A scheduling method for ant colony path rejudgment after particle swarm optimization

Abstract

The present invention relates to a scheduling method of ant colony path re-judgment after particle swarm optimization. In view of the current situation that the passive wideband phased array radar system has wide frequency domain and airspace coverage, large resource consumption, and great difficulty in resource scheduling, and the technical requirements for full airspace tracking of multiple forms and multiple targets on land, sea and air, the present invention is based on the fast solution characteristics of the improved particle swarm optimization algorithm, takes multi-target priority judgment as the goal, considers scheduling efficiency and full-cycle task completion rate under multi-target conditions, improves the particle swarm optimization algorithm, and combines the ant colony algorithm to optimize the multi-target full airspace scheduling process to obtain accurate scheduling results. The present invention utilizes the improved particle swarm optimization algorithm in combination with the ant colony algorithm, and is applied to the scenario when the passive wideband phased array radar system faces a multi-target full airspace search, which overcomes the shortcoming of low flexibility of the fixed template scheduling algorithm in conventional scheduling, and improves the resource utilization rate of the passive wideband phased array radar system. The invention has good versatility and practicality.

Description

Technical Field

[0001] The invention belongs to the technical field of radars and relates to a multi-objective task scheduling method for a broadband phased array system. Background Art

[0002] In the current technological context, broadband phased array systems have excellent performance in wide-area coverage, and can achieve full-airspace tracking of multiple targets through reasonable resource scheduling. For multiple targets and multiple tasks that need to be performed, corresponding efficient scheduling methods need to be designed to achieve the best results. However, according to in-depth research at home and abroad in recent years, it is impossible to perfectly perform all tasks within a limited time under limited conditions. The current solution is mainly to give priority to tasks that meet the conditions, which is the traditional method of arranging tasks according to task priority.

[0003] In "Phased Array Radar Resource Scheduling: Task Scheduling and Performance Evaluation", the authors use task modes and deadlines to calculate the overall priority of tasks and give priority to tasks with higher overall priority; "Multi-beam Adaptive Simulation Scheduling Algorithm for Phased Array Radar" further considers the timeliness of task execution and proposes a benefit-based scheduling strategy; "Time Window-Based Scheduling Method for Multi-Function Radar" gives a strategy for resource scheduling using a multi-nested heuristic algorithm. The calculation is simple, but there is a problem that the optimal solution cannot be found.

[0004] The above strategies can achieve reasonable scheduling of multiple tasks based on task priority to a certain extent, but there are two problems. First, it is often impossible to find the optimal solution due to conditional constraints; second, when the task scale is large, the deviation value of the solution is too large. Summary of the invention

[0005] In order to solve the problem that the traditional priority scheduling strategy in the existing phased array radar system cannot adapt to the tracking of a large number of multiple targets and the deviation in finding the optimal solution is large, the present invention proposes a scheduling method for ant colony path re-judgment after particle swarm optimization, and modifies the weight coefficient in the particle swarm algorithm by adding a priority judgment factor, thereby affecting the iterative formula to form an improved particle swarm algorithm. After the particle swarm algorithm is iterated multiple times to obtain a local solution, it is combined with the ant colony algorithm to find the shortest path to modify the obtained local solution. After fusion calculation, the local optimal solution based on the target priority and the target interception probability is obtained as a scheduling strategy for tracking multiple targets in the entire airspace.

[0006] In order to solve the above technical problems, the present invention is implemented through the following technical solutions:

[0007] Step 1: Determine the priorities of different goals: Design the goal priorities based on the importance, urgency and effectiveness of different goals faced by the system. The task priority of the i-th goal ω i for:

[0008]

[0009] where α i is the importance of the i-th target, β i is the urgency of the i-th goal, γ i is the effectiveness of the i-th objective, α max is the maximum value of all target importances, β max The maximum value of all target urgency;

[0010] Step 2: According to the priority of multiple objectives, the modified particle swarm algorithm dynamic priority weight coefficient w is obtained:

[0011]

[0012] in It is the ratio of the current tracking target task priority to the highest target priority;

[0013] Step 3: According to the logic of the particle swarm algorithm, the local solution of the scheduling task is regarded as a particle. In the tth iteration, particle i continuously updates its position and speed by tracking the individual extreme value pbest(t) and the global extreme value gbest(t) to find the local optimal solution. The formula is:

[0014] ν i (t+1)=w×ν i (t)+c 1 × 1 ×[p best (t)-x i (t)]+c 2 × 2 ×[g best (t)-x i (t)]

[0015] x i (t+1)=x i (t)+ν i (t)

[0016] Where c 1 and c 2 is a learning factor, which is used to adjust the maximum step length of the individual extreme value and global extreme value of the particle during flight; r 1 and r 2 is a random number between (0,1);

[0017] Step 4: After obtaining the local solution, adjust the local solution according to the feedback from the ant colony: Considering that the pheromone concentration is inversely proportional to the path length, the tasks with high pheromone concentration have high detection and interception probability, and the task priority is also higher; adjust the content and order of the tasks to be sent through the updated pheromone concentration table to complete the optimal design of the local solution.

[0018] Furthermore, the ant colony feedback in step 4 is based on the target capture probability as the basis for selecting the shortest path.

[0019] The present invention utilizes an optimized particle swarm algorithm combined with an ant colony algorithm to recalculate the obtained local solution, and rearranges tasks according to the final local optimal solution. The present invention can improve the resource utilization rate and task scheduling success rate of the radar system, and has good versatility and practicality.

[0020] The technical solution of the present invention is further explained below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The schematic diagram is a preferred implementation process of the present invention. DETAILED DESCRIPTION

[0022] The scheduling method of ant colony path re-judgment after particle swarm optimization proposed in the present invention improves particle encoding by increasing the priority of the current tracking target task as the dynamic priority weight coefficient of the particle swarm algorithm; uses the improved particle swarm algorithm to iteratively calculate the local solution of the multi-target task; and then uses the ant colony algorithm to feedback and adjust the obtained local solution to form a local optimal solution, thereby optimizing the multi-target full-airspace tracking scheduling of the broadband phased array system. The present invention redesigns the dynamic weights in the particle swarm algorithm according to the principles of importance, urgency and effectiveness of different targets, adds the ant colony algorithm to optimize the solution after the particle swarm algorithm iteration, and realizes the local optimal solution under the requirements of multi-target full-airspace tracking.

[0023] The preferred implementation flow chart of the present invention is as follows Figure 1 As shown, the implementation steps include:

[0024] (1) Determine the priority of each target according to its characteristics, thus forming an improved particle swarm algorithm

[0025] In the particle swarm algorithm, the size of the inertia weight determines the global and local optimization capabilities of the algorithm. A larger inertia weight can give the algorithm a stronger global search capability, and a smaller inertia weight gives the algorithm better local optimization capabilities. Therefore, when designing the inertia weight, the task priorities of different targets can be taken into consideration in the weight coefficient. When the priority of a certain target is higher, increase the weight coefficient of the task so that this task has better performance in the global search, otherwise it focuses on searching for a better local solution. When designing task priorities, the main things to consider are the importance, urgency, and effectiveness of different targets. Therefore, the design is as follows: The task priority ω of the i-th target i for:

[0026]

[0027] where α i is the importance of the i-th target, β i is the urgency of the i-th goal, γ i is the effectiveness of the i-th objective, α max is the maximum value of all target importances, β max The maximum value of all target urgency;

[0028] According to the priority of multiple targets, the modified particle swarm algorithm dynamic priority weight coefficient w is obtained

[0029]

[0030] in is the ratio of the current tracking target task priority to the highest target priority, w max is the maximum value of the weight coefficient in each iteration, w min is the minimum value of the weight coefficient in each iteration, t max is the maximum number of iterations.

[0031] (2) Using the improved particle swarm algorithm to iteratively calculate the local solution of multi-objective scheduling

[0032] The particle swarm algorithm first needs to initialize the particle swarm population according to different target tasks, and encode the particles according to different tasks. The dimension of the particle is the number of tasks, and each particle represents the scheduling of a target task. If there are n tasks, m processing nodes, and the particle population size is N, then the encoding method of the i-th particle can be expressed by the following vector: i ={p i1 ,p i2 ,…,p ij ,…,p in}, where P i is the encoding method of the i-th particle, p ijrepresents the processing node to which the jth task is scheduled in the i-th scheduling scheme, n is the number of tasks, 1≤i≤N, N is the size of the particle population; at initialization, p ij The value range is 0≤p ij ≤m. m is the number of processing nodes.

[0033] During the tth iteration, particle i tracks the individual extreme value p best (t) and the global extreme value g best (t) continuously updates its position and speed to find the local optimal solution. The formula is:

[0034] ν i (t+1)=w×ν i (t)+c 1 × 1 ×[p best (t)-x i (t)]+c 2 × 2 ×[g best (t)-x i (t)]

[0035] x i (t+1)=x i (t)+ν i (t)

[0036] Where c 1 and c 2 is a learning factor, which is used to adjust the maximum step length of the individual extreme value and global extreme value flight of the particle during the flight process. It is preferred to set c 1 =c 2 =2; r 1 and r 2 is a random number between (0,1);

[0037] (3) Scheduling strategy to form local optimal solution using ant colony feedback

[0038] The principle of the ant colony algorithm is: 1. Ants release pheromones on the path; 2. A path is randomly selected at the intersection, and pheromones related to the length of the path are released when passing through the path; 3. The pheromone concentration is inversely proportional to the length of the path. When the ants encounter the intersection later, they choose the path with a high pheromone concentration; 4. The pheromone concentration of the optimal path becomes larger and larger, and the advantages accumulate; 5. Finally, the optimal path is found.

[0039] According to the target interception probability as the basis for the shortest path selection, in the ant colony algorithm formula, the information heuristic factor is selected as 1, the expectation heuristic factor is selected as 4, the number of ants is 100, the information volatility factor is 0.8, and the residual factor is 0.2. Set the initial information amount τ ij(0) = τ max , in the information calculation formula τ ij (t+n)=(1-ρ)·τ ij (t)+Δτ ij (t), Combining this with the basic formula of the ant colony system (ACS), after one cycle, the pheromone release that finds the shortest path is:

[0040] L=min(L k ),k=1,2,…,m. Among them τ ij (t) is limited to [τ min ,τ max ] According to the pheromone concentrations under each path formed after one cycle, the paths are arranged from high to low, and a ranking table of paths from good to bad can be obtained. According to this table, the tasks of the sending waiting area are rescheduled, and the residence time of the tasks under the paths with high pheromone concentrations is increased to increase the probability of interception, realize a virtuous cycle of positive feedback, and give priority to the key frequency domain airspace with high pheromone concentrations to realize the priority sorting and scheduling of the secondary tasks. Each subsequent task scheduling will update the pheromone concentration table, and reschedule the tasks according to the latest concentration table to achieve tracking of multiple targets in the entire airspace. In this way, the local optimal solution, that is, the determination of the optimal path, is determined, and ultimately the resource utilization rate of the radar system and the improvement of the success rate of task scheduling are achieved.

Claims

1. A scheduling method for re-judging ant colony paths after particle swarm optimization, characterized by: Step 1: Determine the priorities of different goals: Design the goal priorities based on the importance, urgency and effectiveness of different goals; the task priority of the i-th goal ω i for: where α i is the importance of the i-th target, β i is the urgency of the i-th goal, γ i is the effectiveness of the i-th objective, α max is the maximum value of all target importances, β max The maximum value of all target urgency; Step 2: According to the priority of multiple objectives, the modified particle swarm algorithm dynamic priority weight coefficient w is obtained: in It is the ratio of the current tracking target task priority to the highest target priority; Step 3: The local solution of the scheduling task is regarded as a particle. In the tth iteration process, particle i tracks the individual extreme value p best (t) and the global extreme value g best (t) Continuously update its position and speed to find the local optimal solution, including: ν i (t+1)=w×ν i (t)+c1×r1×[p best (t)-x i (t)]+c2×r2×[g best (t)-x i (t)] x i (t+1)=x i (t)+ν i (t) Where c1 and c2 are learning factors, which adjust the maximum step length of individual extreme value and global extreme value flight of particles during flight; r1 and r2 are random numbers between (0,1); Step 4: After obtaining the local solution, adjust the local solution according to the feedback from the ant colony: tasks with high pheromone concentrations have high detection and interception probabilities, and their priorities are also higher; adjust the content and order of tasks to be sent through the updated pheromone concentration table to complete the optimal design of the local solution.

2. The scheduling method of ant colony path re-judgment after particle swarm optimization according to claim 1, characterized in that: The ant colony feedback in step 4 is based on the target capture probability as the basis for selecting the shortest path.

Images