Weapon-target Assignment Problem Solving Based On Particle Swarm Optimization Algorithm

Weapon-Target Assignment problem is a complex optimization problem in multi-platform antiaircraft system. It aims at reasonably assigning defender's weapons to the enemy-targets so that the total expected survival probability of the targets after all the weapon engagements is minimized and the defender's resources and weapon assets can be well protected. This problem sometimes called fire unit assignment or target assignment.The main objective of the research of Weapon-Target Assignment is to make the defender's command and control system can assign weapon unit to the enemy-targets effectively and efficiently so as to eliminate the threatens of the enemy-targets, and to minimize the total loss of the defenders resources and weapon assets. The Weapon-Target Assignment problem is a NP hard problem, and it is characterized by discrete, dynamic, non-linear, stochastic, and huge solution space. With the increase of the number of weapons and number of enemy-targets, the solution space may increase rapidly. It is unrealistic to obtain the optimum solution of the problem with exact methods if the scale of the problem is large. Thus, for the Weapon-Target Assignment problem, especially the large scale problem, developing efficient methods to obtain a optimal weapon-target assignment scheme, thus to improve the efficiency, is still an important subject in the command and control. It also has a great theoretical and practice meaning.Particle swarm optimization is a new developed swarm intelligence based computation method. The basic idea of particle swarm optimization stems from artificial life and evolutionary computation theory. When people investigating the forgiving behavior of swarm of birds, they found each bird only trace the track of its limited number of neighbors, but the final results shows as if the whole swarm is under the control of a central controller. The particle swarm optimization emulates the following scenery: a flock of birds are searching food randomly. There is only one piece of food in this region. And no bird knows where the food is, but they know how far it is away from the food. The most simple and effect strategy is to search the regions of the bird whose position is most close to the food.The traditional particle swarm optimization searches in the real field, but the weapon-target assignment is a discrete combinational optimization problem. Thus, the traditional particle swarm optimization cannot be applied to weapon-target assignment problem directly. By modifying the traditional particle swarm optimization, this paper proposed a new particle swarm optimization algorithm for weapon-target assignment problem. The improvements due to the following aspects:Firstly, this paper designs a particle decoding scheme for weapon-target assignment problem. This scheme can encode particle easily. Meanwhile, this scheme can decode a particle easily and can generate a weapon assignment scheme directly. At the same, this encoding and decoding scheme can guarantee all the problem constraints except constraint two.Secondly, according to the characteristics of the weapon-target assignment problem, this paper designs an objective function to evaluate the performance of the assignment scheme of each particle.Thirdly, because the proposed decoding scheme cannot satisfy all the constraints of the weapon-target assignment problem, this paper designs a new particle initialization method. This method can guarantee the particles in the population satisfy all the constraints in the weapon-target assignment problem.Fourthly, the characteristics of the weapon-target assignment problem make the traditional particle swarm optimization cannot be applied to it directly. Thus, this paper firstly modifies the concept of"velocity"in traditional particle swarm optimization, and defines the"velocity"as"times"that it adjusts to its previous best or global best.Fifthly, in order to calculate the"velocity"of each particle, this paper defines the concept of"distance"between each two particles. This paper defines the"distance"as the weighted sum of the particle difference of the two particle and the objective function values. Based on the"distance"of the particles, this paper proposes a new velocity update method. Finally, in order to update the particles'position according to the particles'velocities, this paper designs an"adjust"operation for each particle adjust its position to its previous best and global best.To show the effectiveness of the proposed algorithm, this paper performs three groups of simulations. The results show that the proposed algorithm is an effectiveness algorithm for Weapon-Target Assignment.