摘要
针对无约束优化问题经典算法,着重采用最速下降法和阻尼牛顿法进行比对,以求解所述的无约束问题。结果表明,采用阻尼牛顿法较最速下降法具有更快的收敛速度,但阻尼牛顿法不能保证收敛性,必须保证合适的初始设置才可以得到正确的结果。
The classical algorithm for unconstrained optimization problems is emphasized by using the steepest descent method and the damped Newton method to solve the unconstrained problem. The results show that the damped Newton method has faster convergence speed than the steepest descent method, but the damped Newton method can not guarantee the convergence, and the correct initial setting must be ensured to get the correct result.
作者
白雪
侯思微
BAI Xue;HOU Siwei(No.92337 of PLA, Dalian 116023, Liaoning, China;The Society of China Shipbuilding Engineering, Beijing 100001, China)
出处
《机电设备》
2019年第5期67-69,75,共4页
Mechanical and Electrical Equipment
关键词
约束优化
最速下降法
阻尼牛顿法
一维搜索
黄金分割法
unconstrained optimization
steepest descent method
damped Newton method
one-dimensional search
golden section method