Multi-Stage Dynamic Target Assignment Method Based on Improved Discrete PSO Algorithm

With regard to the problem of uncertainty of interception results in the dynamic weapon target assignment in modern warfare, a multi-stage dynamic target assignment model with the objective of minimizing the overall threat degree of the targets is established. An improved discrete particle swarm optimization algorithm is designed, which firstly generates the initial population according to the threat degree of the surviving targets combining both ordered allocation and random allocation. Then the inertia weights are controlled in the form of Sigmoid function to balance the global and local search ability of the particles in the pre and post search process; the acceleration coefficients are controlled in the form of hyperbolic tangent function to reduce the probability of falling into the local optimal solution. Finally, the target assignment model is solved by the inverse-order recursion method of stochastic dynamic programming to obtain the sequence of weapon target assignment scheme. Results show that the model and algorithm are significantly superior to the single-phase static target assignment in terms of the overall threat expectation of the targets.