机械产品曳引系统的优化设计方法

Optimal design method of traction system in mechanical products

针对机械产品曳引系统的性能优化,应用非支配排序遗传算法(NSGA-II)求解曳引性能的三目标优化问题,即最大曳引效率、最小曳引功率和最小制动力矩.根据曳引性能优化模型中设计变量分为连续值和离散值的特点,引入浮点数与二进制数混合编码策略.通过改进NSGA-II的二进制交叉、变异规则,保证了设计变量的全局寻优能力和有效性,使得算法一次运行就能够求得分布均匀的Pareto最优解集.实验数据分析表明,采用混合编码策略,NSGA-II算法较线性加权法和Pareto强度进化算法(SPEA)能够获得边界性和分布性更好的Pareto最优前沿.

Aiming at the performance optimization of the traction system in mechanical products, non-dominated sorting genetic algorithm （NSGA-Ⅱ） was used to solve the optimal problem of traction performance with three objectives： maximum traction efficiency, minimum traction power and minimum braking torque. According to the feature that design variables were classified into continuous and discrete values in the optimization model of traction performance, a float/binary hybrid chromosome representation scheme was introduced. The binary crossover and mutation of NSGA-Ⅱ were also improved to ensure the global search capability and the validity of design variables, so a well-distributed Pareto optimal set could be achieved in a single run. Experimental results showed that under the hybrid chromosome representation scheme, the improved NSGA-Ⅱ can acquire Pareto optimal fronts with better boundary and distribution than those of the linear combination and the strength Pareto evolution algorithm （SPEA）.