A planning method for deck sortie scheduling of carrier-based aircraft formations under a lever-less traction transport mode

Abstract

This invention provides a planning method for carrier-based aircraft formation deck sortie scheduling under a lever-less traction transport mode, belonging to the field of carrier-based weapon and equipment support technology. The planning method involves four steps: first, extracting key processes and virtual machines from the carrier-based aircraft sortie operation flow; second, clarifying decision variables and constraints to establish a combinatorial optimization problem; third, constructing a standard trajectory library based on the deck environment and layout; and fourth, applying a genetic algorithm to solve the problem. The beneficial effects of this invention are that by constructing a combinatorial optimization model through reasonable division and analysis of the carrier-based aircraft formation sortie process, combined with trajectory libraries generated for different transport systems, and utilizing a genetic algorithm based on a two-layer encoding method with a fusion delay sortie strategy, reliable and efficient generation of carrier-based aircraft formation sortie scheduling schemes is achieved, ensuring the cooperative collision avoidance requirements in the generated scheme and thus guaranteeing the rationality of the scheduling scheme in practical engineering applications.

Description

Technical Field

[0001] This invention belongs to the field of shipborne weapon and equipment support technology, and relates to a planning method for a shipborne aircraft formation deck sortie scheduling scheme under a lever-less traction and transportation mode. Background Technology

[0002] Aircraft deck sortie scheduling is a complex process optimization problem with numerous constraints. By extracting typical operational procedures, it can be approximated as a flexible job shop scheduling problem (FJSP). However, compared to traditional FJSP, aircraft deck sortie scheduling requires a focus on the collaborative collision avoidance problem among various transport entities (including the aircraft under single-aircraft taxiing, the tractor, and the traction and transport system composed of the two). When using tractors for transport, catapults and tractors must also be considered as scheduling resources; an aircraft can be transported to different catapults via different tractors for launch. Therefore, the aircraft sortie scheduling problem under the traction and transport mode essentially constitutes a combinatorial optimization problem under strongly coupled constraints of "time-space-resource". Proposing a high-fidelity sortie scheduling decision model and constructing a reliable and efficient solution algorithm are key technologies for achieving high-efficiency sortie of carrier air formations, which is crucial for improving the combat capability of carrier battle groups. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a planning method for deck sortie scheduling of carrier-based aircraft formations under a leverless traction and transportation mode. Based on a detailed analysis of the deck sortie operation process, key procedures and virtual machines are extracted, and the decision variables and constraints are analyzed to establish an FJSP model with minimum sortie time as the optimization index. According to the deck environment and layout, trajectories of the traction system and tractor vehicles between different mooring points and preparation points are constructed to form a standard trajectory library. Since the established optimization model contains many complex implicit constraints, a genetic algorithm is considered for reliable solution. Therefore, the encoding method, decoding method, and genetic operator are specifically designed for the problem. In particular, to achieve collision avoidance and coordinated movement of various transportation entities during sortie, a spatiotemporal coordination mechanism based on delayed waiting is proposed.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] A planning method for carrier-based aircraft formation deck sortie scheduling under a lever-less traction mode includes the following steps: First, extracting key processes and virtual machines from the carrier-based aircraft sortie operation flow; second, clarifying decision variables and constraints, and establishing a combinatorial optimization problem; third, constructing a standard trajectory library based on the deck environment and layout; and fourth, applying a genetic algorithm to solve the problem.

[0006] Step 1: Extract key procedures and virtual machines from the carrier-based aircraft sortie operation process.

[0007] The FJSP model, constructed to address the sortie and deployment problem of carrier-based aircraft formations, treats carrier-based aircraft as workpieces and catapults and tractors as processing machines. Efficient resource utilization can be ensured by rationally scheduling the timing of resource usage. Based on the workpiece and machine settings, for a sortie mission involving n carrier-based aircraft, six processing steps can be extracted from the entire sortie process:

[0008] Step 1: The tractor is moved to the corresponding parking position of the carrier-based aircraft. In this stage, the tractor is moved from the parking position corresponding to the departure preparation position to the parking position corresponding to the parking position of the carrier-based aircraft to be deployed. The virtual processing machine used in this step is the tractor. Any tractor can be selected to complete this step. The number of machines that can complete this step is denoted as m1.

[0009] Step 2: Transporting the carrier-based aircraft to its sortie ready position. In this stage, the tractor and the carrier-based aircraft form a traction system, transporting the aircraft to the sortie ready position corresponding to the catapult. The aircraft can be transported to any ready position. The machine used in this stage should be the catapult; however, considering that the carrier-based aircraft only occupies resources when it reaches its ready position during the sortie process, an additional number of virtual machines equal to the number of catapults are set up for this stage. The number of virtual machines that can complete this step is denoted as m2.

[0010] Step 3: Carrier-based aircraft taxi to the catapult. In this stage, the carrier-based aircraft taxis from the sortie preparation position to the corresponding catapult. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as m3, and m3 = m2.

[0011] Step 4: Pre-flight inspection phase for carrier-based aircraft. This phase involves the pilot checking the status of the carrier-based aircraft before takeoff, and the selected catapult performing pre-flight preparations (such as raising the deflector). The virtual machine used in this phase is the catapult, and the number of machines that can complete this phase is denoted as m4, where m4 = m2.

[0012] Step 5: Carrier-based aircraft take off. After confirming that the carrier-based aircraft is in normal condition and the catapult is ready, the carrier-based aircraft is launched. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as m5, where m5 = m2.

[0013] Step 6: Cooling and retraction of deflector plates. After the carrier-based aircraft takes off, the corresponding deflector plates should be cooled and retracted to their initial positions to ensure that subsequent carrier-based aircraft can taxi to the catapult and take off. The virtual machine used in this step is the catapult. The number of machines that can complete this step is denoted as m6, where m6 = m2.

[0014] Step 2: Define the decision variables and constraints, and establish a combinatorial optimization problem.

[0015] Step 2-1: Clarify the variables involved in the scheduling process and their meanings.

[0016] The meanings of the variables involved in the scheduling process are described below:

[0017]

[0018] Step 2-2: Analyze the constraints involved in each stage and describe them as a mathematical model.

[0019] 1): Any operation i of job j can only be assigned to one machine, i.e.

[0020]

[0021] 2): For It should be ensured that the completion time of operation i in work j is greater than the sum of its start time and the processing time required for that operation, i.e.

[0022] C i,j,l >S i,j,l +p i,j,l -V(1-Y i,j,l (2)

[0023] Where * = 1, then V(*) = -∞; otherwise, V(*) = 0.

[0024] 3): Operation i of operation j must be performed after operation i-1.

[0025]

[0026] 4): When the machine used in process 2 is determined, the machine used in subsequent processes should also be determined.

[0027] Y 2,j,l =Y 3,j,l =Y 4,j,l =Y 5,j,l =Y 6,j,l (4)

[0028] 5): Once a tractor is selected for operation j (process 1), the tractor can only serve operation k (process 1) after operation j (process 2) is completed.

[0029] S 1,j,l ≥C 2,k,m +V(1-X 1,j,k,l (5)

[0030] 6) Each sortie stand can hold a maximum of one aircraft. For the same standby stand, a safety time of T is set. safeThis indicates that the carrier-based aircraft preparation position has been occupied, and the process has begun (T3). safe After a certain time, subsequent carrier-based aircraft entering the preparation position will not collide with the previous carrier-based aircraft while in the preparation position.

[0031] C 2,j,l ≥S 3,k,l +T safe +V(1-X 2,j,k,l (6)

[0032] 7): If a catapult is selected for step 3 of operation j, then the catapult must serve the subsequent operation k after the cooling phase is completed.

[0033] S 3,j,l ≥C 6,k,l +V(1-X 3,j,k,l (7)

[0034] 8): Only one carrier-based aircraft is allowed to be launched from the deck at a time.

[0035] (S 4,k,h -C 4,j,l )×(C 4,k,h -S 4,j,l )≥0 (8)

[0036] 9): The trajectory chosen by task j in process i (i≤2) is denoted as Tr. i,j ,right To achieve collision avoidance, the following condition must be met at any time t:

[0037] min||Tr i,j (t)-Tr h,k (t)||≥D safe (9)

[0038] Where D safe A safe distance is required to ensure that the two systems do not collide during dynamic operation.

[0039] Steps 2-3: Establish a combinatorial optimization problem based on the objective function and constraints.

[0040] The objective function is set to minimize the deployment time, i.e., min(max(C) 6,j,l At this point, the following combinatorial optimization problem can be constructed:

[0041]

[0042] Step 3: Construct a standard trajectory library based on the deck environment and layout.

[0043] Based on the division of the carrier-based aircraft sortie process, step 1 involves the movement trajectory of the tractor; step 2 involves the movement trajectory of the traction system consisting of the carrier-based aircraft and the tractor; and steps 3 and 5 involve the taxiing trajectory of the carrier-based aircraft. During the sortie, since other systems are not allowed to pass around the catapult and the sortie runway, a restricted zone must be established. Furthermore, the trajectories involved in steps 3 and 5 are all collision-free fixed trajectories within the restricted zone; therefore, these two stages do not involve trajectory selection or collision avoidance requirements and do not require processing. Once the carrier-based aircraft deck layout is determined, the aircraft parking positions are fixed, and the catapult is also positioned in a fixed location on the deck. Therefore, it is considered to pre-generate a trajectory library of all movement trajectories for the scheduling algorithm to directly call, improving the efficiency of the scheduling plan. For step 1, the tractor needs to generate all trajectories from the sortie preparation position to the parking position, forming a tractor trajectory library, which contains m²×n trajectories, denoted as {Traj}. 1,i,j |i=1,2,…n,j=1,2,…m2};For process 2, it is necessary to generate m2×n trajectories for the traction system from each stop position to all departure preparation positions, forming a trajectory library, which is denoted as {Traj}. 2,i,j |i=1,2,…n,j=1,2,…m2}. It should be noted that collision avoidance for static obstacles must be considered during trajectory generation. Since there may be waiting transport systems in relevant parking spaces, depots, and deployment preparation positions, all relevant depots, parking spaces, deployment preparation positions, and restricted areas are designated as obstacle areas during trajectory library generation. For dynamic collision avoidance, a delayed deployment strategy is used in subsequent decoding to rationally schedule the start time of corresponding procedures to achieve coordinated trajectory arrangement.

[0044] Step 4: Solve the problem using a genetic algorithm.

[0045] Step 4-1: Determine the gene encoding method of the genetic algorithm and generate the initial population.

[0046] Based on the characteristics of carrier-based aircraft sortie missions, and utilizing, for example... Figure 2 The genetic algorithm using the two-layer coding method shown solves the problem. The coding is divided into two layers, using integer coding. The upper layer represents the workpiece and its corresponding process, and the lower layer represents the machine used for the corresponding process in the upper layer. The upper layer has a total of (6 processes × n aircraft) bits of coding. Each job number will appear 6 times in the coding, randomly arranged. The number of times the job number repeats in the coding represents the number of processes. For the lower layer coding, the machines available for the corresponding process are first listed from 1 to m. i First, assign a number, then generate a machine number from the set of machines capable of performing that process based on the process number represented by the upper-level code. Repeat the code generation process N. P Next, generate N P The initial population of individuals.

[0047] Step 4-2: Decode the gene considering cooperative collision avoidance between trajectories and other constraints.

[0048] During the decoding phase, for the trajectories involved in processes 1 and 2, the trajectories are retrieved from the trajectory library based on the current location of the corresponding system and the corresponding virtual machine. Safety distance D safe Set to ΔT*v sys +R dis ΔT represents the collision avoidance detection time precision, v sys R represents the maximum driving speed for different systems. dis To maintain a safe distance for collision avoidance in a static state, the system needs to re-interpolate all scheduled trajectories at the same time interval ΔT. Specifically, for the ith operation of the j-th task, the trajectory Tr... i,j Will be according to time [S i,j,l ,S i,j,l +ΔT,S i,j,l +2ΔT,…,C i,j,l Interpolation is performed to obtain the location information at the corresponding time point. Then, the location information at [S] is retrieved. i,j,l C i,j,l Information on other transport trajectories that are being transported within the time interval is used to determine whether the current trajectory meets the constraints (9) with other trajectories. If not, it is considered that the collision avoidance constraint is not met, and the execution time of the current process is delayed by ΔT, i.e., S. i,j,l =S i,j,l +ΔT, and re-determine the constraints and process execution time delay until the constraint (9) is satisfied. For other constraints, the problem is decoded by applying logical constraints. Finally, the individual is decoded and a scheduling scheme is generated.

[0049] Step 4-3: Determine the population's evolutionary pattern and update the population.

[0050] To address the issue of carrier-based aircraft sorties, the following three population evolution methods are proposed.

[0051] 1) Copy

[0052] The parent generation is selected using a roulette wheel method based on the replication probability, and the genes are replicated.

[0053] 2) Cross

[0054] The parent generation is selected using a roulette wheel method, and crossover points are determined. Gene fragments between these crossover points are then cross-crossed. Only the first layer of coding can be cross-crossed. After cross-crossing, the coding is adjusted to meet the requirements of the dispatch process. Subsequently, the second layer of coding is adjusted according to the machine constraints of the dispatch process to ensure machine compatibility.

[0055] 3) Variation

[0056] The parent generation is selected using a roulette wheel method to determine the mutation point, and then the following mutation methods are applied to the two layers of encoding respectively.

[0057] ① The mutation of the first layer of coding: The first layer of coding randomly selects genes for mutation according to the roulette method. After mutation, the missing process needs to be made up by the extra process. The second layer of coding is adjusted according to the mutation point of the first layer of coding to make the machine correspond.

[0058] ② Variation of the second layer of coding: Since the same process can be completed by different machines, the machines used can be varied. However, it should be noted that once the machine for process 2 is determined, the corresponding machine for subsequent processes should also be determined. Therefore, only the machines used for process 1 and process 2 should be varied. When the machine used for process 2 changes, the machines used for subsequent processes should also be adjusted.

[0059] Step 4-4: Iterate through the population and retain the optimal solution to generate the final scheduling scheme.

[0060] Based on the population change pattern determined in step 4-3, the population is iterated. After each iteration, each individual is decoded according to the decoding method in step 4-2, the objective function value of each individual is recorded, and the current optimal code is updated until the iteration ends. Finally, the optimal code is decoded to generate a scheduling scheme.

[0061] The beneficial effects of this invention are as follows:

[0062] This invention constructs a combinatorial optimization model by rationally dividing and analyzing the carrier-based aircraft formation sortie process. Combined with trajectory libraries generated for different dispatching systems, it utilizes a genetic algorithm based on a two-layer coding method with a fusion delay sortie strategy to reliably and efficiently generate carrier-based aircraft formation sortie scheduling schemes. This ensures the cooperative collision avoidance requirement in the generated scheme, thereby guaranteeing the rationality of the scheduling scheme in practical engineering applications. Attached Figure Description

[0063] Figure 1 The flowchart is for the invention.

[0064] Figure 2 This is a schematic diagram of a two-layer encoding method.

[0065] Figure 3 This is a diagram illustrating the deck layout, aircraft launch arrangements, and catapult usage for an example.

[0066] Figure 4 This is a schematic diagram illustrating the restricted area setting in the embodiment.

[0067] Figure 5 This is a tractor trajectory database used in the embodiment.

[0068] Figure 6 This is the trajectory library for the traction system in the embodiment.

[0069] Figure 7 Schedule time Gantt charts for the corresponding job processes on each virtual machine.

[0070] Figure 8 A Gantt chart showing the time schedule for each carrier-based aircraft's corresponding operational procedures. Detailed Implementation

[0071] The present invention will be further described below with reference to specific embodiments.

[0072] Step 1: Extract key procedures and virtual machines from the carrier-based aircraft sortie operation process.

[0073] The FJSP model, constructed to address the carrier-based aircraft formation deployment and transportation problem, treats carrier-based aircraft as workpieces and catapults and tractors as processing machines. Efficient resource utilization can be ensured by rationally scheduling the timing of resource usage. Based on the workpiece and machine settings, for a deployment of n=12 carrier-based aircraft, six processing steps can be extracted from the entire deployment process:

[0074] Step 1: The tractor is moved to the corresponding parking position of the carrier-based aircraft. In this stage, the tractor is moved from the parking position corresponding to the departure preparation position to the parking position corresponding to the parking position of the carrier-based aircraft to be deployed. The virtual processing machine used in this step is the tractor. Any tractor can be selected to complete this step. The number of machines that can complete this step is denoted as m1. In this embodiment, m1 = 3 is set.

[0075] Step 2: Transporting the carrier-based aircraft to its sortie ready position. In this stage, the tractor and the carrier-based aircraft form a traction system, transporting the aircraft to the sortie ready position corresponding to the catapult. The aircraft can be transported to any ready position. The machine used in this stage should be the catapult; however, considering that the carrier-based aircraft only occupies resources when it reaches its ready position during the sortie process, an additional number of virtual machines equal to the number of catapults are set up for this stage. The number of virtual machines that can complete this step is denoted as m2; in this embodiment, m2 = 3.

[0076] Step 3: Carrier-based aircraft taxi to the catapult. In this stage, the carrier-based aircraft taxis from the sortie preparation position to the corresponding catapult. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as m3, and m3 = m2.

[0077] Step 4: Pre-flight inspection phase for carrier-based aircraft. This phase involves the pilot checking the status of the carrier-based aircraft before takeoff, and the selected catapult performing pre-flight preparations (such as raising the deflector). The virtual machine used in this phase is the catapult, and the number of machines that can complete this phase is denoted as m4, where m4 = m2.

[0078] Step 5: Carrier-based aircraft take off. After confirming that the carrier-based aircraft is in normal condition and the catapult is ready, the carrier-based aircraft is launched. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as m5, where m5 = m2.

[0079] Step 6: Cooling and retraction of deflector plates. After the carrier-based aircraft takes off, the corresponding deflector plates should be cooled and retracted to their initial positions to ensure that subsequent carrier-based aircraft can taxi to the catapult and take off. The virtual machine used in this step is the catapult. The number of machines that can complete this step is denoted as m6, where m6 = m2.

[0080] Step 2: Define the decision variables and constraints, and establish a combinatorial optimization problem.

[0081] Step 2-1: Clarify the variables involved in the scheduling process and their meanings.

[0082] The meanings of the variables involved in the scheduling process are described below:

[0083]

[0084]

[0085] Step 2-2: Analyze the constraints involved in each stage and describe them as a mathematical model.

[0086] 1): Any operation i of job j can only be assigned to one machine, i.e.

[0087]

[0088] 2): For It should be ensured that the completion time of operation i in work j is greater than the sum of its start time and the processing time required for that operation, i.e.

[0089] C i,j,l >S i,j,l +p i,j,l -V(1-Y i,j,l (12)

[0090] Where * = 1, then V(*) = -∞; otherwise, V(*) = 0.

[0091] 3): Operation i of operation j must be performed after operation i-1.

[0092]

[0093] 4): When the machine used in process 2 is determined, the machine used in subsequent processes should also be determined.

[0094] Y 2,j,l =Y 3,j,l =Y 4,j,l =Y 5,j,l =Y 6,j,l (14)

[0095] 5): Once a tractor is selected for operation j (process 1), the tractor can only serve operation k (process 1) after operation j (process 2) is completed.

[0096] S 1,j,l ≥C 2,k,m +V(1-X 1,j,k,l (15)

[0097] 6) Each sortie stand can hold a maximum of one aircraft. For the same standby stand, a safety time of T is set. safe This indicates that the carrier-based aircraft preparation position has been occupied, and the process has begun (T3). safe After a certain time, subsequent carrier-based aircraft entering the preparation position will not collide with the previous carrier-based aircraft while in the preparation position.

[0098] C 2,j,l ≥S 3,k,l +T safe +V(1-X 2,j,k,l (16)

[0099] 7): If a catapult is selected for step 3 of operation j, then the catapult must serve the subsequent operation k after the cooling phase is completed.

[0100] S 3,j,l ≥C 6,k,l +V(1-X 3,j,k,l (17)

[0101] 8): Only one carrier-based aircraft is allowed to be launched from the deck at a time.

[0102]

[0103] 9): The trajectory chosen by task j in process i (i≤2) is denoted as Tr. i,j ,right To achieve collision avoidance, the following condition must be met at any time t:

[0104] min||Tr i,j (t)-Tr h,k (t)||≥D safe (19)

[0105] Where D safe A safe distance is required to ensure that the two systems do not collide during dynamic operation.

[0106] Steps 2-3: Establish a combinatorial optimization problem based on the objective function and constraints.

[0107] The objective function is set to minimize the deployment time, i.e., min(max(C) 6,j,l At this point, the following combinatorial optimization problem can be constructed:

[0108]

[0109] Step 3: Construct a standard trajectory library based on the deck environment and layout.

[0110] Based on the division of the carrier-based aircraft sortie process, step 1 involves the movement trajectory of the tractor; step 2 involves the movement trajectory of the traction system consisting of the carrier-based aircraft and the tractor; and steps 3 and 5 involve the taxiing trajectory of the carrier-based aircraft. During the sortie, since other systems are not allowed to pass around the catapult and the sortie runway, a restricted zone must be established. Furthermore, the trajectories involved in steps 3 and 5 are fixed, collision-free trajectories within the restricted zone; therefore, these two stages do not involve trajectory selection or collision avoidance requirements and do not require processing. Once the carrier-based aircraft deck layout is determined, the aircraft parking positions are fixed, and the catapults are also positioned in fixed locations on the deck. Therefore, it is considered to pre-generate a trajectory library of all movement trajectories for the scheduling algorithm to directly call, improving the efficiency of the scheduling plan. In this embodiment, the carrier-based aircraft deck layout and catapult positions are arranged as follows: Figure 3 As shown, the established no-fly zone is as follows: Figure 4 As shown. For process 1, the tractor needs to generate a tractor trajectory library consisting of all trajectories from the departure preparation position to the parking position. The generated tractor trajectory library is as follows: Figure 5 As shown, there are a total of m²×n trajectories, denoted as {Traj}. 1,i,j |i=1,2,…n,j=1,2,…m2};For process 2, it is necessary to generate m2×n trajectories for the traction system from each stop position to all departure preparation positions, forming a trajectory library, which is denoted as {Traj}. 2,i,j The generated traction system trajectory library is as follows: |i=1,2,…n,j=1,2,…m2} Figure 6 As shown. It should be noted that collision avoidance for static obstacles must be considered during trajectory generation. Since there may be waiting transport systems in relevant parking spaces, depots, and deployment preparation positions, all relevant depots, parking spaces, deployment preparation positions, and restricted areas are designated as obstacle areas during trajectory database generation. For dynamic collision avoidance, a delayed deployment strategy is used in subsequent decoding to rationally schedule the start time of corresponding procedures to achieve coordinated trajectory arrangement.

[0111] Step 4: Solve the problem using a genetic algorithm.

[0112] Step 4-1: Determine the gene encoding method of the genetic algorithm and generate the initial population.

[0113] Based on the characteristics of carrier-based aircraft sortie missions, and utilizing, for example... Figure 2 The genetic algorithm using the two-layer coding method shown solves the problem. The coding is divided into two layers, using integer coding. The upper layer represents the workpiece and its corresponding process, and the lower layer represents the machine used for the corresponding process in the upper layer. The upper layer has a total of (6 processes × n aircraft) bits of coding. Each job number will appear 6 times in the coding, randomly arranged. The number of times the job number repeats in the coding represents the number of processes. For the lower layer coding, the machines available for the corresponding process are first listed from 1 to m. i First, assign a number, then generate a machine number from the set of machines capable of performing that process based on the process number represented by the upper-level code. Repeat the code generation process N. P Next, generate N P The initial population of individuals.

[0114] Step 4-2: Decode the gene considering cooperative collision avoidance between trajectories and other constraints.

[0115] During the decoding phase, for the trajectories involved in processes 1 and 2, the trajectories are retrieved from the trajectory library based on the current location of the corresponding system and the corresponding virtual machine. Safety distance D safe Set to ΔT*v sys +R dis ΔT = 1 represents the collision avoidance detection time precision, v sys =2 represents the maximum driving speed of different systems, R dis =10 represents the safe distance required for the system to achieve collision avoidance in a static state. For cooperative collision avoidance, all scheduled trajectories need to be re-interpolated at the same time interval ΔT, i.e., for the i-th operation of the j-th task, the trajectory Tr... i,j Will be according to time [S i,j,l ,S i,j,l +ΔT,S i,j,l +2ΔT,…,C i,j,l Interpolation is performed to obtain the location information at the corresponding time point. Then, the location information at [S] is retrieved. i,j,l C i,j,l Information on other transport trajectories that are being transported within the time interval is used to determine whether the current trajectory meets the constraints (9) with other trajectories. If not, it is considered that the collision avoidance constraint is not met, and the execution time of the current process is delayed by ΔT, i.e., S. i,j,l =S i,j,l+ΔT, and re-determine the constraints and process execution time delay until the constraint (9) is satisfied. For other constraints, the problem is decoded by applying logical constraints. Finally, the individual is decoded and a scheduling scheme is generated.

[0116] Step 4-3: Determine the population's evolutionary pattern and update the population.

[0117] To address the issue of carrier-based aircraft sorties, the following three population evolution methods are proposed.

[0118] 1) Copy

[0119] The parent generation is selected using a roulette wheel method based on the replication probability, and the genes are replicated.

[0120] 2) Cross

[0121] The parent generation is selected using a roulette wheel method, and crossover points are determined. Gene fragments between these crossover points are then cross-crossed. Only the first layer of coding can be cross-crossed. After cross-crossing, the coding is adjusted to meet the requirements of the dispatch process. Subsequently, the second layer of coding is adjusted according to the machine constraints of the dispatch process to ensure machine compatibility.

[0122] 3) Variation

[0123] The parent generation is selected using a roulette wheel method to determine the mutation point, and then the following mutation methods are applied to the two layers of encoding respectively.

[0124] ① The mutation of the first layer of coding: The first layer of coding randomly selects genes for mutation according to the roulette method. After mutation, the missing process needs to be made up by the extra process. The second layer of coding is adjusted according to the mutation point of the first layer of coding to make the machine correspond.

[0125] ② Variation of the second layer of coding: Since the same process can be completed by different machines, the machines used can be varied. However, it should be noted that once the machine for process 2 is determined, the corresponding machine for subsequent processes should also be determined. Therefore, only the machines used for process 1 and process 2 should be varied. When the machine used for process 2 changes, the machines used for subsequent processes should also be adjusted.

[0126] Step 4-4: Iterate through the population and retain the optimal solution to generate the final scheduling scheme.

[0127] Based on the population change pattern determined in step 4-3, the population is iterated. After each iteration, each individual is decoded according to the decoding method in step 4-2, the objective function value of each individual is recorded, and the current optimal code is updated until the iteration ends. Finally, the optimal code is decoded to generate a scheduling scheme.

[0128] The Gantt charts for the time allocation of each dispatch process on each virtual machine are arranged as follows: Figure 7 As shown, virtual machines 1, 2, and 3 are three tractors used for scheduling, and virtual machines 4, 5, and 6 are used for the processing of the traction system to the preparation position (i.e., process 2), and correspond to catapults 7, 8, and 9 respectively. Virtual machines 7, 8, and 9 are three catapults.

[0129] A Gantt chart showing the time schedule for each sortie procedure of each carrier-based aircraft is shown below. Figure 8 As shown, 1-12 represent different carrier-based aircraft.

[0130] The above-described embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method for planning a deck launch scheduling scheme of a carrier-based aircraft formation in a rodless towing mode, characterized in that, Includes the following steps: Step 1: Extract key procedures and virtual machines from the carrier-based aircraft sortie operation process. In the FJSP model of the problem of the formation of the carrier-based aircraft, the carrier-based aircraft is regarded as the processing workpiece, and the catapult and the tractor are regarded as the processing machine. According to the setting of the workpiece and the machine, for a wave of the take-off mission of the carrier-based aircraft, six processing procedures are extracted from the whole take-off process. In the FJSP model of the problem of the formation of the carrier-based aircraft, the carrier-based aircraft is regarded as the processing workpiece, and the catapult and the tractor are regarded as the processing machine. According to the setting of the workpiece and the machine, for a wave of the take-off mission of the carrier-based aircraft, six processing procedures are extracted from the whole take-off process. Step 2: Define the decision variables and constraints, and establish a combinatorial optimization problem. Step 2-1: Clarify the variables involved in the scheduling process and their meanings. The meanings of the variables involved in the scheduling process are described below: Step 2-2: Analyze the constraints involved in each stage and describe them as a mathematical model. 1): Any assignment process It can only be assigned to one machine, that is: (1) 2): For Work should be guaranteed process The completion time is greater than the sum of its start time and the processing time required for that process, that is: (2) Among them, if but otherwise 3): Homework process It needs to be done in the process Then proceed; (3) 4) When the machine used in process 2 is determined, the machine used in subsequent processes should also be determined; (4) 5): When doing homework Once the tractor unit is selected in step 1, it cannot be used for subsequent operations until step 2 is completed. Process 1 service, namely: (5) 6) Each sortie stand can hold a maximum of one aircraft. For the same standby stand, a safe time is set for... This indicates that the preparation position for carrier-based aircraft has been occupied, and the third step of the process has begun. After a certain time, subsequent carrier-based aircraft entering the preparation position will not collide with the previous carrier-based aircraft in the preparation position; (6) 7): If homework After selecting the catapult in step 3, the catapult must be ready for subsequent operations after the cooling phase. Serve; (7) 8): Only one carrier-based aircraft is allowed to be launched from the deck at a time. , (8) 9): Homework In the process The selected trajectory is denoted as ,right , To achieve collision avoidance, it is necessary to [do something] at any time. satisfy: (9) in, To ensure a safe distance between the two systems to prevent collisions during dynamic operation; Steps 2-3: Establish a combinatorial optimization problem based on the objective function and constraints. The objective function is set to minimize the deployment time, i.e. : (10) Step 3: Construct a standard trajectory library based on the deck environment and layout. Step 1 involves the transport trajectory of the tractor; Step 2 involves the transport trajectory of the traction system consisting of the carrier-based aircraft and the tractor; Steps 3 and 5 involve the taxiing trajectory of the carrier-based aircraft. Since other systems are not allowed to pass around the catapult and the launch runway during carrier-based aircraft sorties, a restricted zone must be established. Furthermore, the trajectories involved in Steps 3 and 5 are all collision-free fixed trajectories within the restricted zone; therefore, these two stages do not involve trajectory selection or collision avoidance requirements and do not require processing. Once the carrier-based aircraft deck layout is determined, the aircraft parking positions are fixed, and the catapult is also positioned in a fixed location on the deck. Therefore, it is considered to pre-generate a trajectory library for all transport trajectories. For process 1, the tractor needs to generate a tractor trajectory library containing all the trajectories from the departure preparation position to the parking position. A trajectory, which is recorded as For process 2, it is necessary to generate the traction system from each stop position to all ready-to-deploy positions. These trajectories form a trajectory library, which is then recorded as... During trajectory generation, collision avoidance for static obstacles needs to be considered. Since there may be waiting transport systems in relevant parking spaces, parking positions, and deployment preparation positions, all relevant parking positions, parking positions, deployment preparation positions, and restricted areas are set as obstacle areas during the trajectory database generation process. For dynamic collision avoidance, collaborative trajectory arrangement is achieved by combining a delayed deployment strategy and reasonably arranging the start time of the corresponding procedures during the subsequent decoding process. Step 4: Solve the problem using a genetic algorithm. Step 4-1: Determine the gene encoding method of the genetic algorithm and generate the initial population. Step 4-2: Decode the gene considering cooperative collision avoidance between trajectories and other constraints. During the decoding phase, for the trajectories involved in processes 1 and 2, the trajectories are retrieved from the trajectory library based on the current location of the corresponding system and the corresponding virtual machine; safety distance. Set as , To avoid collisions and improve detection time accuracy, The maximum driving speed for different systems, To maintain a safe distance for collision avoidance when the system is static; for cooperative collision avoidance, all planned trajectories must be spaced at the same time interval. Perform re-interpolation, that is, for the ... The first assignment Each process involves a trajectory Will be according to time Interpolation is performed to obtain the location information at the corresponding time point; then the data is retrieved from... Information on other transport trajectories involved in the transport operation within the time interval is collected. It is then determined whether the current trajectory satisfies the constraints shown in formula (9) with other trajectories. If not, it is considered that the collision avoidance constraint is not met, and the execution time of the current process is delayed. ,Right now Then, the constraint judgment and process execution time delay are re-performed until the constraint shown in formula (9) is met; for other constraints, the problem is decoded by applying logical constraints; finally, the individual decoding is realized and a scheduling scheme is generated; Step 4-3: Determine the population's evolutionary pattern and update the population. Step 4-4: Iterate through the population and retain the optimal solution to generate the final scheduling scheme; Based on the population change pattern determined in step 4-3, the population is iterated. After each iteration, each individual is decoded according to the decoding method in step 4-2, the objective function value of each individual is recorded, and the current optimal code is updated until the iteration ends. Finally, the optimal code is decoded to generate a scheduling scheme.

2. The planning method for a carrier-based aircraft formation deck sortie scheduling scheme under the leverless traction and transportation mode according to claim 1, characterized in that, The six processing steps in step 1 are as follows: Step 1: The tractor is moved to the corresponding parking position of the carrier-based aircraft. In this stage, the tractor is moved from the parking position corresponding to the departure preparation position to the parking position corresponding to the aircraft to be deployed. The virtual processing machine used in this step is the tractor. Any tractor can be selected to complete this step. The number of machines that can complete this step is denoted as . ; Step 2: Transport the carrier-based aircraft to the sortie ready position; In this stage, the tractor and the carrier-based aircraft form a traction system and transport the carrier-based aircraft to the sortie ready position corresponding to the catapult. The carrier-based aircraft can be transported to any ready position; The machine used in this stage is the catapult, but considering that the carrier-based aircraft only occupies the corresponding resources when it arrives at the ready position during the sortie process, virtual machines with the same number as the catapults are set up for this stage. The number of virtual machines that can complete this process is denoted as . ; Step 3: Carrier-based aircraft taxi to the catapult; In this stage, the carrier-based aircraft taxis from the sortie preparation position to the corresponding catapult. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as . ,have ; Step 4: Pre-flight checks of carrier-based aircraft; This step involves the pilot checking the status of the carrier-based aircraft before takeoff, and the selected catapult performing pre-flight preparations. The virtual machine used in this step is the catapult, and the number of machines capable of completing this step is denoted as [number missing]. , ; Step 5: Carrier-based aircraft sortie. After confirming that the carrier-based aircraft is in normal condition and the catapult is ready, the aircraft is launched. The virtual machine used in this stage is the catapult. The number of machines that can complete this step is denoted as . , ; Step 6: Cooling and Retraction of Deflector Plates. After the carrier-based aircraft takes off, the corresponding deflector plates should be cooled and retracted to their initial positions to ensure that subsequent carrier-based aircraft can taxi to the catapult and take off. The virtual machine used in this step is the catapult. The number of machines that can complete this step is denoted as [number missing]. , .

3. The planning method for a carrier-based aircraft formation deck sortie scheduling scheme under a leverless traction and transportation mode according to claim 1, characterized in that, Step 4-1 specifically involves: solving the problem using a genetic algorithm with a two-layer coding method based on the characteristics of carrier-based aircraft sortie missions; the coding is divided into two layers, using integer coding, with the upper layer representing the workpiece and corresponding process, and the lower layer representing the machine used for the corresponding process in the upper layer; the upper layer has multiple bits of coding, the number of which is 6, the number of processes. Number of carrier-based aircraft deployed Each job number will appear 6 times in the code, and the order will be randomized. The number of times the job number repeats in the code represents the number of processes. For lower-level codes, the corresponding processes will first be assigned machine utilization numbers from 1 to 6. Numbering is performed, and then machine numbers are generated from the set of machines capable of performing the process based on the number of processes represented by the upper-level code; the code generation process is repeated. Next, generate with The initial population of individuals.

4. The planning method for a carrier-based aircraft formation deck sortie scheduling scheme under a leverless traction and transportation mode according to claim 1, characterized in that, In step 4-3, the following three population evolution methods are set to address the issue of carrier-based aircraft sorties: 1) Copy The parent generation is selected using a roulette wheel method based on the replication probability, and the gene is replicated. 2) Cross The parent generation is selected using a roulette wheel method, and the crossover points are determined. Gene fragments between the crossover points are then cross-crossed. Only the first layer of coding can be cross-crossed. After cross-crossing, the coding is adjusted to meet the requirements of the dispatch process. Subsequently, the second layer of coding is adjusted according to the machine constraints of the dispatch process to ensure that the machine is compatible. 3) Variation The parent generation is selected using a roulette wheel method to determine the mutation point, and then the following mutation methods are applied to the two layers of coding respectively; ① The mutation of the first layer of coding: The first layer of coding randomly selects genes for mutation according to the roulette method. After mutation, the missing process needs to be made up by the extra process. The second layer of coding is adjusted according to the mutation points of the first layer of coding to make the machine correspond. ② Variation of the second layer of coding: Since the same process can be completed by different machines, the machines used can be varied. However, it should be noted that once the machine for process 2 is determined, the corresponding machine for subsequent processes should also be determined. Therefore, only the machines used for process 1 and process 2 should be varied. When the machine used for process 2 changes, the machines used for subsequent processes should also be adjusted.

Images