一种新的无约束优化问题的混合算法

An innovation hybrid algorithm on unconstrained optimization problems

将传统的一维搜索方法——成功-失败法与新的Apollonius填充算法相结合,给出一种新的平面上的无约束优化方法。该方法既将成功-失败法推广到了平面上,又将Apollonius填充算法的适用对象由约束问题推广到了无约束问题。数值实验表明,该方法适用于较为复杂的非线性可微凸函数,如果对计算时间要求不高的话,该方法可以适用于具有较高复杂度的非线性可微凸函数,具有一定的实际应用价值。

Combing one dimensional research method—Success-Failure method with novel Apollonius fill algorithm,gives a new kind of algorithm of unconstrained optimization method on the plane.The new algorithm has extended Success-Failure method to the plane.At the same time,the scope of application of Apollonius fill algorithm is extended from constrained optimization problems to unconstrained problems.Numerical experiment results show that this algorithm is suitable for complicated nonlinear differentiable convex function.If the calculation time is not highly required,the algorithm can be applied to any complicated nonlinear differentiable convex function.Whereby indicating that this algorithm is of the highly practical application value.