基于自适应信息遗传算法的路面模量反算研究

近年来,根据FWD实测弯沉盆来反算各层路面结构模量的技术在路面结构的质量评价中广为应用。如何利用FWD反算路面模量也相应成为各方研究的热点。在本文中,作者针对于模量反算中,目标函数在最优值附近表现为大片狭长平坦区域的特点,提出了一类反算新算法——自适应信息遗传算法,提出了根据信息量大小来决定算法是否进行自适应细分模量解空间的机制,以缩小算法后期反算中的搜索空间。此外,本文还改进了实数交叉算子,使用了新的探险策略,从而加强了新算法后期的局部搜索能力。随后的工程实例表明,这新机制、新策略的引入,能够在保持较高反算精度的前提条件下,大大提高模量反算的求解效率,从而更加有利于模量反算的工程实际应用。

Adaptive Informational Genetic Algorithm Based Backcalculation of Pavement Layer Moduli

In recent years,backcalculating the pavement layer moduli based on FWD deflections is widely used for quality evaluation of pavement structures.Since its introduction,the issue of backcalculation of pavement layer moduli has been paid considerably attention in pavement analysis.Due to the existence of a long flat region around the optimum value in objective function,a new backcalculating algorithm,i.e.,adaptive informational genetic algorithm,is proposed in this paper.One advantage of our new algorithm is that it can subdivide the searching moduli space according to the level of information,adaptively,thereby saving a great amount of computing time.Furthermore,a new detective strategy,as well as a modified real-coding copulation operator,is used to improve the local research ability of the new algorithm.The following examples of engineering illustrate that,due to the introduction of new mechanism and strategy,the new algorithm improve the backcalculating efficiency greatly for given precision of results.This means that the new algorithm can better meet the requirements for further engineering application.