Coordinated firepower decision-making method for static UAV swarm combat based on improved particle swarm optimization algorithm

Abstract

The present invention discloses a method for coordinated firepower decision-making in static UAV swarm combat based on an improved particle swarm algorithm. The method comprises the following steps: first, establishing a static weapon target allocation model for a UAV swarm; second, encoding the weapon target allocation scheme; then initializing a particle swarm optimization algorithm; and finally, performing a particle swarm optimization optimization search algorithm and decoding the obtained feasible solution to obtain a possible allocation scheme. The method for coordinated firepower decision-making in static UAV swarm combat based on an improved particle swarm algorithm provided by the present invention has a strong global search capability, can effectively avoid falling into local optimality, can quickly search for a better solution, and can further improve the quality of the solution, effectively solving the problem of excessively long optimization time in existing methods and algorithms.

Description

Technical Field

[0001] The present invention relates to the field of weapon target allocation, and in particular to a method for cooperative firepower decision-making in static UAV cluster combat based on an improved particle swarm algorithm. Background Art

[0002] Coordinated firepower decision-making methods for drone swarm combat are a key component of modern combat command. By studying how to rationally allocate our weapons to existing team members, we can achieve optimal damage effects from our drone weapons on incoming targets. Through firepower decision-making, we minimize our economic losses and protect our supplies. Based on the number of drone weapons and ammunition, the number of incoming targets, the threat level of the incoming targets, the probability of damage from the weapons, and other conditions, we select weapons and ammunition suitable for striking these targets and optimize their allocation. This means comprehensively considering the total amount of various types of weapons and ammunition, and rationally allocating appropriate weapons and ammunition to the targets in the best way possible, ensuring that all targets achieve the desired damage effect.

[0003] When the WTA problem is large in scale, it still cannot effectively solve the problem of significantly prolonged optimization. The objective function of the WTA problem is complex, and the search space grows exponentially with the number of targets, weapon types, and the number of weapons of each type. The space is enormous, and the WTA problem is NP-complete, making it difficult to determine the location of its local optimal region through arithmetic operations. Given the diversity of targets and the large number of weapon types available, existing algorithms are prone to falling into local optimal solutions and consume excessive time, making it difficult to obtain a truly optimal solution.

[0004] Currently, there are many existing optimization algorithms for the static WTA problem in drone swarms, including particle swarm optimization, ant colony optimization, and genetic algorithms. While most of these algorithms can achieve satisfactory solutions, they all suffer from premature convergence and local optima to varying degrees. The primary cause of this problem is that the algorithms often become trapped in local optima during the optimization process, lacking the means to quickly escape from these local optima. Especially when the particle size is large, existing algorithms are prone to falling into local optima or require numerous iterations to reach the optimal solution, making it difficult to quickly converge on the correct allocation solution. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for coordinated firepower decision-making in static UAV cluster combat based on an improved particle swarm optimization algorithm. By encoding particles, improving the value selection method of the particle swarm optimization algorithm weight coefficient, improving the method of updating particle velocity, and updating the method of particle position, the speed of jumping out of the local optimal solution in the solution process is accelerated.

[0006] The method solution for implementing the present invention is as follows: In a first aspect, the present invention provides a method for cooperative firepower decision-making in static UAV cluster combat based on an improved particle swarm algorithm, comprising the following steps:

[0007] Step 1: Establish a static weapon target allocation model for UAV swarm system combat based on the number of UAV swarms and the number of incoming target types, and obtain the corresponding fitness function based on the constraints;

[0008] Step 2: Encode the particles and initialize the particle position and velocity;

[0009] Step 3: Use the improved particle swarm algorithm to perform marching optimization and continuously update the search speed and position of particles in each dimensional direction until the iteration termination condition is reached.

[0010] In a second aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the first aspect when executing the program.

[0011] Compared with the existing methods, the present invention has the following significant advantages:

[0012] (1) The present invention constructs a mathematical model for the coordinated firepower decision-making method of static UAV swarm combat based on the improved particle swarm algorithm, ensuring that the threat level of each target and the probability of damage to each target by the weapon are calculated, meeting the concept of consistency of the weapon target allocation system, thereby maximizing the benefits of the overall allocation system;

[0013] (2) This invention uses a particle swarm optimization algorithm, which is improved on the original one. It effectively avoids the algorithm from falling into the local optimum during the optimization process, making the difference between the longest and shortest optimization times smaller, and can effectively solve the problem of excessive optimization time. It improves the efficiency of UAV clusters in firepower decision-making and subsequent tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of the collaborative firepower decision-making method for static UAV cluster combat based on the improved particle swarm algorithm of the present invention.

[0015] Figure 2 Schematic diagram of the particle encoding scheme in the present invention.

[0016] Figure 3 It is a convergence comparison diagram in a specific example of the present invention. DETAILED DESCRIPTION

[0017] like Figure 1 As shown in FIG, a method for cooperative firepower decision-making in static UAV cluster combat based on an improved particle swarm algorithm includes the following steps:

[0018] Step 1: Establish a static weapon target allocation model for UAV swarm system combat based on the number of UAV swarms and the number of incoming target types, and obtain the corresponding fitness function based on the constraints;

[0019] Step 2: Encode the particles and initialize the particle position and velocity;

[0020] Step 3: Use the improved particle swarm algorithm to perform marching optimization and continuously update the search speed and position of particles in each dimension until the iteration termination condition is reached;

[0021] Furthermore, the establishment of the drone swarm weapon target allocation model described in step 1 is as follows:

[0022] The threat level of the incoming target is analyzed, and a model f1 with the maximum damage coefficient caused by drone swarm combat to the target is established. The function is expressed as:

[0023]

[0024] Among them, v ij Indicates the threat level of the incoming target to our drone, x ij = 0 means that our i-th UAV attacks the j-th target without allocating the weapons carried by the UAV; otherwise, x ij =1 means allocating weapons to attack and destroy, p ij Indicates the probability of the drone damaging the incoming target;

[0025] The constraints on the model are as follows:

[0026] Constraint 1: The maximum number of weapons that the i-th UAV platform can use;

[0027]

[0028] Among them, r i represents the upper limit of the number of weapons that the i-th UAV platform can use;

[0029] Constraint 2: The minimum weapon that the i-th UAV platform can use;

[0030]

[0031] Constraint 3: The maximum number of drone weapons required for each incoming target;

[0032]

[0033] Among them, s j It represents the upper limit of the number of drone weapons that can be assigned to each incoming target;

[0034] Constraint 4: The maximum number of drone weapons required for each incoming target;

[0035]

[0036] Constraint 5: When allocating drone weapons for combat, the actual number allocated cannot exceed the total number of drone swarm weapons;

[0037]

[0038] Furthermore, the fitness function expression in the improved particle swarm algorithm is as follows:

[0039]

[0040] Among them, minF(x ij ,δ k ) represents the minimum value of the fitness function, x ij Indicates whether drone weapons are assigned, δ k is the penalty factor, and δ k >0; α=β=χ=η=μ=σ=2.

[0041] Furthermore, the particle coding scheme described in step 2 is based on assigning incoming targets to the weapons carried by the arranged drone cluster in the form of integer codes, thereby performing coding processing, and the length D of the code is the sum of all weapons in the drone cluster.

[0042] Furthermore, the method for updating the search speed and position of the particle in each dimensional direction in step 3 is:

[0043] v id (t+1)=ωv id (t)+c1r1(p id -x id (t))+c2r2(p gd -x id (t))

[0044] x id (t+1)=x id (t)+v id (t+1)

[0045] Where, v id (t+1) represents the speed of the particle moving at the t+1th time, x id(t+1) is the position of the particle at the t+1th iteration; ω is the inertia factor of the particle movement, and its value is non-negative. When the inertia factor value is relatively large, the particle has a strong ability to find the optimal solution globally. When the inertia factor value is small, the particle has a weak ability to find the optimal solution globally. ω can be adjusted to achieve local or global optimal search. id is the optimal solution that the particle can find at the tth iteration, and p best Indicates that p gd is the optimal solution found in the entire particle swarm, using g best represents. c1 and c2 are acceleration factors. r1 and r2 are random numbers in [0,1].

[0046] Furthermore, the inertia factor w of particle movement is improved, and a linearly decreasing inertia weight formula is established, as follows:

[0047]

[0048] Among them, maxG and curG represent the maximum number of iterations of the current algorithm and the number of iterations currently running, respectively, and are generally taken as w max =0.9,w min =0.2.

[0049] Furthermore, based on the definition of the “distance” between particles, the search speed of particles in each dimension is redefined:

[0050]

[0051] Among them, S(x i ,x j ) are two particles x i and x j The similarity function of .

[0052] dis(p i -x i )=h[ρ|f(p i )-f(x i )|] / C+υ(DS(P i ,x i )) / D]

[0053] Among them, ρ and υ are two positive numbers, and ρ+υ=1, which are used to adjust the difference in function fitness between two particles and the encoding difference between two particles, respectively. i ) and f(x i ) are all fitness values; h is the acceleration factor, which is a positive integer; C is the maximum fitness function value in the current group; and D is the particle dimension.

[0054] v i =int[ωv i +c1r1dis(p i -x i )+c2r2dis(p g -x i )]

[0055] Among them, int[·] means taking the integer part of this formula, p i and p g are the optimal solutions p of the individual extreme values of the particles best and the global optimal solution g best The other parameter variables are the same as those of the PSO algorithm, and their overall meanings are the same as those of the particle swarm algorithm.

[0056] Furthermore, the updated particle velocity can be synchronously applied to the updated particle position, and the particle position formula is updated as follows:

[0057]

[0058] Among them, MaxT and CurT represent the maximum number of iterations of the current algorithm and the number of iterations currently running, respectively.

[0059] Furthermore, the feasible solution of the improved particle swarm algorithm is decoded in an integer manner based on the attack targets and weapons to obtain the optimal allocation plan.

[0060] The following will describe in detail a method for coordinated firepower decision-making in static UAV cluster combat based on an improved particle swarm algorithm of the present invention in combination with the accompanying drawings and specific embodiments.

[0061] This paper applies a method for coordinated firepower decision-making in static UAV swarm combat, based on an improved particle swarm optimization algorithm, to the static UAV weapon-target assignment problem. Based on the same context, different cases were set up, each optimized using the method described in this paper. The resulting optimized models and algorithms were compared with those of existing methods.

[0062] The scenario involves 10 drones, each carrying one weapon, and eight incoming targets. A total of 10 weapons are used to strike eight targets. Table 1 shows the available quantity of each weapon type and the maximum number of weapons that can be used per target. Table 2 shows the threat level of each target to the drones. Table 3 shows the known probability of damage to the targets by each weapon type.

[0063] Table 1

[0064]

[0065] Table 2

[0066]

[0067] Table 3

[0068]

[0069] Using Matlab software, under the same hardware and software environment, the PSO algorithm before improvement and the improved PSO algorithm were used to program and simulate the WTA problem, and the average optimization time (unit: second), average number of iterations and average minimum value of the objective function (fitness function) of the conventional method and the method of the present invention were recorded.

[0070] Reference Figure 1 The flow chart of the optimization program is designed and written. First, the static UAV weapon target allocation model and constraints are determined according to the above conditions:

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077] Among them, r i It represents the upper limit of the number of weapons that the i-th UAV platform can use, s j It represents the upper limit number of drone weapons that can be assigned to each incoming target, and the function model satisfies the constraints.

[0078] Next, according to Figure 2 The feasible solution is integer-encoded, the encoding dimension is D, the number of individuals in the drone cluster is m, r1+r2+…+r m =D. A weapon can only strike a target once.

[0079] Initialize particle position, velocity, and set p best 、g best and the number of iterations.

[0080] Substitute the current position of each particle into the fitness function to calculate the fitness value of each particle. The calculation method is:

[0081]

[0082] Among them, δ k is the penalty factor, and δ k >0. Generally, α=β=χ=eta=μ=σ=2.

[0083] Then for each particle p i Evaluate and then update p best and g best .

[0084] When updating the particle velocity, use the definition of "distance" to recalculate the individual particle position p i and the individual optimal solution p best , global optimal solution g best The distance between them is calculated as:

[0085]

[0086] Among them, S(x i ,x j ) are two particles x i and x j The similarity function of .

[0087] dis(p i -x i )=k[α|f(p i )-f(x i )|] / C+β(DS(P i ,x i )) / D]

[0088] Among them, α and β are two positive numbers, and α+β=1, which are used to adjust the difference in function fitness between two particles and the encoding difference between two particles, respectively. i ) and f(x i ) are fitness values. k is the acceleration factor, which is a positive integer. C is the maximum fitness function value in the current group. D is the particle dimension.

[0089] The update of the search speed of the particle in each dimension is as follows:

[0090] v i =int[ωv i +c1r1dis(p i -x i )+c2r2dis(p g -x i )]

[0091] Among them, int[·] means taking the integer part of this formula, p i and p g are the optimal solutions p of the individual extreme values of the particlesbest and the global optimal solution g best The other parameter variables are the same as those of the PSO algorithm, and their overall meanings are the same as those of the particle swarm algorithm.

[0092] The update of the search position of the particle in each dimension is as follows:

[0093]

[0094] Among them, MaxT and CurT represent the maximum number of iterations of the current algorithm and the number of iterations currently running, respectively.

[0095] Continuously update the search speed and position of the particle in each dimension until the iteration termination condition is reached. When the termination condition is met, stop iterating and output g best , and obtain the corresponding decision value x through encoding and decoding ij , traverse the entire matrix record to update the decision matrix X, so as to complete the firepower decision of the drone cluster. Otherwise, return to the fitness function to recalculate and repeat the operation.

[0096] Table 4 accurately records the best allocation scheme obtained by using the improved particle swarm algorithm:

[0097] Table 4

[0098]

[0099] Table 5 accurately records the results of the average time, average number of iterations and average minimum value of the objective function of the two algorithms when the program is run 100 times.

[0100] Table 5

[0101]

[0102] The above comparative analysis demonstrates that the proposed method for coordinated firepower decision-making in static UAV swarm combat, based on an improved particle swarm optimization algorithm, effectively addresses the time-consuming optimization problem encountered in existing algorithms. It also improves solution quality, achieves better convergence, and is less likely to fall into local optima. It also provides a solution for rapid UAV swarm allocation, improving mission efficiency.

[0103] Figure 3 The results show that the improved particle swarm algorithm converges faster than the general particle swarm algorithm using the method of the present invention, and the number of iterations of the improved algorithm is smaller than that before the improvement. The improved algorithm is easier to converge and has better results.

[0104] The above content describes in detail the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm optimization algorithm, characterized by: The steps include: Step 1: Establish a static weapon target allocation model for UAV swarm system combat based on the number of UAV swarms and the number of incoming target types, and obtain the corresponding fitness function based on the constraints; The establishment of the static weapon target allocation model for UAV swarm system combat is as follows: The threat level of the incoming target is analyzed, and a model f1 with the maximum damage coefficient caused by drone swarm combat to the target is established. The function is expressed as: Among them, v ij Indicates the threat level of the incoming target to our drone, x ij = 0 means that our i-th UAV attacks the j-th target without allocating the weapons carried by the UAV; otherwise, x ij =1 means allocating weapons to attack and destroy, p ij Indicates the probability of the drone damaging the incoming target; The constraints on the model are as follows: Constraint 1: The maximum number of weapons that the i-th UAV platform can use; Among them, r i represents the upper limit of the number of weapons that the i-th UAV platform can use; Constraint 2: The minimum weapon that the i-th UAV platform can use; Constraint 3: The maximum number of drone weapons required for each incoming target; Among them, s j It represents the upper limit of the number of drone weapons that can be assigned to each incoming target; Constraint 4: The maximum number of drone weapons required for each incoming target; Constraint 5: When allocating drone weapons for combat, the actual number allocated cannot exceed the total number of drone swarm weapons; The fitness function expression in the improved particle swarm optimization algorithm is as follows: Among them, min F(x ij ,δ k ) represents the minimum value of the fitness function, x ij Indicates whether drone weapons are assigned, δ k is the penalty factor, and δ k >0;α=β=χ=eta=μ=σ=2; Step 2: Encode the particles and initialize the particle position and velocity; Step 3: Use the improved particle swarm algorithm to perform marching optimization and continuously update the search speed and position of particles in each dimensional direction until the iteration termination condition is reached.

2. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm optimization algorithm according to claim 1 is characterized in that: The particle coding scheme in step 2 is based on assigning the incoming targets to the weapons carried by the arranged drone cluster in the form of integer codes, and then performing coding processing. The length of the code is the sum of all the weapons in the drone cluster.

3. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm optimization algorithm according to claim 1 is characterized in that: The method for updating the search speed and position of particles in each dimension in step 3 is: v id (t+1)=ωv id (t)+c1h1(p id -x id (t))+c2h2(p gd -x id (t)) x id (t+1)=x id (t)+v id (t+1) Where, v id (t+1) represents the speed of the particle moving at the t+1th time, x id (t+1) is the position of the particle at the t+1th iteration; ω is the inertia factor of the particle movement, its value is non-negative, and ω is adjusted to achieve local or global optimal search; p id is the optimal solution that the particle can find at the tth iteration, and p best Indicates that p gd is the optimal solution found in the entire particle swarm, using g best Indicates; c1 and c2 are acceleration factors; h1 and h2 are random numbers in the range [0,1].

4. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm algorithm according to claim 3 is characterized in that: Improve the inertia factor w of particle movement and establish a linearly decreasing inertia weight formula, as follows: Among them, maxG and curG represent the maximum number of iterations of the current algorithm and the number of iterations currently running, respectively, and the value is w max =0.9,w min =0.

2.

5. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm algorithm according to claim 3 is characterized in that: According to the definition of "distance" between particles, the search speed of particles in each dimension is redefined: Among them, S(x i ,x j ) are two particles x i and x j Similarity function of dis(p i -x i )=h[ρ|f(p i )-f(x i )|] / C+υ(D-S(p i ,x i )) / D] Among them, ρ and υ are two positive numbers, and ρ+υ=1, which are used to adjust the difference in function fitness between two particles and the encoding difference between two particles, respectively. i ) and f(x i ) are all fitness values; h is the acceleration factor, which is a positive integer; C is the maximum fitness function value in the current group; and D is the particle dimension; v i =int[ωv i +c1h1dis(p i -x i )+c2h2dis(p g -x i )] Among them, int[·] means taking the integer part of this formula, p i and p g are the optimal solutions p of the individual extreme values of the particles best and the global optimal solution g best ; The other parameter variables are the same as those of the PSO algorithm, and their overall meanings are the same as those of the particle swarm algorithm.

6. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm algorithm according to claim 5 is characterized in that: The updated particle velocity can be synchronously applied to the updated particle position, and the particle position formula is updated as follows: Among them, MaxT and CurT represent the maximum number of iterations of the current algorithm and the number of iterations currently running, respectively.

7. The method for cooperative firepower decision-making in static UAV swarm combat based on improved particle swarm algorithm according to claim 6 is characterized in that: The feasible solution of the optimization algorithm is decoded in an integer manner based on the attack targets and weapons to obtain the best allocation plan.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method according to any one of claims 1 to 7 are implemented.

Images