Application of Improved NSGA-II Algorithm in Matching Optimization for Tractor Powertrain

Abstract. To optimize matching of the tractor powertrain and improve the performance of the tractor, a novel matching optimization method for tractor powertrain was proposed based on the improved non-dominated sorting genetic algorithm-II. The normal distribution crossover operator and the differential evolution mutation operator based on the differential evolutionary algorithm were introduced to expand the spatial search range and improve the uniformity of population distribution. Subsequently, the optimization model of transmission ratios was established with constraints such as vehicle speed, ratios of gear ratios, driving adhesion restriction, and so on. In this model, gear ratios were taken as input variables, and the optimization objective was to the lowest drive power loss rate and the lowest specific fuel consumption loss rate. The proposed algorithm was used to optimize the tractor transmission ratios and compared with the original NSGA-II. The experimental results show that after optimized by improved NSGA-II, the drive power loss rate and the specific fuel consumption loss rate of the tractor could be theoretically reduced by 42.62% and 63.80% than before, respectively, which is better than NSGA-II. The overall performance of the tractor has been improved obviously which verifies the effectiveness of the improved NSGA-II algorithm.