Killer chain construction method based on particle swarm optimization

Abstract

The invention discloses a killing chain construction method based on a particle swarm algorithm, and belongs to the technical field of cooperative combat systems. The method comprises the following steps: dividing a killing chain construction problem into four links of detection, positioning, interference and strike which are executed in sequence, respectively establishing corresponding task allocation models, and taking maximization of efficiency of each link as a target function; and solving each link model by adopting an improved decimal discrete particle swarm algorithm to obtain an optimal task allocation scheme, and sequentially combining all link schemes to complete the construction of the killing chain. According to the method, the modeling process is simplified, the calculation efficiency is improved, the method is suitable for a large-scale combat scene, an optimization scheme can be quickly generated, and the requirements of actual combat for real-time performance and accuracy are effectively met.

Description

Technical Field

[0001] This invention relates to the field of cooperative combat system technology, and specifically to a kill chain construction method based on particle swarm optimization algorithm. Background Technology

[0002] The kill chain, a product of information warfare, is an integrated operational mission chain encompassing reconnaissance, surveillance, intelligence, computing, communication, command, control, and strike elements, achieving "from detection to destruction." Specifically, it's a closed-loop mission pattern where, against a specific type of target, each link element operates sequentially and interdependently based on a pre-planned fixed architecture, producing a linear kill effect on the target. With the development of information technology, modern warfare is rapidly evolving into information warfare, shifting from traditional high-tech platform competition to weapon system confrontation. The fulfillment of operational system missions depends on a series of kill chains; therefore, research on kill chains is of great significance. The key to the formation and application of kill chains lies in optimizing operational tasks and resource scheduling, achieving the matching of different operational tasks with varying performance of operational resources. While researchers both domestically and internationally have conducted extensive theoretical research on these issues, shortcomings remain. Some literature uses traversal methods to generate kill chains, which is limited in the scale of application scenarios. Other literature presents complex kill chain modeling and calculation processes, making it difficult to cope with rapidly changing battlefield situations.

[0003] Reference 1: Gao Baohui, Hu Hai, Zhong Zhitong, et al. A method for constructing kill chains for space-based guided anti-ship strikes [J]. Journal of Command and Control, 2024, 10(02): 184-196. This paper proposes a method for constructing kill chains for space-based guided anti-ship strikes. The general operational process of space-based guided anti-ship strike kill chains is clarified, a kill chain model is established, and a kill chain construction algorithm is designed, namely, traversing and finding all kill chains for striking each target, and sorting and outputting the kill chains according to the performance index evaluation method. However, this method has an excessive computational load when facing large-scale combat scenarios, making it difficult to obtain results quickly.

[0004] Reference 2: Wan Silai, Wang Guoxin, Ming Zhenjun, et al. Kill chain modeling and optimization method based on AGE-MOEA [J]. Acta Ordnanceica Sinica, 2024, 45(8): 2617-2628. A multi-objective optimization mathematical model for kill chain design is proposed, with equipment usage constraints, kill chain relationship constraints, and damage threshold constraints as constraints, to realize the mathematical representation of the kill chain design problem. A kill chain optimization design method based on AGE-MOEA is proposed to solve the multi-objective optimization mathematical model for kill chain design. The modeling and calculation process of the kill chain in this method is relatively complex, and it is difficult to cope with the rapidly changing battlefield situation in practical applications.

[0005] The above research has the following problems:

[0006] First, the use of a traversal method to generate the kill chain is limited in the scale of application scenarios;

[0007] Secondly, the modeling and calculation of the kill chain is quite complex and difficult to cope with rapidly changing battlefield situations. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a kill chain construction method based on particle swarm optimization (PSO) algorithm, belonging to the technical field of cooperative combat systems. Addressing the deficiencies of existing technologies, primarily the complexity of kill chain modeling and calculation, poor real-time performance, and the low computational efficiency of traversal methods, making them unsuitable for large-scale combat scenarios, this invention divides the kill chain construction problem into four sequentially executed stages: detection, localization, jamming, and attack. Corresponding task allocation models are established for each stage, with maximizing the effectiveness of each stage as the objective function. An improved decimal discrete particle swarm optimization algorithm is used to solve the model for each stage, obtaining the optimal task allocation scheme. The schemes for each stage are then sequentially combined to complete the kill chain construction. This invention simplifies the modeling process, improves computational efficiency, is applicable to large-scale combat scenarios, can quickly generate optimized schemes, and effectively meets the real-time and accuracy requirements of actual combat.

[0009] A kill chain construction method based on particle swarm optimization includes the following steps:

[0010] Step 1, Kill Chain Stage Division: The kill chain is divided into four stages: detection stage, location stage, jamming stage, and strike stage.

[0011] The detection phase: The application of sensors to detect unknown areas is the foundation for the location, jamming, and strike phases; the sensors include radar sensors, ESM sensors, and photoelectric sensors;

[0012] The positioning phase involves accurately and continuously locating the target based on the target information returned from the detection phase.

[0013] The jamming phase: The jammer releases jamming on the target to cover the breakthrough formation's breakthrough;

[0014] The strike phase: The target is struck based on its location;

[0015] Step 2, Kill Chain Problem Modeling:

[0016] For the detection phase, localization phase, jamming phase, and strike phase, separate detection task allocation models, localization task allocation models, jamming task allocation models, and strike task allocation models are established for each phase.

[0017] Step 3, Kill Chain Construction Algorithm:

[0018] The kill chain construction algorithm is a kill chain construction method based on particle swarm optimization, and the particle swarm optimization algorithm is a decimal discrete particle swarm optimization algorithm.

[0019] Step 4: The decimal discrete particle swarm optimization algorithm is used to solve the detection task allocation model, the localization task allocation model, the jamming task allocation model, and the strike task allocation model to obtain the optimal allocation schemes for detection tasks, localization tasks, jamming tasks, and strike tasks, respectively.

[0020] Step 5: Combine the optimal allocation schemes for detection, location, jamming, and strike missions in sequence to construct a kill chain.

[0021] Furthermore, in step 2, the method for constructing the detection task allocation model is as follows:

[0022] The detection methods for the detection mission include ESM dual-machine detection, radar single-machine or radar dual-machine detection, and electro-optical single-machine or electro-optical dual-machine detection; ESM dual-machine detection is used in the long-range detection phase, and electro-optical or radar detection is used in the short-range phase; the long range is a position greater than 50km, and the short range is a position less than or equal to 50km;

[0023] The exploration mission allocation scheme is as follows ,in, The aircraft number used to perform ESM detection; The aircraft number used for photoelectric detection; The aircraft number used for radar detection; For the flag bit: x7 can be either 1 or 2;

[0024] like If the value is 1, the detection task allocation scheme is ESM dual-machine detection and photoelectric detection;

[0025] like If the value is 2, the detection task allocation scheme is ESM dual-machine detection and radar detection;

[0026] The objective function s of the detection task allocation model is to select a detection scheme with the goal of maximizing detection effectiveness. det for:

[0027] ;

[0028] Where: n' is the number of sensor types currently enabled; This enhances the dual-machine detection performance of the ESM. For photoelectric detection efficiency; for Radar detection effectiveness; ; ; ; The number of ESM sensors; This refers to the number of photoelectric sensors; The number of radars;

[0029] like Then photoelectric single-machine detection is used; if This method uses single-unit radar detection.

[0030] Furthermore, in step 2, the method for constructing the location task allocation model is as follows:

[0031] Positioning methods include ESM dual-machine positioning, photoelectric dual-machine positioning, and radar single-machine or dual-machine positioning;

[0032] The location task allocation scheme (Loc) is represented as an n×3 matrix:

[0033] ;

[0034] The nth row represents the resources allocated to our side for the nth target; and This indicates the machine number used to locate the target. That is, single-machine positioning; otherwise, dual-machine positioning; x 13 and x n3 These are all flag bits, with values ​​ranging from 1, 2, or 3. A value of 1 selects ESM positioning; a value of 2 selects photoelectric positioning; and a value of 3 selects radar positioning. 11 Indicates the platform number of the first platform performing the positioning task; x 12 This indicates the second platform number that performed the positioning task;

[0035] The objective function of the location task allocation model is s. loc :

[0036] ;

[0037] ;

[0038] Where n is the target quantity; Improve the positioning efficiency of ESM dual-machine system; For photoelectric positioning efficiency; To improve radar positioning efficiency; ; ; ;like In this case, radar single-unit positioning is used; j represents the j-th target;

[0039] Furthermore, in step 2, the method for constructing the interference task allocation model is as follows:

[0040] In a scenario where our jammers are fewer than the number of targets, assume that each jammer can jam at most two targets, and each target is jammed by only one jammer.

[0041] First, allocate our resources based on the distance to the target as the objective function, and interfere with the nearest target. For the remaining targets, allocate our resources based on the interference effectiveness when our aircraft interferes with both the already interfered target and the target to be interfered with.

[0042] The interference task allocation scheme is represented as follows: ,in, This indicates the jammer number used to jam the nth target;

[0043] Let m be the number of our jamming devices; and n be the number of targets. This represents the distance between our platform i and the target j; This represents the interference effectiveness of platform i relative to target j; This indicates that target j is being interfered with by our platform i;

[0044] The objective function of the task allocation model is s int This includes two scenarios:

[0045] (1) If :

[0046] ;

[0047] (2) If :

[0048] ;

[0049] The objective function s int The constraints are:

[0050] ;

[0051] in, This indicates the maximum number of radars that a single jammer can jam.

[0052] Furthermore, in step 2, the method for constructing the strike mission allocation model is as follows:

[0053] Assume that each platform can attack a maximum of two targets, each target is attacked by only one platform, and one missile is launched for each attack.

[0054] The strike mission allocation scheme is represented as follows: ,in, This indicates the platform number used by our side to strike the nth target.

[0055] The objective function of the strike mission allocation model for:

[0056] ;

[0057] Where m represents the number of our fighter jets; This represents the overall situational advantage value of platform i relative to target j; This indicates that target j was attacked by our platform i;

[0058] Objective function of the strike mission allocation model That is, to achieve the highest combat effectiveness with the least attack cost;

[0059] The objective function The constraints are:

[0060] ;

[0061] in, This indicates the number of our platforms required to destroy target j; The maximum number of targets that can be attacked for each platform.

[0062] Furthermore, in step 3, the process of improving the discrete particle swarm algorithm to a decimal discrete particle swarm algorithm is as follows:

[0063] Aircraft are numbered using natural numbers, and the position vector of each particle represents the aircraft number performing a specific mission. The particle velocity update formula for the decimal discrete particle swarm optimization algorithm is the same as that for the binary discrete particle swarm optimization algorithm, and the particle position... The updated formula is as follows:

[0064]

[0065] The round() function rounds the integers according to the rounding rules. yes Updated particle positions, It is the current particle position; It is a sigmoid function mapping of particle update velocity;

[0066] ;

[0067] Particle position The range of values ​​for is as follows:

[0068] ;

[0069] The position of the j-th dimension particle The maximum value of is the maximum number of our platforms executing each stage of the task; the particle position in the j-th dimension represents the task allocation scheme for the j-th target; e is a constant; v ij This indicates the update speed of the position.

[0070] Furthermore, in step 4, the process of obtaining the optimal allocation scheme for the detection tasks is as follows:

[0071] Step 4.1.1: Input the mission area coordinates, available platform coordinates, sensor type, detection radius, and platform velocity parameters into the detection mission allocation model, and solve it using the particle swarm optimization algorithm;

[0072] Step 4.1.2: Generate random particle position vectors, which are feasible solutions for the detection task allocation scheme;

[0073] Step 4.1.3, calculate Corresponding ESM detection performance Find Corresponding photoelectric detection efficiency Find Corresponding radar detection effectiveness ;

[0074] Compare and :like ≥ , Set to 1; if < , Set to 2;

[0075] The fitness value of a particle is calculated, which represents the effectiveness of the current detection scheme.

[0076] Step 4.1.4 compares the individual historical optimal value of a particle with the global optimal value of the swarm, causing the particle to continuously update its position and move closer to the optimal solution.

[0077] In step 4.1.5, when the maximum number of iterations is reached, the algorithm terminates and the optimal allocation scheme for the probe tasks is obtained; otherwise, return to step 4.1.3.

[0078] Furthermore, in step 4, the process of obtaining the optimal allocation scheme for the location task is as follows:

[0079] The optimal allocation scheme for the localization task is obtained by using the particle swarm optimization algorithm;

[0080] Step 4.2.1: Read the friendly platform information from the battlefield situation; the friendly platform information includes the number of friendly platforms, longitude, latitude, and altitude;

[0081] Step 4.2.2: Read the target information for detection; the target information includes the number of targets, the approximate target accuracy, latitude, and altitude;

[0082] Step 4.2.3: Input target information; using the particle swarm optimization algorithm, with ESM positioning performance as the objective function, derive the optimal ESM dual-machine positioning scheme and positioning performance. The optimal photoelectric positioning scheme and positioning performance are derived by taking photoelectric positioning efficiency as the objective function. The optimal radar positioning scheme and positioning effectiveness are derived by using radar positioning effectiveness as the objective function. ;

[0083] Compare , and Select the optimal allocation scheme for the positioning task and place the flag position on the corresponding number; if the sensor allocation sequence is empty, the positioning efficiency is 0 and the positioning allocation scheme is empty.

[0084] Step 4.2.4: Determine whether the number of times the selected platform has been reused has exceeded (excluding "equal to") the maximum number of location targets. If the number of targets is not exceeded, input the information for the next target and continue the allocation; if the number of targets is exceeded, remove the queue from the allocation sequence and return to step 4.2.3 for reallocation.

[0085] Step 4.2.5: Determine whether all targets have been assigned a positioning scheme: if yes, the algorithm ends; otherwise, return to step 4.2.3 to continue the assignment.

[0086] Furthermore, in step 4, the process of obtaining the optimal allocation scheme for the interference tasks is as follows:

[0087] Step 4.3.1: Read the information of surviving friendly platforms from the enemy-friendly situation in the scenario simulation system, including the longitude, latitude, altitude and azimuth of the friendly platforms, as alternative platforms for interference targets;

[0088] Step 4.3.2: Identify the interference target. The interference target is the target being attacked in the current stage. Read the target information obtained from sensor cooperative detection and cooperative localization, including the enemy target's longitude, latitude, altitude, and azimuth.

[0089] Step 4.3.3: First, select the target that is currently under attack and has not been assigned. Use the distance between our machine and the currently selected target as the objective function, and use the particle swarm algorithm to prioritize the target that is closest to our machine for interference.

[0090] Step 4.3.4: For targets that have not been allocated our resources in the allocation with distance as the objective function, the particle swarm optimization algorithm is used to select our resources with the interference effectiveness when our machine interferes with both the already interfered targets and the targets to be interfered with as the objective function.

[0091] Restriction: If the angle between my aircraft and the target to be jammed and the assigned jamming target is less than the maximum jamming angle, then both aircraft will be jammed simultaneously; otherwise, both aircraft cannot be jammed simultaneously, and the jamming effectiveness will be 0.

[0092] Step 4.3.5: Determine whether all targets have been assigned a positioning scheme: if yes, the algorithm ends; otherwise, return to step 4.3.4 to continue the assignment.

[0093] Furthermore, in step 4, the process of obtaining the optimal allocation scheme for strike missions is as follows:

[0094] The target allocation criterion of the strike mission allocation model is to minimize the cost while meeting the damage requirements. The steps of the strike mission allocation algorithm are as follows:

[0095] Step 4.4.1: Read the friendly platform information from the enemy and friendly situation in the scenario simulation system, including the longitude, latitude, altitude, speed, heading and number of weapons of the friendly platform;

[0096] Step 4.4.2: Read the target information obtained from sensor cooperative detection and cooperative localization, including the enemy target's longitude, latitude, altitude, speed, and heading;

[0097] Step 4.4.3: Calculate the threat index by weighting the angle threat index, speed threat index, altitude threat index, and distance threat index of the targets to be assigned to our platform; sort the targets according to the sum of their threat indices to all our platforms, and assign them in the order of the sorted targets.

[0098] Step 4.4.4: Using the particle swarm optimization algorithm, with the strike effectiveness of our aircraft against the target as the objective function, we allocate our resources to the target, and guidance is completed by the aircraft carrying out the strike.

[0099] Step 4.4.5: Determine whether the number of times the selected platform can be reused has exceeded the maximum number of targets to be hit: If it has not exceeded the maximum number of targets to be hit, enter the information of the next target and continue the allocation; if it has exceeded the maximum number of targets to be allocated, remove the platform from the allocation sequence and return to step 4.4.4 for reallocation.

[0100] Step 4.4.6: Determine whether all targets have been allocated our resources: if yes, the algorithm ends; otherwise, return to step 4.4.4 to continue allocation.

[0101] Furthermore, in step 5, the kill chain targets the target. The kill chain is represented as:

[0102] ;

[0103] Among them, targeting The kill chain; The carrier number represents the aircraft that will carry out the collaborative detection mission, which is the optimal detection scheme derived by the particle swarm algorithm. The carrier number represents the optimal positioning scheme obtained by the particle swarm algorithm. The carrier number representing the aircraft performing the jamming task is the optimal jamming scheme derived by the particle swarm algorithm. The aircraft number representing the carrier mission is the optimal strike plan derived by the particle swarm optimization algorithm.

[0104] A kill chain construction system based on particle swarm optimization algorithm is provided. The kill chain construction method based on particle swarm optimization algorithm is implemented to realize the kill chain construction based on particle swarm optimization algorithm. It is divided into two modules, and steps 1 to 5 are executed respectively.

[0105] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the kill chain construction method based on particle swarm optimization algorithm to achieve kill chain construction based on particle swarm optimization algorithm.

[0106] A computer-readable storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, the aforementioned kill chain construction method based on particle swarm optimization algorithm is implemented to realize kill chain construction based on particle swarm optimization algorithm.

[0107] Compared with the prior art, the significant advantage of this invention is that the modeling process is simple and easy to implement, and the introduction of particle swarm optimization algorithm makes it applicable to large-scale combat scenarios, enabling rapid generation of kill chain construction schemes and effectively meeting the real-time and accuracy requirements of actual combat. Attached Figure Description

[0108] Figure 1 It is a diagram showing the stages and modeling of the kill chain;

[0109] Figure 2 This is a flowchart of the detection task allocation algorithm;

[0110] Figure 3 This is a flowchart of the location task allocation algorithm;

[0111] Figure 4 This is a flowchart of the interference task allocation algorithm;

[0112] Figure 5 This is a flowchart of the task allocation algorithm. Detailed Implementation

[0113] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0114] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0115] Step S1, Kill Chain Stage Division and Modeling:

[0116] The kill chain construction problem is modeled as a task allocation problem involving four sequentially executed stages: detection, localization, jamming, and strike. For each of the detection, localization, jamming, and strike stages, a corresponding task allocation model is established, with each model taking the maximization of its stage effectiveness as the objective function.

[0117] In step S1, the kill chain modeling process is as follows:

[0118] The kill chain refers to the closed-loop operational sequence of striking a specific target or object, supported by intelligence and based on command and control information. The basic attributes of the kill chain include closure, orderliness, and dependency, with its core objective being to contribute to the effectiveness of achieving the kill effect. In the 1990s, based on the OODA loop theory, the ordered link consisting of six operational stages—find, fix, track, target, engage, assess (F2T2EA)—was defined as the kill chain for analyzing the attack process on a target, attracting widespread attention from scholars.

[0119] Unlike the six-stage kill chain (detection-location-tracking-aiming-engagement-assessment), this invention simplifies and consolidates the kill chain into four stages—detection, location, jamming, and strike—to emphasize efficiency. The assessment of kill chain effectiveness serves as the objective function throughout the decision-making process, supporting resource allocation.

[0120] The meanings of each stage are as follows:

[0121] Detection phase: Using sensors such as radar, ESM sensors, and photoelectric sensors to detect unknown areas is the foundation for the location, jamming, and strike phases.

[0122] Localization phase: Based on the target information transmitted back from the detection phase, the target is located accurately and continuously;

[0123] Jamming Phase: Jammers jam the target to cover the penetration formation's breakthrough.

[0124] Strike phase: Strike the target based on its location;

[0125] The above four stages can cover reconnaissance, surveillance, intelligence, computing, communication, command, control, and strike elements, realizing an integrated combat mission chain "from detection to destruction". Therefore, the kill chain can be constructed by completing the task allocation of the detection, location, jamming and strike stages according to the execution order.

[0126] (1) Task allocation model in the exploration phase

[0127] Based on factors such as platform location, sensor performance, and target area, a detection allocation model is established to generate a collaborative detection scheme, determining the platforms and sensor operating modes participating in the collaborative detection. To ensure ESM detection accuracy, a dual-UAV ESM detection method is adopted; electro-optical detection can be performed single-aircraft or dual-aircraft; radar detection has two operating modes: collaborative range extension and non-collaborative. The collaborative range extension mode requires two UAVs to cooperate, which can increase the detection range. This invention only considers airspace coverage when allocating detection platforms to the target area. Based on the detection coverage area, time, and detection cost, a detection scheme is generated, and the detection effectiveness is output. In the entire combat process, to achieve stealthy penetration and ensure the safety of manned aircraft, the formation first activates ESM collaborative detection, and then selects to activate radar sensors or electro-optical sensors for replacement based on detection effectiveness.

[0128] The exploration mission allocation scheme is represented as follows: ,in:

[0129] The aircraft number used to perform ESM detection;

[0130] The aircraft number used for photoelectric detection;

[0131] The aircraft number used for radar detection;

[0132] For flag bits:

[0133] like If the value is 1, the detection scheme is ESM dual-machine detection + photoelectric detection;

[0134] like If the value is 2, the detection scheme is ESM dual-machine detection + radar detection;

[0135] The objective function of the detection task allocation model, which selects a detection scheme with the goal of maximizing detection effectiveness, is:

[0136]

[0137] in:

[0138] 'Indicates the number of sensor types currently enabled;

[0139] This enhances the dual-machine detection performance of the ESM.

[0140] For photoelectric detection efficiency;

[0141] for Radar detection effectiveness;

[0142] ;

[0143] ;

[0144] ;

[0145] in:

[0146] The number of ESM sensors;

[0147] This refers to the number of photoelectric sensors;

[0148] The number of radars;

[0149] like Then photoelectric single-machine detection is used; if This method uses single-unit radar detection.

[0150] (2) Task allocation model in the positioning phase

[0151] Location refers to determining the precise position of a target, tracking and locking onto enemy targets in order to launch an attack or take other actions. Detection and location are interrelated in air combat. Detection provides information on the presence and initial position of the target, while location, through further analysis and calculation, determines the target's precise position. Together, they provide crucial information for operational decisions and actions in air combat. The location methods employed in this invention include ESM dual-aircraft location, electro-optical dual-aircraft location, and radar single-aircraft or dual-aircraft location.

[0152] The location task allocation scheme (Loc) is represented as an n×3 matrix:

[0153]

[0154] The nth row represents the resources allocated to our side for the nth target:

[0155] and This indicates the machine number used to locate the target;

[0156] like This refers to single-machine positioning;

[0157] Otherwise, it is a dual-machine positioning system;

[0158] For flag bits:

[0159] like If so, then select ESM detection;

[0160] like In this case, photoelectric detection is selected;

[0161] like If so, then radar detection will be selected;

[0162] The objective function of the location task allocation model is:

[0163]

[0164]

[0165] in:

[0166] n is the target quantity;

[0167] Improve the positioning efficiency of ESM dual-machine system;

[0168] For photoelectric positioning efficiency;

[0169] To improve radar positioning efficiency;

[0170] ;

[0171] ;

[0172] ;

[0173] like It uses radar single-unit positioning.

[0174] (3) Task allocation model during the interference phase

[0175] Based on factors such as platform performance, sensor performance, and target radar information, the cover formation selects mission platforms to form a jamming mission formation, achieving coordinated jamming of the target. During coordinated jamming, the jamming power and the required number of platforms will vary depending on the jamming suppression distance to meet the jamming area requirements. Therefore, it is necessary to select appropriate mission platforms through task allocation to achieve optimal configuration under constraints of jamming effectiveness, energy loss, and the number of cooperating platforms.

[0176] In a scenario where our jammers are fewer than the number of targets, each jammer can jam at most two targets, and each target is jammed by only one jammer. First, using the distance to the target as the objective function, we allocate our resources using particle swarm optimization (PSO) to jam the nearest target. For the remaining targets, we use PSO to select our resources based on the jamming effectiveness when our machine jams both the already jammed target and the target to be jammed.

[0177] The interference task allocation scheme is represented as follows: ,in This indicates the jammer number used to jam the nth target;

[0178] set up:

[0179] m represents the number of our jamming devices;

[0180] n is the number of targets;

[0181] This represents the distance between platform i and target j;

[0182] This represents the interference effectiveness of platform i relative to target j;

[0183] This indicates that target j is being interfered with by our platform i;

[0184] The objective function of the task allocation model is discussed in two cases:

[0185] like

[0186]

[0187] like

[0188]

[0189] The constraints are:

[0190]

[0191] in This indicates the maximum number of radars that a single jammer can jam. .

[0192] (4) Task allocation model during the strike phase

[0193] This paper establishes a mathematical programming model and uses the particle swarm optimization algorithm to optimize the allocation of combat resources based on the performance of our weapons, thereby improving combat effectiveness and reducing attack costs. It is assumed that each platform can attack at most two targets, each target is attacked by only one platform, and one missile is launched per attack.

[0194] The strike mission allocation scheme is represented as follows: ,in This indicates the platform number used by our side to strike the nth target.

[0195] The objective function of the strike mission allocation model is:

[0196]

[0197] in:

[0198] m represents the number of our fighter jets;

[0199] n is the number of enemy targets;

[0200] This represents the overall situational advantage value of platform i relative to target j;

[0201] This indicates that target j was attacked by our platform i;

[0202] That is, to achieve the highest combat effectiveness with the least attack cost;

[0203] The constraints are:

[0204]

[0205] in:

[0206] This indicates the number of our platforms required to destroy target j. ;

[0207] The maximum number of targets that can be attacked for each platform, here. .

[0208] Step S2, Design of Kill Chain Construction Method Based on Particle Swarm Optimization: Improve the discrete particle swarm optimization algorithm to a decimal discrete particle swarm optimization algorithm, and use the decimal discrete particle swarm optimization algorithm to solve the task allocation model of each link established in step 1, and obtain the optimal allocation scheme of detection, localization, jamming and strike tasks respectively; combine the optimal allocation schemes of detection, localization, jamming and strike tasks in sequence to complete the construction of the kill chain.

[0209] In step S2, the kill chain construction algorithm is as follows:

[0210] First, the decimal discrete particle swarm optimization algorithm.

[0211] Unlike traditional methods that use the standard Particle Swarm Optimization (PSO) algorithm for task allocation, this invention innovatively applies a decimal discrete PSO algorithm to achieve rapid allocation and complete construction of tasks at each stage of the kill chain. Compared to the standard PSO algorithm, which typically represents the execution plan of each task as a string of binary code (marking whether aircraft in a certain order execute a task by setting corresponding positions to 0 or 1), this invention directly uses natural numbers (decimal) to encode the aircraft numbers executing a certain type of task. When dealing with a large number of aircraft, this method can significantly shorten the encoding length, reduce computational complexity, and thus improve algorithm efficiency and scalability.

[0212] The particle position update formula is:

[0213]

[0214] The round() function rounds the integer according to the rules for rounding to the nearest integer. It is the particle's position at the next moment. It is the current particle position. It is a sigmoid function mapping of particle update velocity. The calculation formula and the range of values ​​for the particle position are as follows:

[0215]

[0216] It represents the maximum value of the j-th dimension x, that is, the maximum number of our platform executing a certain task.

[0217] Second, application scenarios

[0218] (1) Combat scenario setting

[0219] In a coordinated air defense mission scenario, our nine aircraft, each carrying six missiles, include:

[0220] Three manned aircraft carrying radar;

[0221] Four drones equipped with ESM and radar;

[0222] Two drones equipped with electro-optical systems;

[0223] The enemy has 6 aircraft, each carrying 6 missiles, and the type and number of sensors they carry are unknown. The air combat zone is at an altitude of 8 kilometers, with both sides' aircraft entering from their respective air combat positions at opposite ends.

[0224] (2) Sensor parameter settings

[0225] Detection radius of ESM and photoelectric sensors: 80km

[0226] Radar detection radius: 160km

[0227] Aircraft speed: 0.35 km / s

[0228] Photoelectric detection angle θ: 10°

[0229] Size of the area to be detected: 50km long, 30km wide

[0230] (3) Particle Swarm Algorithm Parameter Settings

[0231] Number of particles: 30

[0232] Maximum number of iterations: 20

[0233] Acceleration constants c1 and c2: 1.5

[0234] Inertia weight w: dynamically changes from 0.6 to 0.8

[0235] Third, the design of the detection task allocation algorithm.

[0236] (1) Detection task allocation algorithm flow

[0237] The particle swarm optimization algorithm is used to derive the probe task allocation scheme. The steps of the algorithm are as follows:

[0238] S3-1: Input parameters such as mission area coordinates, available platform coordinates, sensor type, and detection radius;

[0239] S3-2: Generate random particle position vectors, which is a feasible solution for the detection task allocation scheme;

[0240] S3-3: Find Corresponding ESM detection performance , Corresponding photoelectric detection efficiency , Corresponding radar detection effectiveness .Compare and ,like , Set to 1, if , Set to 2; calculate the particle's fitness value, i.e., the effectiveness of the current detection scheme;

[0241] S3-4: Compare the individual historical optimal value of the particle with the global optimal value of the group to make the particle continuously update its position and move closer to the optimal solution;

[0242] S3-5: When the maximum number of iterations is reached, the algorithm terminates and the optimal allocation scheme for the probe tasks is obtained; otherwise, return to S3-3.

[0243] (2) Detection performance calculation function

[0244] The formula for calculating the effectiveness of the detection scheme is as follows:

[0245]

[0246] in, To detect time efficiency, To explore cost-effectiveness, It is the weight, let .

[0247] The formula for calculating detection time performance is as follows:

[0248]

[0249] in For the detection time, For host speed, The distance between the host and the task area when the sensor is activated.

[0250] The formula for calculating the cost-effectiveness of detection is as follows:

[0251]

[0252] in, To determine the number of our aircraft used. This refers to the total number of our aircraft.

[0253] (3) Example of detection task allocation algorithm

[0254] Analysis of our detection resources revealed the following: ESM: 4 UAVs; Electro-optical: 2 UAVs; Radar: 3 manned aircraft and 4 UAVs. ;

[0255] Search space: , .

[0256] The coordinates of each sensor position are shown in Table 1.

[0257] Table 1. Coordinates of each sensor

[0258]

[0259] The algorithm was run using MATLAB R2018b, and the calculation results are as follows: Global optimal position vector: global_best=[1, 2, 2, 1, 6, 6, 2], corresponding to a global optimal objective function value of 4.2318. The detection scheme at this time is ESM detection + radar single-unit detection. }

[0260] Fourth, the design of the localization task allocation algorithm.

[0261] (1) Location task allocation algorithm process

[0262] The particle swarm optimization algorithm is used to derive the localization task allocation scheme. The steps of the algorithm are as follows:

[0263] S4-1: Read information about friendly platforms from the battlefield situation, including the number of friendly platforms, longitude, latitude, and altitude;

[0264] S4-2: Read target information for reconnaissance, including the number of targets, rough target accuracy, latitude, and altitude;

[0265] S4-3: Input Target Based on the information, the particle swarm optimization algorithm is applied with ESM positioning performance as the objective function to derive the optimal ESM dual-machine positioning scheme and positioning performance. The optimal photoelectric positioning scheme and positioning performance are derived by taking photoelectric positioning efficiency as the objective function. The optimal radar positioning scheme and positioning effectiveness are derived by using radar positioning effectiveness as the objective function. The efficiency values ​​of the three schemes are compared, and the optimal positioning scheme is selected as the final positioning scheme, with the marker position placed at the corresponding number. If the sensor's assigned sequence is empty, the positioning efficiency is 0, and the positioning allocation scheme is empty.

[0266] S4-4: Determine whether the number of times the selected formation has been reused has exceeded the maximum number of target locations. If there is no excess, input the information of the next target to continue the allocation; if there is excess, remove the queue from the allocation sequence and return to S4-3 for reallocation.

[0267] S4-5: Determine whether all targets have been assigned a positioning scheme. If so, the algorithm ends; otherwise, return to S4-3 to continue the assignment.

[0268] (2) Positioning effectiveness calculation function

[0269] Target The formula for calculating positioning effectiveness is as follows:

[0270]

[0271] in, For accuracy performance, it is related to the accuracy of the sensor; To determine the cost; Let be the weight, ; For time cost, The calculation formula is as follows:

[0272]

[0273] Where t represents time loss, which is the time for adjusting the lateral spacing between the two machines or the radar sensing response time.

[0274] (3) Example of a location task allocation algorithm

[0275] With the positions of all sensors and the aircraft speed remaining constant, the maximum number of targets that each formation can locate is [number missing]. Set to 2, the error on both sides of the target azimuth angle measured by ESM. The value is set to 0.15 rad, the photoelectric positioning error Qe is set to 2 km, and the radar sensing response time is set to... Set to 5 microseconds, the horizontal distance between the master station and the auxiliary station. The range is 16 km. The coordinates of the detected target location are shown in Table 2.

[0276] Table 2 Target coordinates obtained from detection

[0277]

[0278] The algorithm was run using MATLAB R2018b software, and the calculation results are shown in Table 3:

[0279] surface The calculated positioning task allocation scheme and positioning performance

[0280]

[0281] All possible combinations can be obtained through enumeration algorithms. The top five localization task allocation schemes with the best localization performance for each target are shown in Tables 4 to 9:

[0282] Table 4 shows the top five positioning task allocation schemes for target 1, obtained using the enumeration method.

[0283]

[0284] Table 4 shows that the optimal solution for locating target 1 is... and .

[0285] Table 5 shows the top five positioning task allocation schemes for target 2, obtained using the enumeration method.

[0286]

[0287] Table 5 shows that the optimal solution for locating target 2 is... and .

[0288] Table 6 shows the top five positioning task allocation schemes for target 3, obtained by the enumeration method.

[0289]

[0290] As can be seen from Table 6, due to and The number of times it has been used has exceeded the maximum number of targets that can be located per formation. Therefore, the optimal solution for locating target 3 is... and .

[0291] Table 7 shows the top five positioning task allocation schemes for target 4, obtained through enumeration.

[0292]

[0293] As can be seen from Table 7, due to and The number of times it has been used has exceeded the maximum number of targets that can be located per formation. Therefore, the optimal solution for locating target 4 is... and .

[0294] Table 8 shows the top five positioning task allocation schemes for target 5, obtained using the enumeration method.

[0295]

[0296] As can be seen from Table 8, due to and The number of times it has been used has exceeded the maximum number of targets that can be located per formation. ,so and Having been removed from the allocation sequence, the optimal solution for locating target 5 is: and .

[0297] Table 9 shows the top five positioning task allocation schemes for target 6, obtained using the enumeration method.

[0298]

[0299] As can be seen from Table 9, due to and The number of times it has been used has exceeded the maximum number of targets that can be located per formation. ,so and Having been removed from the allocation sequence, the optimal solution for locating target 6 is: and .

[0300] In summary, the localization task allocation scheme obtained by the particle swarm optimization algorithm is exactly the same as the optimal solution obtained by the enumeration method, and the algorithm result is the optimal solution.

[0301] Fifth, design of interference task allocation algorithm

[0302] (1) Interference task allocation algorithm flow

[0303] The specific steps of the interference task allocation algorithm are as follows:

[0304] S5-1: Reads information about surviving friendly platforms from the enemy-friendly situation in the scenario simulation system, including the longitude, latitude, altitude, azimuth, etc. of the friendly platforms, as alternative platforms that can be used to interfere with targets.

[0305] S5-2: Identify the jamming target, which is the target of the attack in the current stage. Read the target information obtained from sensor cooperative detection and cooperative positioning, including the enemy target's longitude, latitude, altitude, and azimuth.

[0306] S5-3: First, select the target that is currently under attack and has not been assigned. Using the distance between the standby machine and the currently selected target as the objective function, use the particle swarm algorithm to prioritize the target closest to the machine for interference.

[0307] S5-4: For targets not allocated our resources in the distance-based allocation, we use the particle swarm optimization algorithm to select our resources, with the interference effectiveness when our machine interferes with both the already interfered target and the target to be interfered with as the objective function. Note the constraint: If the angle between our machine and this target and the previously allocated interference targets is less than the maximum interference angle, we can interfere with both machines simultaneously; otherwise, we cannot interfere with both machines simultaneously, and the interference effectiveness is set to 0.

[0308] S5-5: Determine whether all targets have been assigned a positioning scheme. If so, the algorithm ends; otherwise, return to step 4 to continue the assignment.

[0309] (2) Interference effectiveness calculation function

[0310] By calculating the interference power Maximum interference power of radar The ratio of the two values ​​is used to obtain the interference effectiveness.

[0311]

[0312] Interference power calculation method:

[0313]

[0314] in, For radar transmit pulse power, For the transmit antenna gain, The distance between my aircraft and the jammed target. For the radar cross-section of the target, For antenna gain, This represents the polarization mismatch loss (typically taken as 0.5). The distance between my aircraft and the target to be jammed. The equivalent power of the interference signal. In order for the jammer to observe the angle between the jammed target and the target to be jammed, this invention specifies Effective when the temperature is less than 3°.

[0315] (3) Example of interference task allocation algorithm

[0316] With the positions of all sensors and the aircraft speed remaining constant, the maximum number of targets each jammer can jam is [number missing]. Assuming the target location is the same as in 2.4.3, and that the radar transmit power of the target is 25kW, the transmit antenna gain is 30dB, and the radar cross-section is... 62 The radar beamwidth is 5°, the maximum angle between two targets being jammed is 3°, and the jamming antenna gain is... 30dB, polarization coefficient The suppression coefficient is 0.5. The value is 4, and the maximum interference power is 40kW.

[0317] The algorithm was run using MATLAB R2018b software, and the globally optimal solution, global_best, is shown in Table 10. Specifically, jammer 3 interferes with targets 1 and 6, jammer 1 interferes with targets 2 and 5, jammer 2 interferes with target 3, and jammer 4 interferes with target 4. Jammer 3 is closest to target 1 at 43.6005 km, jammer 1 is closest to target 2 at 51.6237 km, jammer 2 is closest to target 3 at 47.0133 km, and jammer 4 is closest to target 4 at 97 km. When target 5 and target 2 are both interfered with by jammer 1, the highest efficiency is 1.4140. When target 6 and target 1 are both interfered with by jammer 3, the highest efficiency is 0.0004.

[0318] Table 10 shows the allocation schemes obtained by the particle swarm optimization algorithm and their corresponding distances and performance.

[0319]

[0320] Table 11 shows the optimal allocation schemes obtained by the enumeration method and their corresponding distances or efficiencies.

[0321]

[0322] The optimal allocation scheme and corresponding distance or efficiency obtained by the enumeration method are shown in Table 11. As can be seen from the results of the enumeration method, the particle swarm optimization algorithm can obtain the optimal solution and the optimal disturbance allocation scheme.

[0323] Sixth, design of the strike mission allocation algorithm.

[0324] (1) Strike mission allocation algorithm flow

[0325] The target allocation criterion of the strike mission allocation model is to minimize the cost while meeting the damage requirements. The steps of the strike mission allocation algorithm are as follows:

[0326] S6-1: Read friendly platform information from the enemy-friendly situation in the scenario simulation system, including the friendly platform's longitude, latitude, altitude, speed, heading, and number of weapons.

[0327] S6-2: Read target information obtained from sensor-assisted detection and cooperative localization, including enemy target longitude, latitude, altitude, speed, and heading.

[0328] S6-3: A weighted threat index is calculated based on the angle threat index, speed threat index, altitude threat index, and distance threat index of the targets to be assigned to our platform. Targets are then ranked according to the sum of their threat indices across all our platforms, and assigned to platforms sequentially according to this ranking.

[0329] S6-4: Using the particle swarm optimization algorithm with the strike effectiveness of our aircraft against the target as the objective function, our resources are allocated to the target, and guidance is completed by the aircraft carrying out the strike.

[0330] S6-5: Determine whether the number of times the selected platform has been reused has exceeded the maximum number of targets. If there is no excess, input the information of the next target to continue the allocation; if there is excess, remove the platform from the allocation sequence and return to step 4 for reallocation.

[0331] S6-6: Determine whether all targets have been allocated our resources. If so, the algorithm ends; otherwise, return to step 4 to continue allocation.

[0332] (2) Strike effectiveness calculation function

[0333] Strike effectiveness is represented by strike advantage, including angular advantage index, speed advantage index, height advantage index, and distance advantage index.

[0334] It is a relative distance; The difference between the target altitude and the altitude of our aircraft; and These represent the velocity vectors of our aircraft and the target aircraft, respectively. This is the azimuth angle of my aircraft; The target aircraft's approach angle.

[0335] The advantage in striking is the weighted sum of the various advantage indices, that is:

[0336]

[0337] in As the weights, the fuzzy analytic hierarchy process is used to obtain... .

[0338] The angle advantage index is calculated as follows:

[0339]

[0340] The speed advantage index is calculated as follows:

[0341]

[0342] The height advantage index is calculated as follows:

[0343]

[0344] When calculating the range dominance index, the maximum detection range of our aircraft needs to be considered. Maximum target detection range Our missile's maximum range Maximum range of the target .

[0345]

[0346] , , The parameter can take the following four values:

[0347] Scenario 1:

[0348]

[0349] Scenario 2:

[0350]

[0351] Scenario 3:

[0352]

[0353] Scenario 4:

[0354]

[0355] In addition, in cases 1 and 2 In cases 3 and 4 .

[0356] The threat level of a target aircraft to our aircraft is the strike advantage of a target aircraft against our aircraft, and the calculation method is the same as the above formula.

[0357] (3) Example of a strike mission allocation algorithm

[0358] The target location is the same as in 2.4.3, and my aircraft's maximum detection range is [not specified]. The maximum detection range for enemy aircraft is 160km. The maximum range of the air-to-air missiles carried by our aircraft is 159.999 km. The maximum range of enemy aircraft carrying air-to-air missiles is 95km. It is 80km / h. My aircraft's speed is... shaft and The projected velocities on the axis are shown in Table 2. The enemy aircraft's speed is... shaft and The velocities of the projection on the axis are shown in Table 13.

[0359] Table 12 Projection of my aircraft's velocity onto the coordinate axes

[0360]

[0361] Table 13 Projection of enemy aircraft speeds onto the coordinate axes

[0362]

[0363] Using MATLAB R2018b software to run the algorithm, the strike plan and strike effectiveness obtained by the particle swarm optimization algorithm are shown in Table 14, namely, aircraft No. 7 strikes enemy aircraft No. 1 and No. 2, aircraft No. 8 strikes enemy aircraft No. 3 and No. 4, aircraft No. 9 strikes enemy aircraft No. 5, and aircraft No. 5 strikes enemy aircraft No. 6.

[0364] Table 14 shows the attack schemes derived from the particle swarm optimization algorithm.

[0365]

[0366] The ranking of our aircraft's strike advantage over enemy aircraft, obtained by enumeration, is shown in Table 15, and the ranked strike advantage values ​​are shown in Table 16.

[0367] Table 15 Ranking of our aircraft's advantages in striking enemy aircraft

[0368]

[0369] Table 16 shows the sorted superiority values ​​of our aircraft against enemy aircraft.

[0370]

[0371] As shown in Tables 15 and 16, the particle swarm optimization algorithm can obtain a suboptimal solution with a total effectiveness of 0.5705. The optimal solution obtained by the enumeration method is [8, 8, 7, 7, 9, 1] with a total effectiveness of 0.5741 and a relative error of 0.6271%.

[0372] Seventh, the construction of the kill chain

[0373] (1) Kill chain construction method

[0374] The optimal allocation schemes for each stage are combined sequentially to construct the kill chain. (Target) The kill chain can be represented as: ,in This indicates the aircraft designation for the cooperative detection mission, representing the optimal detection scheme derived from the particle swarm optimization algorithm. This represents the carrier aircraft number that performs the localization task, which is the optimal localization scheme derived by the particle swarm optimization algorithm. The identifier represents the carrier aircraft that performs the jamming mission, which is the optimal jamming scheme derived by the particle swarm optimization algorithm. The aircraft number representing the carrier mission is the optimal strike plan derived by the particle swarm optimization algorithm.

[0375] (2) Example of killing chain construction

[0376] By combining the allocation instances of the detection phase, the positioning phase, the jamming phase, and the strike phase, and combining them in the order of execution, the kill chain construction result is obtained.

[0377] Table 17 Examples of Kill Chain Construction

[0378]

[0379] The above embodiments are merely preferred embodiments of the present invention. Those skilled in the art can make several improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for constructing a kill chain based on a particle swarm algorithm, characterized in that, Includes the following steps: Step 1, Kill Chain Stage Division: The kill chain is divided into four stages: detection stage, location stage, jamming stage, and strike stage. The detection phase: The application of sensors to detect unknown areas is the foundation for the location, jamming, and strike phases; the sensors include radar sensors, ESM sensors, and photoelectric sensors; The positioning phase involves accurately and continuously locating the target based on the target information returned from the detection phase. The jamming phase: The jammer releases jamming on the target to cover the breakthrough formation's breakthrough; The strike phase: The target is struck based on its location; Step 2, Kill Chain Problem Modeling: For the detection phase, localization phase, jamming phase, and strike phase, separate detection task allocation models, localization task allocation models, jamming task allocation models, and strike task allocation models are established for each phase. Step 3, Kill Chain Construction Algorithm: The kill chain construction algorithm is a kill chain construction method based on particle swarm optimization, and the particle swarm optimization algorithm is a decimal discrete particle swarm optimization algorithm. Step 4: The decimal discrete particle swarm optimization algorithm is used to solve the detection task allocation model, the localization task allocation model, the jamming task allocation model, and the strike task allocation model to obtain the optimal allocation schemes for detection tasks, localization tasks, jamming tasks, and strike tasks, respectively. Step 5: Combine the optimal allocation schemes for detection, location, jamming, and strike missions in sequence to construct a kill chain.

2. The method of claim 1, wherein, In step 2, the method for constructing the detection task allocation model is as follows: The detection methods for the detection mission include ESM dual-machine detection, radar single-machine or radar dual-machine detection, and electro-optical single-machine or electro-optical dual-machine detection; The probe task allocation scheme is wherein, the number of the airplane for performing the ESM probe; the number of the airplane for performing the photoelectric probe; the number of the airplane for performing the radar probe; is a flag bit: x7 is 1 or 2; like If the value is 1, the detection task allocation scheme is ESM dual-machine detection and photoelectric detection; like If the value is 2, the detection task allocation scheme is ESM dual-machine detection and radar detection; The detection scheme is selected with the goal of maximizing detection performance, the objective function s of the detection task allocation model is: det s = ∑i∈V∑j∈V∑k∈V∑l∈V∑m∈V∑n∈V∑ ； Where: n' is the number of sensor types currently enabled; This enhances the dual-machine detection performance of the ESM. For photoelectric detection efficiency; for Radar detection effectiveness; ; ; ; The number of ESM sensors; This refers to the number of photoelectric sensors; The number of radars; like Then photoelectric single-machine detection is used; if This method uses single-unit radar detection.

3. The kill chain construction method based on particle swarm optimization algorithm according to claim 1, characterized in that, In step 2, the method for constructing the location task allocation model is as follows: Positioning methods include ESM dual-machine positioning, photoelectric dual-machine positioning, and radar single-machine or dual-machine positioning; The location task allocation scheme (Loc) is represented as an n×3 matrix: ； The nth row represents the resources allocated to our side for the nth target; and This indicates the machine number used to locate the target. That is, single-machine positioning; otherwise, dual-machine positioning; x 13 and x n3 These are all flag bits, with values ​​ranging from 1, 2, or 3. A value of 1 selects ESM positioning; a value of 2 selects photoelectric positioning; and a value of 3 selects radar positioning. 11 Indicates the platform number of the first platform performing the positioning task; x 12 This indicates the second platform number that performed the positioning task; The objective function of the positioning task allocation model is s loc : ； ； Where n is the target quantity; Improve the positioning efficiency of ESM dual-machine system; For photoelectric positioning efficiency; To improve radar positioning efficiency; ; ; ;like In this case, radar single-unit positioning is used; j represents the j-th target.

4. The kill chain construction method based on particle swarm optimization algorithm according to claim 1, characterized in that, In step 2, the method for constructing the interference task allocation model is as follows: In a scenario where our jammers are fewer than the number of targets, assume that each jammer can jam at most two targets, and each target is jammed by only one jammer. First, allocate our resources based on the distance to the target as the objective function, and interfere with the nearest target. For the remaining targets, allocate our resources based on the interference effectiveness when our aircraft interferes with both the already interfered target and the target to be interfered with. The interference task allocation scheme is represented as follows: ,in, This indicates the jammer number used to jam the nth target; Let m be the number of our jamming devices; and n be the number of targets. This represents the distance between platform i and target j; This represents the interference effectiveness of platform i relative to target j; This indicates that target j is being interfered with by our platform i; Objective function s of the task allocation model int Includes two cases: like : ； like : ； The objective function s int with the constraint that: ； in, This indicates the maximum number of radars that a single jammer can jam.

5. The kill chain construction method based on particle swarm optimization algorithm according to claim 1, characterized in that, In step 2, the method for constructing the strike mission allocation model is as follows: Assume that each platform can attack a maximum of two targets, each target is attacked by only one platform, and one missile is launched for each attack. The strike mission allocation scheme is represented as follows: ,in, This indicates the platform number used by our side to strike the nth target. The objective function of the strike mission allocation model for: ； Where m represents the number of our fighter jets; This represents the overall situational advantage value of platform i relative to target j; This indicates that target j was attacked by our platform i; Objective function of the strike mission allocation model That is, to achieve the highest combat effectiveness with the least attack cost; The objective function The constraints are: ； in, This represents the number of platforms required to destroy target j; The maximum number of targets that can be attacked for each platform.

6. The method for constructing a kill chain based on particle swarm optimization as described in claim 1, characterized in that, In step 3, the process of improving the discrete particle swarm algorithm to a decimal discrete particle swarm algorithm is as follows: Aircraft are numbered using natural numbers, and the position vector of each particle represents the aircraft number performing a specific mission. The particle velocity update formula for the decimal discrete particle swarm optimization algorithm is the same as that for the binary discrete particle swarm optimization algorithm, and the particle position... The updated formula is as follows: The round() function rounds the integers according to the rounding rules. yes Updated particle positions, It is the current particle position; It is a sigmoid function mapping of particle update velocity; ； Particle position The range of values ​​for is as follows: ； The position of the j-th dimension particle The maximum value of is the maximum number of our platforms executing each stage of the task; the particle position in the j-th dimension represents the task allocation scheme for the j-th target; e is a constant; v ij This indicates the update speed of the position.

7. The method for constructing a kill chain based on particle swarm optimization as described in claim 1, characterized in that, In step 4, the process of obtaining the optimal allocation scheme for the detection tasks is as follows: Step 4.1.1: Input the mission area coordinates, available platform coordinates, sensor type, detection radius, and platform velocity parameters into the detection mission allocation model, and solve it using the particle swarm optimization algorithm; Step 4.1.2: Generate random particle position vectors, which are feasible solutions for the detection task allocation scheme; Step 4.1.3, calculate Corresponding ESM detection performance Find Corresponding photoelectric detection efficiency Find Corresponding radar detection effectiveness ; Compare and :like ≥ , Set to 1; if < , Set to 2; The fitness value of a particle is calculated, which represents the effectiveness of the current detection scheme. Step 4.1.4 compares the individual historical optimal value of a particle with the global optimal value of the swarm, causing the particle to continuously update its position and move closer to the optimal solution. Step 4.1.5: When the maximum number of iterations is reached, the algorithm terminates, and the optimal allocation scheme for the probe tasks is obtained. Otherwise, return to step 4.1.

3.

8. The method for constructing a kill chain based on particle swarm optimization as described in claim 1, characterized in that, In step 4, the process of obtaining the optimal allocation scheme for the location task is as follows: The optimal allocation scheme for the localization task is obtained by using the particle swarm optimization algorithm; Step 4.2.1: Read the friendly platform information from the battlefield situation; the friendly platform information includes the number of friendly platforms, longitude, latitude, and altitude; Step 4.2.2: Read the target information for detection; the target information includes the number of targets, the approximate target accuracy, latitude, and altitude; Step 4.2.3, input the target information; Using the particle swarm optimization algorithm, with ESM positioning performance as the objective function, the optimal ESM dual-machine positioning scheme and positioning performance are derived. The optimal photoelectric positioning scheme and positioning performance are derived by taking photoelectric positioning efficiency as the objective function. The optimal radar positioning scheme and positioning effectiveness are derived by using radar positioning effectiveness as the objective function. ; Compare , and Select the optimal allocation scheme for the positioning task and place the flag position on the corresponding number; if the sensor allocation sequence is empty, the positioning efficiency is 0 and the positioning allocation scheme is empty. Step 4.2.4: Determine whether the number of times the selected platform has been reused has exceeded the maximum number of location targets. If the number of targets is not exceeded, input the information for the next target and continue the allocation; if the number of targets is exceeded, remove the queue from the allocation sequence and return to step 4.2.3 for reallocation. Step 4.2.5: Determine whether all targets have been assigned a positioning scheme: if yes, the algorithm ends; otherwise, return to step 4.2.3 to continue the assignment.

9. The method for constructing a kill chain based on particle swarm optimization as described in claim 1, characterized in that, In step 4, the process of obtaining the optimal allocation scheme for the interference tasks is as follows: Step 4.3.1: Read the information of surviving friendly platforms from the enemy-friendly situation in the scenario simulation system, including the longitude, latitude, altitude and azimuth of the friendly platforms, as alternative platforms for interference targets; Step 4.3.2: Identify the interference target. The interference target is the target being attacked in the current stage. Read the target information obtained from sensor cooperative detection and cooperative localization, including the enemy target's longitude, latitude, altitude, and azimuth. Step 4.3.3: First, select the target that is currently under attack and has not been assigned. Use the distance between our machine and the currently selected target as the objective function, and use the particle swarm algorithm to prioritize the target that is closest to our machine for interference. Step 4.3.4: For targets that have not been allocated our resources in the allocation with distance as the objective function, the particle swarm optimization algorithm is used to select our resources with the interference effectiveness when our machine interferes with both the already interfered targets and the targets to be interfered with as the objective function. Step 4.3.5: Determine if all targets have been assigned a positioning scheme: if yes, the algorithm ends; otherwise, return to step 4.3.4 to continue the assignment. The process for obtaining the optimal allocation scheme for the strike missions is as follows: The target allocation criterion of the strike mission allocation model is to minimize the cost while meeting the damage requirements. The steps of the strike mission allocation algorithm are as follows: Step 4.4.1: Read the friendly platform information from the enemy and friendly situation in the scenario simulation system, including the friendly platform's longitude, latitude, altitude, speed, heading, and number of weapons; Step 4.4.2: Read the target information obtained from sensor cooperative detection and cooperative localization, including the enemy target's longitude, latitude, altitude, speed, and heading; Step 4.4.3: Calculate the threat index by weighting the angle threat index, speed threat index, altitude threat index, and distance threat index of the targets to be assigned to our platform; sort the targets according to the sum of their threat indices to all our platforms, and assign them in the order of the sorted targets. Step 4.4.4: Using the particle swarm optimization algorithm, with the strike effectiveness of our aircraft against the target as the objective function, we allocate our resources to the target, and guidance is completed by the aircraft carrying out the strike. Step 4.4.5: Determine whether the number of times the selected platform can be reused has exceeded the maximum number of targets to be hit. If it has not exceeded the maximum number of targets, enter the information of the next target and continue to allocate targets. If the number of requests exceeds the limit, the platform will be removed from the allocation sequence, and the process will return to step 4.4.4 for reallocation. Step 4.4.6: Determine whether all targets have been allocated our resources: if yes, the algorithm ends; otherwise, return to step 4.4.4 to continue allocation.

10. The method for constructing a kill chain based on particle swarm optimization as described in claim 1, characterized in that, In step 5, the kill chain targets the target. The kill chain is represented as: ； Among them, targeting The kill chain; The carrier number represents the aircraft that will carry out the collaborative detection mission, which is the optimal detection scheme derived by the particle swarm algorithm. The carrier number represents the optimal positioning scheme obtained by the particle swarm algorithm. The carrier number representing the aircraft performing the jamming task is the optimal jamming scheme derived by the particle swarm algorithm. The aircraft designation represents the carrier aircraft tasked with the strike mission, and is the optimal strike plan derived from the particle swarm optimization algorithm.

Images