关键词: 差分进化算法;控制参数自适应;轮胎模型;参数识别

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.

Key words: differential evolution;self-adaptive control parameters;tire model;parameters identification