【摘要】 本文提出一种可实现非线性方程最优拟合的新方法——缩张算法(Contraction-Expansion Algorithm,简称C-E算法)。该算法由若干循环组成,每一循环都包含收缩和扩张两个步骤。因而,可以在多维起始空间的内、外漫游搜索目标函数,并利用搜索过程所反馈的信息调整搜索中心和步长;通过较少次C-E循环的搜索,便能逼近给定非线性方程的最适参数值。该算法不需要给出方程的导数或偏导数,减少了计算的复杂性,也可能推广应用于解决其它目标函数的最优化问题。更多还原

【Abstract】 The paper developed a new method, Contraction-Expansion Algorithm, whichcan optimize the fitting of the nonlinear equation. The algorithm is formed with a number of circles, each with contraction and expansion two stages, so it is called C-E algorithm. C-E algorithm searches for the parametric points of better objective function in and around initial area,and it optimizes searching process through adjusting the central point and step length by the feedback information. With several circles, it can reach global optimal fitting under any initials. Thisalgorithm has the simplicity of free of derivatives of the equation. C-E algorithm can also be usedin other objective optimization problems.更多还原