Abstract: During the course of selective assembly of internal-combustion engine linkage of crank-rod, in order to achieve multi-objective that includes mating several pairs of parts in respective assembly precision demands at the same time, making the surplus parts minimum and increasing the mate rate and the assembly quality of the parts within the linkage of crank-rod, an objective function of multi-objective selective assembly is built and restriction conditions are defined. Then considering the speciality of this combinatorial optimization problem, the conventional hybrid algorithm based on genetic algorithm and simulated annealing algorithm is improved, and a method, to take a directional mutation when forming the initial population, is presented to boost the average fitness of the initial population, meanwhile, specific coding operation, specific crossover operator and specific mutation operator are designed, thereby a multi-objective selective assembly method of the linkage of crank-rod is built. Finally the validity of such method is proved by a practical example.

Key words: Directional mutation, Genetic algorithm, Linkage of crank-rod, Multi-objective, Selective assembly, Continuous generating grinding, Feed rate, Gear, Noise, Tooth surface texture