【摘要】 针对传统差分进化法固定的控制参数设置与进化策略难以适应复杂多变的问题的状况,提出一种新的自适应差分进化方法,并将其应用于魔术公式(Magic formula,MF)轮胎模型的参数辨识中以解决其参数辨识难的问题。该方法结合基于成功进化个体的控制参数选择策略以及基于双审判矢量的进化策略,实现控制参数的有效自适应。通过对纯侧偏工况下轮胎数据的侧向力和回正力矩参数辨识,证明该方法比另外两种先进的自适应差分进化算法IOA和SspDE具备更好的全局优化与快速收敛能力,也比传统的数值优化Levenberg-Marquardt方法识别精度更高,是辨识MF轮胎模型参数的一种有效手段。更多还原

【Abstract】 According to the situation that the fixed control parameters and evolutionary strategy of traditional differential evolution cannot be adapted to complex problems,a new self-adaptive differential evolution(NSADE) is presented,and further utilized to overcome the difficulty of identifying the parameters of magic formula(MF) tire model.The algorithm combines the selection strategy of control parameters based on the successfully evolutionary individuals and the evolutionary strategy based on dual trial vectors to achieve the effective self-adaptation of control parameters.By the identification of lateral force parameters and aligning moment parameters for pure slip,the new algorithm is demonstrated to have more powerful ability of global optimization and fast convergence than the other two advanced self-adaptive differential evolution,IMMa optimization algorithm(IOA) and differential evolution with self-adapting strategy and control parameters(SspDE),and the identified results are more accurate than that of the traditional Levenberg-Marquardt method,which indicates NSADE is an efficient method to identify the parameters of MF tire model.更多还原