Cooperative multiple target assignment using Monte Carlo sampling
Air Combat Decision-Making for Coordinated Multiple Target assignment is an important yet difficult problem in the modern information warfare. Existing methods, such as neural network, genetic algorithm, ant colony algorithm, particle swarm optimization algorithm and auction algorithm, used to resolve this problem are mainly centralized. In order to meet the needs of modern information warfare, a distributed cooperation method using Monte Carlo sampling is proposed. In this method, agent updating algorithm uses Probability Collectives proposed by NASA Research Center, in which probability of actions is updated independently. Each agent estimates the expected value of global utility function by memorizing the value of global utility function using Monte Carlo sampling, regardless of the other agents' choice. By this method, system has a high anti-jamming capability when the communication is interfered. Experiments show the proposed method can estimate the expected value of global utility function effectively and make the system achieve good performance.
Author's Name: Zhang, X., Yu, W., Liang, J., Liu, B.
Volume: Volume 6
Issues: Issue 4
Keywords: Air combat decision-making, Monte Carlo sampling, Probability collectives