路面结构模量反演的蚁群算法研究

Back-calculation of Modulus for Pavements Based on Ant Colony Algorithm

路面结构模量反算是一个复杂的非线性优化问题.通过对待反算参数进行离散,将路面结构参数反算问题转化为组合优化问题,建立了基于蚁群算法的路面结构参数智能反演模型.通过理论和实测弯沉盆的反算及分析,获得了满意的反算结果.研究表明,采用蚁群算法进行路面结构层模量反演,有效地解决了常规最优化算法的初始值和局部收敛的问题,不需多次调用正分析程序,提高了计算效率,是一种精度好、速度快、结果稳定且全局收敛的模量反算方法.

The back analysis of modulus for pavements is a typical complicated nonlinear function optimization problem.The question is transferred into finding the shortest paths for ants according to extreme conditions of optimization problems.A mathematical model for back-calculation of pavement layer modulus was established based on the Ant Colony Algorithm with the paths of ants responsible to the modulus of pavement.A set of numerical experiments were conducted to illustrate the methodology.The results achieved were compared with those previously obtained by simulation procedures.The study shows that the problems of initial values and partial convergences for conventional optimum algorithms are solved efficiently by using ant colony algorithm method to back calculate layer modulus of pavements.Therefore,using ant colony algorithm method to back-calculate layer modulus of pavement is great in accuracy,fast in speeds and stable in results and wide in convergence.