论文部分内容阅读
天基预警过程可以看作一种多维离散时间序列监控与预测问题,其调度的决策要素、优化目标和约束条件较多,故往往采用智能优化算法求解该非线性优化问题.而它们在指定时间内却是概率性收敛到Pareto解集.对此,提出基于贝叶斯方法提供多类别决策树挖掘调度中的启发信息,以及引入局部搜索算子等方法提高智能优化算法的快速性和鲁棒性.预警仿真实验表明融入上述方法的免疫克隆选择算法收敛性能提高了10.1%,遗传算法提高了9.8%.
The space-based early-warning process can be regarded as a multi-dimensional discrete time series monitoring and forecasting problem, and its scheduling decision-making elements, optimization goals and constraints are often more, so intelligent optimization algorithms are often used to solve this nonlinear optimization problem, It converges probabilistically to Pareto solution sets.In this paper, we propose a Bayesian approach to provide heuristic information in mining scheduling with multi-category decision trees and introduce local search operators to improve the speed and robustness of intelligent optimization algorithms The simulation results of early warning show that the immune clonal selection algorithm incorporating the above method improves the convergence performance by 10.1% and the genetic algorithm by 9.8%.