基于混合算法的机车二系弹簧载荷调整优化方法

Optimization Model for Adjustment of Locomotive Secondary Spring Load Based on Hybrid Algorithm

针对电力机车二系悬挂调簧分析数学模型算法的优化问题,提出一种遗传算法(GA)与蚂蚁算法(AA)相结合的混合优化算法。其基本思想是:首先采用遗传算法以较少的进化代数进行全局快速随机搜索,获得若干可能的(近似)优化解,以此生成蚂蚁算法初始信息素分布,再用后者求得全局优化精确解。对国产SS3B和SS9型机车的应用结果表明,对同一车体进行多次优化计算试验,混合优化算法的搜索寻优过程均能稳健一致地收敛到全局优化解,可明显缩短二系支承载荷调整调簧计算所需时间,使调簧试验的实时性大为提高。对于二系为高圆簧的SS9型机车,混合算法平均用时比迭代算法和单一遗传算法分别减少约74%和29%。

Aimed at the optimization of the electric locomotive＇s secondary spring loads adjustment, a hybrid optimal algorithm based on the combination use of Genetic Algorithm （GA） and Ant Algorithm （AA） was presented. With its ability of quick and global stochastic searching, GA is firstly adopted to find numbers of approximate optimal candidates, by which initial pheromone distribution is generated. And then precise optimized solution to the problem is finally obtained by using AA. Results of the application to SS9 and SS3b locomotives show that the new approach has the ability to find global optimal solution to the problem with high stability and consistency, and enhances real-time performance of fast secondary load test by remarkably reducing computation time. Compared with the iterative method and single-GA algorithm, the average computing time required by hybrid algorithm is reduced by 74% and 29% respectively for the optimal load adjusting of SS9 locomotive with flexi-coil secondary springs.