transfer collision avoidance constraint and the transfer equipment transfer guarantee timing constraint; the objective function of the scheduling model includes the minimum fleet transfer completion time, the minimum load balancing index and the minimum transfer time of the transfer group.

[0045] Specifically, the objective functions of the scheduling model include:

[0046] Minimize transfer completion time: Due to the complex aircraft dispatch process and complicated procedures, the faster the aircraft transfer is completed, the more time can be saved for subsequent processes. The minimum transfer completion time is:

[0047]

[0048] Among them, the time for the i-th aircraft to arrive at the target parking position is Indicates that the fleet transfer completion time C max Indicates the maximum completion time of fleet transfer, that is, the time when the last aircraft completes the transfer.

[0049] Minimize the load variance of the transfer group: In actual transfer support, in order to balance the workload of each transfer equipment and transfer personnel and ensure that personnel are fully energized, it is necessary to reasonably allocate transfer tasks. Therefore, the sum of the load variances of the transfer group (LBT) is used as the load balancing indicator. The minimum load variance of the transfer group is:

[0050]

[0051] Among them, γ k represents the weight of the kth (k∈kz) transport group, TB kl is the kth(k∈kz)th class lth(l∈lz k ) transfer group load, which is the weighted sum of the aircraft transfer time and the tractor transfer support time, is calculated as follows:

[0052]

[0053] Among them, TR(P ij ,P i(j+1) )for is the weight of the kth (k∈kz) transport group, is the load average of the group.

[0054] Minimize the transfer time of the transfer group: Since the present invention does not consider the interference of the transfer equipment on the aircraft transfer, the transfer time of the transfer equipment is minimized as much as possible to reduce its disturbance to the platform plane and ensure smooth aircraft transfer. The minimum transfer time of the transfer group is:

[0055]

[0056] In the transshipment operation, the transshipment completion time is taken as the first optimization objective, and the load variance and the transshipment group transfer time are determined as the second optimization objective or the third optimization objective according to the optimization requirements.

[0057] The constraints of the scheduling model include:

[0058] Uniqueness constraint: At any stage, an aircraft can only be supported by one transfer equipment. The uniqueness constraint is:

[0059]

[0060] Aircraft transfer timing constraints: The time constraints for transferring an aircraft from the initial parking position to the target parking position. When the transfer equipment is an elevator, the one-way lifting time te of the elevator must be considered. Each time the aircraft enters and exits the parking position, the aircraft mooring and unmooring time tx must be considered. The aircraft transfer timing constraints are:

[0061]

[0062]

[0063] Aircraft transfer collision avoidance constraints: The occurrence of a transfer collision can be viewed as the intersection of space and time. Since the path library of the present invention is established, when an aircraft collision occurs, a priority-based delayed departure strategy is used to avoid collisions. First, there must be a feasible path from an initial parking position to a target parking position. The aircraft transfer collision avoidance constraints are:

[0064]

[0065] Among them, Road t (P ij ,P i(j+1) )for The abbreviation of represents the spatial occupancy point set of the i-th aircraft in the reference deployment state q, and the multi-aircraft transport collision avoidance constraint can be expressed as:

[0066]

[0067] The time constraint for transferring equipment must be considered. This cost is also the main basis for aircraft to select support equipment during decoding. The time constraint for transferring equipment is:

[0068]

[0069] Decision variable constraints: Xz ijl 、Yz ijeg are all 0-1 decision variables, namely:

[0070]

[0071] The scheduling model parameters are defined as follows:

[0072] I is the set of aircraft to be transferred, I = {1, 2, ..., Nf};

[0073] Jz i The mission set for aircraft i, Jz i ={1,2,…,Nz i};

[0074] Jz i ′ is the set of aircraft i tasks that include the virtual end task, Jz i ′=Jz i ∪{Nz i +1};

[0075] Nz i is the number of transfer stages for aircraft i, and Nz i ≤3;

[0076] O ij is the transfer process j of aircraft i;

[0077] P ij is the initial position of aircraft i in stage j;

[0078] M ij The set of available transfer equipment for aircraft i at stage j;

[0079] Kz is the set of transport group categories;

[0080] Lz k is the set of transport groups of type k, Lz k ={1,2,…,|Lz k |};

[0081]

[0082] TR(p,q,Ap t (p,q)) is the transit time from parking position p to q at time t;

[0083] tx is the tethering / untethering time;

[0084] For the kth type of equipment, process O ij After completion, the parking position is transferred to process O eg Time at the parking stand before the start;

[0085] BM is a maximum constant used for penalty term;

[0086] The decision variables are defined as follows:

[0087] Sz ij For the transfer process O ij The start time of the operation;

[0088] Ez ij For the transfer process O ij The end time of the job;

[0089]

[0090]

[0091] Based on the above modeling analysis, although FTSCA can be abstracted as a hybrid flow shop scheduling problem, it is more complex. First, the start time of aircraft transfer is not only constrained by the transfer equipment, as in the hybrid flow shop scheduling problem, but also needs to consider whether there is a feasible path during the transfer process and whether there will be collisions along the path. Second, in the hybrid flow shop problem, the processes of different workpieces do not interfere with each other and the flow directions are consistent. However, in the FTSCA problem of the present invention, the aircraft transfer flow directions are not unique, and interference may occur between the various stages of different aircraft. For example, the third stage of an outbound aircraft and the first stage of an inbound aircraft both occur on the platform, and the two may collide. To efficiently solve the complex FTSCA problem, the present invention improves the Grey Wolf Optimization (GWO) algorithm. Compared with other swarm intelligence algorithms, this algorithm has a simpler structure, fewer parameters, and is relatively easy to implement. Moreover, as the algorithm continues to iterate, it can provide continuous feedback on the optimization results and adaptively adjust the convergence factor. This achieves a good balance between local and global optimization, and has good solution speed and accuracy. The present invention designs an improved Grey Wolf Algorithm (IGWO) based on the standard GWO algorithm and targeting the scheduling model characteristics of FTSCA.

[0092] Step 106, initialize the gray wolf population according to the in-and-out transfer task parameters, use the roulette wheel method to select α wolf, β wolf or δ wolf in the gray wolf population for neighborhood optimization, obtain the neighborhood gray wolf individual, and update the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individual, α wolf, β wolf and δ wolf.

[0093] There is a strict social hierarchy in the gray wolf group. From top to bottom, the social hierarchy of the gray wolf group is divided into the leader, think tank, command implementer, and command executor, which are represented by α, β, δ, and ω respectively. The α wolf is mainly responsible for making decisions on group actions, the β wolf is mainly responsible for assisting the α wolf in making decisions, the δ wolf is mainly responsible for reconnaissance, sentry and other tasks, and the ω wolf is responsible for balancing the internal relationship of the group. This strict hierarchy plays an important role in the hunting process of the gray wolf population. Figure 2 The figure shows the hunting process of a gray wolf pack. The entire hunting process is completed under the leadership of the alpha wolf. The gray wolf hunting process includes surrounding the prey, hunting and attacking the prey. During the process of surrounding the prey, the gray wolf pack's behavior in the process of encircling the prey is defined as follows:

[0094]

[0095]

[0096] in, Indicates the distance between the target prey and the current individual, is the update formula for the current gray wolf position, t refers to the current iteration number, and are the position vector of the gray wolf and the position vector of the prey, and is the coefficient vector, is the convergence factor, whose value is affected by the number of iterations. As the number of iterations increases, it decreases linearly from 2 to 0. and The modulus of takes a random number between [0,1]. During the hunting process, the gray wolf can identify the prey and then determine its location and call on the group to surround it. When the gray wolf finds the prey, the β wolf and the δ wolf, led by the α wolf, guide the wolf pack to surround the prey, and the other individuals of the wolf pack move in the direction of the α wolf, the β wolf and the δ wolf. When attacking the prey, the gray wolf group starts to attack after completing the prey positioning and surrounding operations. To simulate the process of approaching the prey, The value gradually decreases, thereby reducing The fluctuation range of , that is, as the number of iterations increases, When the value decreases linearly from 2 to 0, the corresponding The value also changes in the interval [-a, a]. When , the algorithm falls into the local optimum. When , the gray wolf looks for the global optimal solution.

[0097] Due to the characteristics of discrete coding, the above gray wolf hunting process cannot be directly applied. Therefore, it is necessary to discretize the gray wolf optimization algorithm to adapt to the FTSCA model.

[0098] Step 108, iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, stop the iteration and output the inbound and outbound transfer scheduling plan represented by the current α wolf; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0099] In the above-mentioned aircraft in-and-out transfer scheduling method based on the gray wolf algorithm, by calling a pre-built scheduling model, the in-and-out storage and platform transfer between the platform and the hangar are collaboratively planned, and the gray wolf hunting process of the gray wolf algorithm with a simple structure is discretized to solve the complex scheduling model. Specifically, the neighborhood optimization of the α wolf, β wolf or δ wolf in the gray wolf population is performed to obtain the neighborhood gray wolf individuals. According to the fitness function values ​​corresponding to the neighborhood gray wolf individuals, α wolf, β wolf and δ wolf, the positions of α wolf, β wolf and δ wolf are updated, and this process is iterated to output the optimal in-and-out transfer scheduling plan, determine the transfer sequence and transfer resource allocation, and improve the transfer efficiency. The embodiment of the present invention can improve the update efficiency of the optimal solution, enhance the optimization ability of the algorithm, and thus efficiently execute the in-and-out transfer scheduling of the fleet.

[0100] In one embodiment, initializing a gray wolf population based on inbound and outbound transfer task parameters includes: obtaining a transfer priority sequence based on the number of aircraft to be transferred, the aircraft numbers of the aircraft to be transferred, and the order in which each aircraft to be transferred performs the transfer operation; encoding the transfer priority sequence to obtain gray wolf individuals; and randomly generating multiple gray wolf individuals to initialize the gray wolf population.

[0101] In this embodiment, regardless of whether it is inbound or outbound transfer, since the hangar tractor and the platform tractor are of the same type, the speed, acceleration and other parameters are consistent during the transfer process, and they are equivalent parallel machines, which will not affect the transfer time. In addition, due to the small space of the offshore platform hangar and the limited structure, the choice of elevator is relatively fixed. Therefore, the decision object that needs to be solved in this problem is the timing of the transfer operation, and the encoding method is Among them, x j The number of the aircraft with the jth transfer priority in the code.

[0102] In one embodiment, during the individual decoding process, a pre-constraint strategy is used to search for aircraft i and aircraft j in the individual to determine whether the following situations exist: Case 1: If the unique path interference parking stand set (including the initial parking stand and the target parking stand) of aircraft j at a certain stage includes the initial parking stand of aircraft i, then i has priority over j; Case 2: If the unique path interference parking stand set (including the initial parking stand and the target parking stand) of aircraft j at a certain stage includes the target parking stand of aircraft i, then j has priority over i. In this embodiment, due to improper priority allocation, a blocking situation may occur in which the plan cannot complete the transfer. Therefore, before optimization, priority pre-constraints can be performed based on the transfer task, and the priority sequence of the individuals can be changed to avoid mutual blocking that makes the plan infeasible, thereby improving the efficiency of optimization.

[0103] In one embodiment, before the roulette wheel method is used to select the α wolf, β wolf or δ wolf in the gray wolf population to perform neighborhood optimization and obtain the neighborhood gray wolf individuals, it also includes: using a preset decoding rule to perform a decoding operation on each gray wolf individual in the initialized gray wolf population, and calculating the fitness function value of each gray wolf individual based on the in-and-out transfer scheduling plan and fitness function generated after the individual decoding; the fitness function includes the fleet transfer completion time, load balancing index and transfer group transfer time; the fitness function value of each gray wolf individual is arranged in ascending order, and the fitness function values ​​of the top three are found to obtain the positions of α wolf, β wolf and δ wolf, and the positions of β wolf and δ wolf are initialized respectively. In this embodiment, in order to further improve the optimization efficiency of the algorithm, after each round of iterative new population evaluation, the position of the head wolf α is initialized. To retain the position of wolf β and δ and Compared with the initialization of α, β and δ in the standard GWO algorithm, the global leader retains the historical optimal solution, so that the algorithm can generate more neighborhood solutions based on the historical optimal solution. This operation can greatly improve the update efficiency of the optimal solution and enhance the algorithm's optimization ability.

[0104] In addition, the decoding process includes transferring the priority sequence of the aircraft and generating a specific plan. According to the individual coded priorities, the transfer operation starts from the first aircraft. First, it is determined whether there is a transfer path for the aircraft. If so, it is checked to see if it meets the various constraints of the scheduling model, and finally a complete plan is formed. However, due to the blocking relationship between parking spaces, some priority sequences cannot complete the transfer implementation, resulting in transfer failure. For example, after a higher priority aircraft enters the space, it blocks a lower priority aircraft from leaving. In order to successfully implement decoding, the set to be scheduled W is defined. g , scheduled set S g , where W g Initially, it is list priority encoding. The decoding process is as follows:

[0105] Step 1. Initialization

[0106] Step 2: When index>|W g |, go to step 6.

[0107] Step 3. Select the aircraft number i to be dispatched * =W g (index), corresponding to the transfer phase task To determine whether there is an available transfer route, retrieve the set of blocked aircraft I in the available transfer routes at each stage of the transfer task. obs , and determine whether these aircraft have been dispatched If the dispatch is not completed, the current transfer route is not available, set index = index + 1, and go to step 2; if the dispatch is completed, search I obs The latest time for an aircraft to leave the parking stand This time is the earliest possible transfer time for the current aircraft. In addition, it is also necessary to retrieve S g Is the target parking position of the aircraft that has completed the transfer the blocked parking position of the current aircraft's only available path? If so, determine whether the blocked aircraft's parking time is earlier than If so, set index = index + 1 and go to step 2.

[0108] Step 4: In order to find the aircraft transfer process, ensure each stage The optimal timing that satisfies both the transport collision avoidance constraint and the equipment transfer guarantee constraint is as follows:

[0109] Step 4.1, search in The tractor that can reach the parking space where the required transfer aircraft is located at the time is selected. If there are multiple tractors that can reach the parking space, the tractor with the smallest load is selected according to the second optimization goal. When the second optimization goal is the transfer time of the transfer equipment, the tractor with the shortest transfer guarantee time is selected. When there is no tractor available, the tractor with the smallest load is selected. When the time arrives, the preferred The earliest tractor that arrives after time. The arrival time of the tractor is recorded as

[0110] Step 4.2: The earliest start time for the aircraft transfer is The collision detection is performed based on this moment. First, the available transfer paths are searched. If there are multiple available paths, the path with the shorter time consumption is selected first. The collision detection between this path and the transferred aircraft is performed with a detection step length of Δt=1s. If the distance between the two aircraft is less than the safety threshold, a delayed departure strategy is adopted for the aircraft with lower priority. (delay multiple k = 10), repeat this step for the delayed departure aircraft until there is no collision with the completed transfer aircraft during the transfer process, and the transfer completion time is

[0111] Step 5: When the aircraft transfer is successful, update the status parameters of the parking space, transfer equipment, and aircraft. g =W g -{i *},S g =S g ∪{i *}, reset index = 1, and go back to step 2.

[0112] Step 6: After the transfer is completed, output the transfer plan and calculate the objective function value.

[0113] The present invention uses a weighted sum method to deal with multi-objective optimization problems. Taking the transport group load variance as the second optimization objective and the transport group transfer time as the third optimization objective as an example, the individual fitness is calculated as follows

[0114]

[0115] in, and They are the weighting coefficients of the second optimization objective and the third optimization objective respectively. Each level of optimization objective must have a clear order of magnitude distinction to avoid interference. g | is a penalty term, BM is a very large constant. If there are aircraft that have not completed the scheduling, the solution is not feasible. |W g | is not 0, then the fitness of the scheme approaches infinity.

[0116] In one embodiment, a roulette wheel method is used to select α wolves, β wolves, or δ wolves from a gray wolf population for neighborhood optimization to obtain neighborhood gray wolf individuals, including: assigning weights to α wolves, β wolves, and δ wolves respectively according to the fitness function values ​​of α wolves, β wolves, and δ wolves in the gray wolf population; a roulette wheel method is used to select α wolves, β wolves, or δ wolves from the gray wolf population, and a joint mutation operation is performed on the selected gray wolf individuals to obtain neighborhood gray wolf individuals.

[0117] In this embodiment, due to the characteristics of discrete coding, the gray wolf hunting process of the standard gray wolf optimization algorithm cannot be directly applied. Therefore, the present invention designs a neighborhood optimization mechanism. and Represent the current positions of α, β and δ respectively, and α, β and δ are the three solutions with the highest fitness of the population. According to their fitness ranking, the three individuals are given weights P α 、P β 、P δ , and ∑P = 1, and then a roulette wheel is used to select one of the individuals for the subsequent neighborhood optimization operation. The neighborhood individuals are generated using a joint mutation mechanism. The fitness of the new individual is evaluated and compared with the fitness of α, β, and δ, and the positions of α, β, and δ are updated. The population size is Ng. The basic process of this mechanism is shown in the following table:

[0118] Table 1 Domain optimization mechanism

[0119]

[0120] Table 2 Fitness update function

[0121]

[0122] Based on α, β and δ wolves, a joint mutation operation is performed to generate neighborhood solutions corresponding to α, β and δ wolves. The joint mutation is divided into three steps, namely, field re-insertion operation, two-point crossover operation and field inversion operation, as shown in Figure 3 The field re-insertion operation diagram shown in FIG. 1 includes: for a certain gray wolf individual, randomly select a field with a length not greater than a′ at any position, extract it, and re-insert it into any position of the remaining fields, where a′ is the convergence factor; Figure 4 The two-point crossover operation diagram shown in FIG2 includes: for a certain individual, randomly selecting two different positions in the field and exchanging the data of the two positions; Figure 5 The field reversal operation includes: for an individual, randomly selecting a field whose length is not greater than a′ at any position, and reversing the order of the selected sequence. The specific algorithm is shown in the following table:

[0123] Table 1. Joint mutation mechanisms

[0124]

[0125] In the Grey Wolf Optimization Algorithm, the convergence factor controls the balance between global search and local search. Based on this, the convergence factor a′ is designed as [1,s·D max +1], where s represents the learning factor and s∈[0,1], D max for The dimension, that is, the number of aircraft to be scheduled, a′ determines the maximum dimension of the subsequent joint mutation mechanism and local search mechanism on the individual. Due to the blocking relationship between aircraft, the primary factor affecting the fitness of the scheduling sequence is the order of aircraft scheduling. When the operation field length is small, the original field sequence retained by the operation is short, and the aircraft scheduling order changes relatively large, which can be considered to have a strong global search ability; when the operation field length is large, the original field sequence retained by the operation is long, and the aircraft scheduling order changes relatively small, which can be considered to have a strong local search ability. According to this feature, as the number of iterations increases, let a′ gradually increase from 1 to [s·D max +1], in order to effectively balance the global exploration and local development capabilities of the algorithm, a′ is increased using an exponential function, as follows:

[0126]

[0127] Where t is the current iteration number, T is the maximum iteration number, and [·] is the rounding operation.

[0128] The series of algorithm improvements mentioned above are all based on α, β, and δ wolves. This operation has the advantage of fully searching the neighborhood of the optimal solution and finding the local optimal solution. However, this also weakens the global optimization ability. Once the algorithm is stuck in a local optimal solution in the later stages, it is difficult to escape on its own. To balance local optimization and global optimization, this paper designs a chaotic search mechanism, which improves the algorithm from two aspects, increasing the search breadth and depth.

[0129] In one embodiment, after updating the positions of the gray wolf α, β and δ wolves according to the fitness function values ​​corresponding to the neighborhood gray wolf individuals, α wolf, β wolf and δ wolf, it also includes: updating the positions of the remaining gray wolf individuals in the current gray wolf population according to the updated positions of the α wolf, β wolf and δ wolf to update the current gray wolf population; reconstructing the population according to the updated population to obtain a reconstructed gray wolf population; the population reconstruction step includes: extracting multiple gray wolf individuals from the updated population, and replacing the remaining gray wolf individuals in the population with randomly generated gray wolf individuals, and obtaining a reconstructed gray wolf population according to the multiple gray wolf individuals and the randomly generated gray wolf individuals.

[0130] In this embodiment, during the algorithm iteration process, 50% of the individuals in each generation of the population are generated by the previous operations, and 50% of the individuals are randomly generated. This ensures that half of the individuals in each generation are located in the local optimal neighborhood, and half of the individuals are randomly distributed in the solution space, which greatly increases the algorithm search breadth and can effectively enhance the algorithm's global search capability and prevent it from falling into the local optimum.

[0131] In one embodiment, when the optimal solution of the population is not updated after K cycles, an iterative greedy operation is introduced, which is divided into two stages: destruction and reconstruction. In the destruction stage, the position of the α wolf is extracted. Randomly select a′ tasks and divide the original sequence into and Two parts, of which Contains a′ tasks, a′ is the convergence factor, Contains (D max -a′) tasks; in the reconstruction phase, Each task in is inserted into , select Insert each task The sequence with the lowest fitness is retained, and the next task is inserted based on the new sequence until the reconstruction is complete. The process is shown in the following table.

[0132] Table 3 Iterative greedy operations

[0133]

[0134] In a specific embodiment, Figure 6As shown, a flow chart of the IGWO algorithm is provided. The improved gray wolf algorithm includes population initialization, pre-constraint retrieval, fierce wolf selection, joint mutation, population reconstruction and iterative greed. First, the transfer priority of the aircraft to be transferred under the current mission wave is encoded, and multiple gray wolf individuals are generated by encoding to initialize the gray wolf population; then, a pre-constraint retrieval is performed on each individual in the current gray wolf population to avoid the problem of infeasibility of the solution caused by path blocking due to interference with the parking space during transfer, which can optimize the population individuals and improve the optimization efficiency; when performing fierce wolf selection, the current gray wolf population is first evaluated. During the evaluation, the gray wolf individuals are solved using the decoding operation to obtain the in-and-out transfer scheduling plan. The fitness function value of the corresponding gray wolf individual is calculated according to the in-and-out transfer scheduling plan and the fitness function, and the fitness function values ​​are sorted in ascending order. The top three gray wolf individuals are α wolf, β wolf and δ wolf, respectively. α wolf, β wolf and δ wolf are fierce wolves, and α wolf is the alpha wolf. According to the global alpha wolf mechanism, the position of α wolf is retained, and the positions of β wolf and δ wolf are initialized so that the global alpha wolf The historical optimal solution is retained so that the algorithm can generate more neighborhood solutions based on the historical optimal solution to improve the updating efficiency of the optimal solution; then, one individual among α wolf, β wolf and δ wolf is selected for joint mutation to realize neighborhood optimization. The convergence factor is introduced during the joint mutation to balance the global exploration and local development capabilities of the algorithm, and the search breadth of the algorithm is increased through population reconstruction. When the optimal solution of the population is not updated after K cycles, the convergence factor is used for iterative greed, which can increase the search depth of the algorithm and avoid the algorithm from falling into the local optimal solution. When t>T, the inbound and outbound transfer scheduling plan corresponding to the current α wolf individual is output to perform the inbound and outbound transfer scheduling of the fleet. The above-mentioned IGWO algorithm has strong optimization ability and fast convergence speed. By optimizing the solution of the complex FTSCA model through the IGWO algorithm, the quality of the optimal solution can be improved, thereby efficiently performing the inbound and outbound transfer scheduling of the fleet.

[0135] In a specific embodiment, in order to verify the model algorithm of this article, a hangar and platform with P1 to P8 platform temporary parking spaces, A1-A16 platform parking spaces, and R1-R18 hangar parking spaces are used as simulated offshore platforms for simulation experiments. To ensure no loss of generality, the simulation cases of the present invention include outbound transfer, inbound transfer, and platform transfer, and the experiment is conducted in the same way as the aircraft is ready to be dispatched again. The parking spaces P1 to P8 are temporary parking spaces for the platform, which are used for temporary parking of recovered aircraft. Since there are only eight parking spaces, the maximum number of recoveries in a single wave is 8. The dispatch mode simulates three situations: 8 aircraft, 12 aircraft, and 16 aircraft, and the optimization target parameters are: γ=[0.5,0,0.5],ω lb=[0.2,0,0.2]. The present invention uses MATLAB2020a as the experimental platform, the simulation environment is Windows 10 64-bit operating system, Intel(R)Core(TM)i7-10750HCPU@2.60GHz 2.59GHz, and memory 16G. In the 12-aircraft group's warehousing and platform transfer tasks, 6 of the 8 recovered aircraft are transferred into the warehouse and transferred to the hangar for maintenance, 2 are transferred to the platform and dispatched again, and 10 aircraft are taken out of the hangar to supplement the needs of the next wave of dispatches, in order to simulate the full staffing state. The rest of the transfer positions not involved are equipped with parked aircraft. In terms of transfer equipment, 4 tractors are configured in the hangar and the platform, and 2 elevators are set up. The aircraft mooring and unmooring time are both 30s, and the one-way transfer time of the elevator is 70s. In order to adjust the algorithm parameters, the present invention adopts L 16 (4 5 ) Orthogonal table is used to optimize the algorithm parameters. The five factors and the four levels of each factor are population size Np = {10, 30, 50, 100}, learning factor s = {0.25, 0.5, 0.75, 1}, joint mutation probability P m1 ={0.05,0.3,0.5,1}, P m2 ={0.05, 0.3, 0.5, 1}, iterative greedy starting threshold K = {3, 5, 10, 20}, simulation experiments were conducted for the above 16 parameter combinations, each parameter combination was simulated under the 12-machine dispatch mission, each experiment was conducted 20 times, the termination condition was the scheduling evaluation number of 20000, the average value of the first optimization target transfer completion time was taken, and the optimal solution was: population size Np = 50, learning factor s = 0.75, joint mutation probability P m1 =1.0, P m2 =0.5, iterative greedy starting threshold K=20.

[0136] Select the above optimal parameter settings and conduct simulation experiments again, such as Figure 7 As shown in the figure, a schematic diagram of the change trend of the optimization index of the 12-machine dispatch transfer scheduling is provided, where (a) is a schematic diagram of the change trend of the transfer completion time, (b) is a schematic diagram of the change trend of the transfer group load variance, (c) is a schematic diagram of the change trend of the transfer group transfer time, and the transfer completion time C max=43.12min, the load variance of the transfer group LBT = 4.408, and the cumulative transfer time of the transfer equipment on the ship deck CTT = 39.63min. (a) shows that the algorithm has a strong global optimization capability in the initial stage and can achieve rapid convergence within 1500 evaluations. As the algorithm iterates, the algorithm falls into a local optimum between 1500 and 3500 evaluations. This is because the convergence factor value is still relatively small and the algorithm's deep search capability is insufficient. As the number of evaluations continues to increase, the convergence factor value continues to increase, and the field length of the algorithm's joint mutation and iterative greedy operability continues to increase, causing the algorithm's search depth to continue to increase. Therefore, when the number of evaluations reaches more than 3500, the first optimization objective converges to a suboptimal solution. Combined with (b) and (c), it can be seen that under the conditions of this case, the second optimization objective successfully converges to the optimal solution when the number of evaluations reaches 4100. When the second optimization objective reaches the optimal solution, the third optimization objective is also at a relatively good solution. At this time, the transfer solution is the optimal solution for this case.

[0137] Figure 8 and Figure 9 They are the Gantt charts for aircraft fleet in and out of storage and platform collaborative transfer scheduling and the Gantt charts for transfer equipment scheduling under the optimal plan of 12 aircraft dispatching. Figure 8 In the Gantt box, "DM-1", "HM-1" and "EM-1" represent the platform tractor, hangar tractor and aircraft elevator, and the gray Gantt box represents the aircraft's mooring and unmooring. Figure 9 In the figure, the vertical axis is the code of the transfer equipment, and Figure 8 The expression is consistent with that in [1]. The ij in the Gantt box represents the j-th transfer process of aircraft i. The striped Gantt box represents the time it takes for the transfer equipment to arrive at the aircraft's parking position and wait for the aircraft to complete mooring and unmooring. The gray Gantt box represents the transfer time consumed by the transfer equipment to reach the parking position of the aircraft to be transferred. Figure 8 and Figure 9 The figure shows the transfer sequence, distribution of transfer equipment, and occupancy time of each aircraft during the fleet's entry and exit and platform collaborative transfer. It can be seen from the figure that the scheme meets all the constraints in the model and is mutually verified with the model.

[0138] In a specific embodiment, Figure 10As shown in the figure, a schematic diagram of the convergence curve comparison of the completion time of 12-machine dispatching transfer scheduling is provided. To verify the performance of this algorithm, the algorithm optimization effect is compared based on the above-mentioned cases of 8-machine dispatching, 12-machine dispatching, and 16-machine dispatching. The present invention selects several excellent algorithms for solving the problem of hybrid flow shop scheduling for comparative analysis, including GWO, GA (Genetic Algorithm) and HBV (Hybrid Biogeographic optimization based on Variable neighborhood search). In the algorithm performance comparison, only C max For consideration, the four algorithms were run 20 times in each of three case studies. IGWO successfully found the optimal solution in all three cases. Comparing the three cases, the gap between the optimal solution and the mean gradually widened as the scale of aircraft transfers increased, indicating that the probability of the algorithm finding the optimal solution gradually decreased. The data in the table also shows that the difference between the worst-case and the optimal solution also gradually increased, indicating that the robustness of the algorithm also gradually deteriorated with the increase in transfer scale. Comparing the other three algorithms, with the exception of IGWO, only the HBV algorithm found the optimal solution in the 8- and 16-aircraft sortie cases; none of the other algorithms found the optimal solution. Comparing the mean and worst-case values ​​of the various algorithms across the three case studies, the algorithm of the present invention still performed best in both respects, indicating that the optimization and robustness of these three algorithms also lag behind those of the algorithm of the present invention. Based on this comparison, the IGWO algorithm designed by the present invention outperforms the other algorithms in both optimization and robustness, with the algorithm ranking as follows: IGWO > HBV > GWO > GA.

[0139] It should be understood that although Figure 1 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. In addition, Figure 1 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 sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0140] In one embodiment, Figure 11 As shown, a device for aircraft in-and-out transfer scheduling based on the Grey Wolf Algorithm is provided, comprising: a parameter acquisition module 1102, a model calling module 1104, a population update module 1106 and a transfer scheduling module 1108, wherein:

[0141] The parameter acquisition module 1102 is used to obtain the current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred;

[0142] Model calling module 1104 is used to call a pre-built scheduling model; the constraints of the scheduling model include equipment allocation exclusivity constraints, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective functions of the scheduling model include minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group;

[0143] The population update module 1106 is used to initialize the gray wolf population according to the inbound and outbound transfer task parameters, select α wolf, β wolf or δ wolf from the gray wolf population by using a roulette wheel method to perform neighborhood optimization to obtain a neighborhood gray wolf individual, and update the positions of α wolf, β wolf and δ wolf according to the fitness function values ​​corresponding to the neighborhood gray wolf individual, α wolf, β wolf and δ wolf;

[0144] The transfer scheduling module 1108 is used to iteratively update the positions of α wolf, β wolf and δ wolf. When the pre-selected iteration stop condition is met, the iteration is stopped and the inbound and outbound transfer scheduling plan represented by the current α wolf is output; the inbound and outbound transfer scheduling plan is used for aircraft inbound and outbound transfer scheduling.

[0145] In one embodiment, the population update module 1106 is further used to assign weights to α wolf, β wolf and δ wolf in the gray wolf population according to the fitness function values ​​of α wolf, β wolf and δ wolf in the gray wolf population; select α wolf, β wolf or δ wolf in the gray wolf population by a roulette wheel method, perform a joint mutation operation on the selected gray wolf individuals, and obtain neighborhood gray wolf individuals.

[0146] In one embodiment, the population update module 1106 is further configured to perform a combined mutation operation including a field re-insertion operation, a two-point crossover operation, and a field reversal operation; the field re-insertion operation includes extracting fields from the selected gray wolf individuals whose length is not greater than a convergence factor and inserting the fields into the remaining fields; the convergence factor is:

[0147]

[0148] Where t is the current iteration number, T is the maximum iteration number, and [·] is a rounding operation. The field reversal operation includes selecting fields whose length is not greater than the convergence factor in the selected gray wolf individuals and reversing the order of the fields.

[0149] In one embodiment, the population update module 1106 is also used to perform a decoding operation on each gray wolf individual in the initialized gray wolf population using a preset decoding rule, and calculate the fitness function value of each gray wolf individual based on the in-and-out transfer scheduling plan and fitness function generated after the individual decoding; the fitness function includes the fleet transfer completion time, load balancing index and transfer group transfer time; the fitness function value of each gray wolf individual is arranged in ascending order, and the fitness function values ​​of the top three are found to obtain the positions of α wolf, β wolf and δ wolf; and the positions of β wolf and δ wolf are initialized respectively.

[0150] In one embodiment, the population update module 1106 is further used to obtain a transfer priority sequence based on the number of aircraft to be transferred, the aircraft numbers of the aircraft to be transferred, and the order in which each aircraft to be transferred performs the transfer operation; encode the transfer priority sequence to obtain gray wolf individuals; and randomly generate multiple gray wolf individuals to initialize the gray wolf population.

[0151] In one embodiment, the population update module 1106 is also used to perform an iterative greedy operation if the optimal solution of the current gray wolf population has not been updated after K iterations; the iterative greedy operation includes: extracting the position of the α wolf in the current gray wolf population, and randomly dividing the transfer priority sequence corresponding to the α wolf according to the convergence factor to obtain a first transfer priority sequence and a second transfer priority sequence; the number of aircraft to be transferred in the first transfer priority sequence is the value corresponding to the convergence factor; inserting the aircraft to be transferred in the first transfer priority sequence into different positions of the second transfer priority sequence in sequence, and calculating the fitness function value corresponding to the current second task sequence after each insertion of the aircraft to be transferred into a different position, and updating the second transfer priority sequence according to the second transfer priority sequence corresponding to the lowest fitness function value; iteratively updating the second transfer priority sequence, and outputting the reconstructed transfer priority sequence when the pre-set iteration stop condition is met; and updating the position of the α wolf in the current gray wolf population according to the reconstructed transfer priority sequence.

[0152] In one embodiment, the population update module 1106 is further used to update the positions of the remaining gray wolf individuals in the current gray wolf population based on the updated positions of the α wolf, β wolf and δ wolf to update the current gray wolf population; reconstruct the population based on the updated population to obtain a reconstructed gray wolf population; the population reconstruction step includes: extracting multiple gray wolf individuals from the updated population, and replacing the remaining gray wolf individuals in the population with randomly generated gray wolf individuals, and obtaining a reconstructed gray wolf population based on the multiple gray wolf individuals and the randomly generated gray wolf individuals.

[0153] Regarding the specific limitations of the aircraft in-and-out transfer scheduling device based on the Gray Wolf Algorithm, please refer to the limitations of the aircraft in-and-out transfer scheduling method based on the Gray Wolf Algorithm above, which will not be repeated here. The various modules in the above-mentioned aircraft in-and-out transfer scheduling device based on the Gray Wolf Algorithm 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 the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0154] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 12 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for scheduling aircraft in and out of the warehouse transfer based on the gray wolf algorithm is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0155] Those skilled in the art will understand that Figure 12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0156] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the method in the above embodiment when executing the computer program.

[0157] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in the above embodiment are implemented.

[0158] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. 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 can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various 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 (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0159] The technical features of the above embodiments can 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.

[0160] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for scheduling aircraft inbound and outbound transfers based on the Grey Wolf algorithm, characterized in that: The method comprises: Obtaining current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred; Invoking a pre-built scheduling model; the constraints of the scheduling model include equipment allocation exclusivity constraints, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group; The gray wolf population is initialized according to the inbound and outbound transport task parameters, and the roulette wheel method is used to select the gray wolf population. Wolf, Wolf or The wolf performs neighborhood optimization and obtains the neighborhood gray wolf individual. According to the neighborhood gray wolf individual, Wolf, Wolf and The fitness function value corresponding to the wolf is updated Wolf, Wolf and The location of the wolf; Iterative Updates Wolf, Wolf and The position of the wolf, when the pre-set iteration stop condition is met, the iteration stops and the current The wolf represents the inbound and outbound transfer scheduling plan; the inbound and outbound transfer scheduling plan is used to perform aircraft inbound and outbound transfer scheduling; In the roulette wheel method of selecting gray wolf populations Wolf, Wolf or Before the wolf performs neighborhood optimization and obtains the neighboring gray wolf individuals, it also includes: A decoding operation is performed on each individual gray wolf in the initialized gray wolf population using a preset decoding rule, and the fitness function value of each gray wolf individual is calculated based on the inbound and outbound transfer scheduling plan and fitness function generated after the individual decoding; the fitness function includes the group transfer completion time, load balancing index and transfer group transfer time; Arrange the fitness function values ​​of each gray wolf individual in ascending order, find the fitness function values ​​of the top three in the ranking, and obtain Wolf, Wolf and The location of the wolf and the Wolf and the Initialize the wolf's position; The transfer equipment transfer guarantee timing constraints are: in, For the transfer process The end time of the job, For the k Equipment by process After completion, the parking position is transferred to the process The time of parking before starting, For the transfer process The start time of the operation, is a very large constant, Gather for the transfer aircraft, For aircraft i The task set, , For aircraft i The transfer process j .

2. The method according to claim 1, characterized in that Roulette wheel method was used to select the gray wolf population. Wolf, Wolf or The steps for the wolf to perform neighborhood optimization and obtain the neighboring gray wolf individuals include: According to the gray wolf population Wolf, Wolf and The fitness function value of the wolf is assigned Wolf, Wolf and Wolf weight; Roulette wheel method was used to select the gray wolf population. Wolf, Wolf or Wolf, perform joint mutation operation on the selected gray wolf individuals to obtain neighborhood gray wolf individuals.

3. The method according to claim 2, characterized in that The joint mutation operation includes a field re-insertion operation, a two-point crossover operation, and a field reversal operation; The field re-insertion operation includes extracting fields whose length is not greater than a convergence factor from the selected gray wolf individuals and inserting the fields into the remaining fields; the convergence factor is: in, is the convergence factor, is the learning factor, and , is the number of aircraft to be dispatched, is the current iteration number, is the maximum number of iterations, is the rounding operation; The field reversal operation includes selecting fields whose lengths are not greater than the convergence factor in the selected gray wolf individuals and reversing the order of the fields.

4. The method according to claim 1, wherein Initializing the gray wolf population according to the inbound and outbound transport task parameters includes: Obtaining a transfer priority sequence according to the number of aircraft to be transferred, the aircraft numbers of the aircraft to be transferred, and the order in which each aircraft to be transferred performs the transfer operation; Encoding the transport priority sequence to obtain gray wolf individuals; Multiple randomly generated gray wolf individuals are used to initialize the gray wolf population.

5. The method according to claim 3, characterized in that The method further comprises: If the optimal solution of the current gray wolf population is If the update is not completed after the first iteration, the iterative greedy operation is performed; the iterative greedy operation includes: Extract the current gray wolf population The position of the wolf, according to the convergence factor The transfer priority sequence corresponding to the wolf is randomly divided to obtain a first transfer priority sequence and a second transfer priority sequence; the number of aircraft to be transferred in the first transfer priority sequence is the value corresponding to the convergence factor; Inserting the aircraft to be transferred in the first transfer priority sequence into different positions of the second transfer priority sequence in sequence, calculating the fitness function value corresponding to the current second task sequence each time the aircraft to be transferred is inserted into the different position, and updating the second transfer priority sequence according to the second transfer priority sequence corresponding to the lowest fitness function value; Iteratively updating the second transport priority sequence, and outputting a reconstructed transport priority sequence when a preset iteration stop condition is satisfied; Update the current gray wolf population according to the reconstructed transport priority sequence Wolf's location.

6. The method according to claim 1, characterized in that According to the neighborhood gray wolf individuals, Wolf, Wolf and The fitness function value corresponding to the wolf is updated Wolf, Wolf and After the wolf's position, it also includes: According to the updated Wolf, Wolf and The position of the wolf, updates the position of the remaining gray wolf individuals in the current gray wolf population to update the current gray wolf population; The population is reconstructed based on the updated population to obtain a reconstructed gray wolf population; the population reconstruction step includes: A plurality of gray wolf individuals are extracted from the updated population, and the remaining gray wolf individuals in the population are replaced with randomly generated gray wolf individuals, and a reconstructed gray wolf population is obtained based on the plurality of gray wolf individuals and the randomly generated gray wolf individuals.

7. An aircraft in-and-out transfer scheduling device based on the Grey Wolf algorithm, characterized in that: The device comprises: A parameter acquisition module is used to obtain the current inbound and outbound transfer task parameters; the inbound and outbound transfer task parameters include the number of aircraft to be transferred and the aircraft numbers of the aircraft to be transferred; A model calling module is used to call a pre-built scheduling model; the constraints of the scheduling model include exclusive constraints on equipment allocation, aircraft transfer timing constraints, aircraft transfer collision avoidance constraints, and transfer equipment transfer guarantee timing constraints; the objective function of the scheduling model includes minimizing the fleet transfer completion time, minimizing the load balancing index, and minimizing the transfer time of the transfer group; The population update module is used to initialize the gray wolf population according to the inbound and outbound transport task parameters, and select the gray wolf population in a roulette wheel manner. Wolf, Wolf or The wolf performs neighborhood optimization and obtains the neighborhood gray wolf individual. According to the neighborhood gray wolf individual, Wolf, Wolf and The fitness function value corresponding to the wolf is updated Wolf, Wolf and The location of the wolf; Transfer scheduling module for iterative updates Wolf, Wolf and The position of the wolf, when the pre-set iteration stop condition is met, the iteration stops and the current The wolf represents the inbound and outbound transfer scheduling plan; the inbound and outbound transfer scheduling plan is used to perform aircraft inbound and outbound transfer scheduling; The population update module is also used to perform decoding operations on each individual gray wolf in the initialized gray wolf population using a preset decoding rule, and calculate the fitness function value of each individual gray wolf based on the in-and-out transfer scheduling plan and fitness function generated after individual decoding; the fitness function includes the fleet transfer completion time, load balancing index and transfer group transfer time; the fitness function values ​​of each individual gray wolf are arranged in ascending order, and the fitness function values ​​of the top three are found to obtain Wolf, Wolf and The wolf's position and the Wolf and the Initialize the wolf's position; The transfer equipment transfer guarantee timing constraints are: in, For the transfer process The end time of the job, For the k Equipment by process After completion, the parking position is transferred to the process The time of parking before starting, For the transfer process The start time of the operation, is a very large constant, Gather for the transfer aircraft, For aircraft i The task set, , For aircraft i The transfer process j .

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 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 method according to any one of claims 1 to 6 are implemented.

Images