Abstract: Aiming at “black box” of neural nets, this paper proposes a Particle Swarm Optimization(PSO) algorithm to extract gear-shifting rules from automatic transmissions, analyzes how to encode extracted rules into particle swarms, discusses the process through which the optimal rules are generated by PSO’s velocity-position model, compares the performance between basic and adaptive PSOs in terms of their abilities to search for the optimal solution. Experimental results demonstrate the effectiveness of the algorithm.

Key words: gear-shifting rule, rule extraction, particle swarm optimization algorithm

摘要： 针对神经网络“黑箱”模型的缺陷，利用粒子群优化的换档规则提取算法，将规则编码为粒子的方法，通过粒子群优化算法的“位置-速度”搜索模型生成换档规则集。实验分析了标准粒子群与惯性递减粒子群在最优解搜索过程中的性能差异，并验证了该方法的有效性。

关键词: 换档规则, 规则提取, 粒子群优化算法

CLC Number: 

• TP393