The load sharing characteristics and lightweight were considered as the objectives

so a multi-objective optimization design of torque-split transmission system was conducted. First of all

a non-linear dynamics model of torque-split transmission system was established. Load sharing coefficients under different powers and speeds were calculated to measure load sharing characteristics of torque-split transmission system. Afterwards

a multi-objective optimization model including minimum load sharing coefficient and mass was developed with the torque-split transmission system parameters selected as design variables

and the effects of different working conditions were taken into consideration. In order to improve the solution efficiency

a modified fitness approximation mechanism non-dominated sorting genetic algorithm (FA-NSGA) was promoted. Three benchmark functions were used to test the convergence and effectiveness of FA-NSGA. The test results show that satisfactory optimal solutions in all three test functions are achieved by using FA-NSGA

and more than 60% of the computation times of real fitness can be reduced at the same time. Finally

a given instance was solved by using FA-NSGA

and the satisfactory design variables were selected from the Pareto optimal solution. Compared with the reference design

the selected one reduces the load sharing coefficient by 0.05

and the mass by 3.57kg at the same time.