基于改进多目标蝙蝠优化算法的大规模电子干扰任务分配

Task allocation for large-scale electronic warfare jamming based on improved multi-objective bat algorithm optimization

摘要:

作战过程中,作战分组复杂,电子战装备种类繁多,为指挥员的指挥行动带来了极大的挑战。为了解决这一问题,首先针对多分群、多类型装备和多作战目标建立了电子战干扰任务分配(electronic-warfare jamming task assignment, EJTA)模型。该模型为复杂的多目标组合优化问题,具有多决策变量和多目标函数的特征。接着,提出了一种基于角度分解的改进多目标蝙蝠优化算法(multi-object improved bat algorithm optimization based on angle decomposition, MOIBA/AD)。此算法将传统基于分解的多目标进化算法(multi-objective evolutionary algorithm based on decomposition,MOEA/D)的平面空间分解策略改进为角度空间分解策略,降低了个体处理难度和算法复杂度;并将种群更新策略通过改进的蝙蝠算法具有的寻优特点进行增强,使之不易进入局部最优,有效提高了算法的种群更新效果。最后, 将MOIBA/AD与几种经典的和最新提出的多目标进化算法进行对比,分别对2种不同规模的EJTA模型进行求解。通过3种性能指标的对比,表明MOIBA/AD能够有效求解EJTA 模型,并能够较好地保持Pareto解集的分布性。

Abstract:

During operations, the complexity of combat formations and the diversity of electronic warfare equipment have posed significant challenges to commanders? command actions. To address this issue, firstly, an electronic-warfare jamming task assignment (EJTA) model was developed for handling multiple clusters, diverse equipment types, and numerous operational objectives. This model is a complex multi-objective combinatorial optimization issue, characterized by numerous decision variables and multiple objective functions. Subsequently, a multi-object improved bat algorithm optimization based on angle decomposition(MOIBA/AD) was proposed. This algorithm enhances the traditional multi objective evolutionary algorithm based on decomposition(MOEA/D) by replacing its plane space decomposition strategy with an angle-based decomposition approach, thereby reducing individual processing complexity and algorithmic overhead. Furthermore, the population update strategy was refined through the optimization characteristics of the bat algorithm, making it less likely to converge to local optima and significantly improving the effectiveness of the algorithms population update. Finally, comparing MOIBA/AD with several classic and recently proposed multi-objective evolutionary algorithms, the EJTA model of two different scales was solved respectively. Through the comparison of three performance metrics, it has demonstrated that MOIBA/AD can effectively solve the EJTA model and maintain the distribution of the Pareto optimal set well.