Abstract: Based on interval numbers decision with uncertain information, the attack-defends game model of unmanned combat aerial vehicle is established. The combat situation is introduced into interval payoff matrix, and the distribution of the point within the interval is reformed and transformed into four-parameter interval number. A new decision method of target assignment combined with relative entropy is put forward, which has the advantages of high differentiation degrees during the decision-making process. Then, the method to solve the optimization problem is denoted and the Nash equilibrium is attained by adaptive PSO method. The simulation result confirms the feasibility and validity of this method used in unmanned combat aerial vehicle game theory.

Key words: Unmanned Combat Aerial Vehicle(UCAV), interval numbers decision, relative entropy, Nash equilibrium, particle swarm optimization

摘要： 针对不确定信息下的无人机攻防博弈问题，建立了基于区间数的无人机攻防博弈模型。将态势信息引入支付区间矩阵，重构了支付区间内点的分布并将其转化为四参数区间数。结合信息论提出了一种基于相对熵的目标分配决策模型，提高了对不同策略的区分度。给出了参数不确定优化问题的求解方法，并使用自适应粒子群算法求解了纳什均衡。仿真结果验证了该方法在无人机攻防博弈中的可行性和有效性。

关键词: 无人作战飞机, 区间数决策, 相对熵, 纳什均衡, 粒子群优化