摘要
为有效解决频率指派问题,提出了一种解决该问题的曲面拟合Multi-Quadric算法,算法将随机指派方案及其对应的干扰值作为多元散乱数据采样点,以此为基础进行多元散乱数据拟合,通过对拟合曲面极小值的寻找从而完成对频率指派问题的优化。优化结果可直接应用于实际工程,也可作为其它优化算法的初始解。算法在电台数量规模较大的应用中体现出良好的性能,算法结果作为蚁群、遗传算法的初始解,后继算法收敛速度明显提高。
The multi-quadric surface fitting algorithm is applied to solve the frequency assignment problem. The algorithm takes randomly assigned solutions and their corresponding interference values as scattered multiple sampling points, and multivariate scattered data is fitted based on them. The optimization of frequency assignment problems is completed by finding the minimum of fitted surface. The solution correspond to the minimum is optimal solution. Optimized results of the multi-quadric surface fitting algorithm is directly applied to actual engineering project, and also is used as the initial input of other optimization algo- rithms. Optimized results of the multi-quadric surface fitting algorithm is used as the initial solution, algorithms like ant colony algorithm and genetic algorithm reflects high efficiency in the application that have large-scale radio stations and convergence rate is improved significantly.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第10期3853-3857,共5页
Computer Engineering and Design
基金
航空科学基金项目(20100796004)
关键词
频率指派
曲面拟合
多元插值
Multi—Quadric算法
单形调优法
Key words： frequency assignment
surface fitting
multivariate interpolation
Muhi-Quadric algorithm
simplex designs in op-timization
