舰船电子对抗

软件化雷达中现场可编程门阵列(FPGA)构件在动态可重构场景下划分成多个核心资源同时部署多个任务时,要兼顾芯片的资源利用率与任务运行时间。针对该问题,首先对需要部署的任务进行拆分,并建立资源利用以及任务运行时间的数学模型;接着,用NSGA-Ⅱ来对问题模型求解;最后,将基于方向的启发式交叉算子(DBHX)和组合变异方法应用于非支配排序后的遗传算法,增强了算法的优化能力。结果表明,优化后的解集的评价指标GD值、IGD均小于优化前。且优化后平均资源利用率提高3.5%,平均任务运行时间均衡度提高27%。通过仿真验证,对比NSGA-Ⅱ算法优化前与优化后的结果,在解集质量上验证了改进算法的有效性,为用户部署提供了科学依据。

When field programmable gate array(FPGA) components of software-defined radar are divided into multiple core resources and multiple tasks are deployed at the same time in the dynamic reconfigurable scenario,the resource utilization of the chip and the task execution time should be taken into account.To address this problem,this paper firstly decomposes the tasks deployed and constructs mathematical models for resource allocation and task execution time.Subsequently,NSGA-Ⅱ is used to solve the problem model.Finally,a direction-based heuristic crossover operator(DBHX) and a combined mutation method are applied to the genetic algorithm after non-dominated sorting,which enhances the optimization ability of the algorithm.The results demonstrate that the GD and IGD values of the optimized solution set are significantly reduced compared to those before optimization.Furthermore,the average resource utilization rate increases by 3.5% and the average task execution time balance improves by 27% after optimization.Through simulation verification,the results before and after NSGA-Ⅱ algorithm optimization are compared,and the effectiveness of the improved algorithm is verified in terms of solution set quality,which provides a scientific basis for user deployment.