摘要
本文将演化算法与模拟退火算法集成,提出一种适用的全局演化局部模拟优化技术,该技术使用演化算法在解空间通过较少代数的演化为模拟退火算法提供一个良好的算法构形,在此基础上,通过退火方式寻找全局最优解。本文探讨了该算法在汽车操纵稳定性评价与结构优化中的应用问题。优化以最小化汽车操纵稳定性综合评价指标为目标对评价参数与结构参数进行优化。使用优化参数在双移线条件下对四自由度汽车动力学模型的操纵稳定性进行仿真研究。通过仿真对比分析了演化算法与模拟退火算法,本文使用的算法继承了这两种算法的优点。研究结果表明:使用全局演化局部模拟优化技术所得综合评价指标值最小,耗用CPU时间也最少,而且所得优化参数使得人-车闭环系统跟随预期路径的精度提高约58.31％,而使综合评价指标值下降约37.35％。
Global evolution local simulation annealing algorithm, a global optimum algorithm which integrates simulation annealing algorithm with genetic algorithm, is presented in this paper. Through less generations' evolution, the genetic algorithm provides a good variable configuration for the simulation annealing algorithm, and the global optimal variable configuration can be got based on this method. Application of the algorithm to the automobile maneuverability evaluation and structure parameters optimization is studied. Optimization of the structure parameters is based on the minimization of the comprehensive evaluation index. Simulation study on the 4-DOF vehicle dynamic model in the condition of double-lane change is performed by using the optimal parameters. The study shows that, by using the algorithm, the CPU time is minimum, the precision of the driver-road closed-loop system following the predictive route is increased by 58.31% and the comprehensive evaluation index is decreased by 37.35%.
出处
《汽车工程》
EI
CSCD
北大核心
2001年第2期73-77,共5页
Automotive Engineering
基金
高等学校重点实验室访问学者基金资助项目
国家自然科学基金资助(59975041)
