应用差分进化–神经网络模型的杀爆弹瞄准点分配方法

In order to improve the damage effectiveness of multiple anti-explosive shells against targets without increasing computational time consumption, a multiple aiming points planning method incorporating dynamic explosive power calculation was established. Accurate calculation of target damage area caused by multiple blast-fragmentation warheads was made using structured grid generation method for area targets, and computational results were verified. Based on simulation calculation data of multiple dynamic explosion damage effects, neural network method was used to establish a computational agent model of single ammunition for area target damaged area. Under the same calculation conditions, its calculation time was 1 000 times shorter than that of non-agent model. Based on this, the differential evolution algorithm was used to plan the aiming point and terminal ballistic parameters of multiple blast-fragmentation warheads. Case analysis shows that this aiming point planning method has greatly improved the damage efficiency compared with the traditional method with damage radius as the input, and the minimum increase is 25.5%, and the single planning time does not exceed 3 seconds, solving the contradiction between the complexity of the damage efficiency model and the computational time in aiming point planning.