Structural optimization for mechanical elements by modified immune genetic algorithm

In view of the low convergence of the current immune genetic algorithm(IGA),6 inherent drawbacks were found existing in the current rationale of IGA.For these reasons,6 new immune strategies which were corresponded to above drawbacks were put forward to accelerate the convergence speed and precision.The relationship between real-coded antibodies and variables vector was also presented,and the complicated algorithm mechanism was intuitively described by graphic,which made people understood easily and practically operate on IGA.The stochastic perturbation method of the reliability was combined with its sensitivity analysis to deduce their function formulas on mechanical parts whose probability distributions of random parameters were of arbitrary shape.The two-objective mathematical models were created to minimize the sensitivity of reliability with respect to designing variables and the volume.A new model was creatively proposed that could realize the dynamic balance between objectives,by which the objectives were smoothly combined with the image set method.Finally,the shafts of 3 stages planetary gear in shield tunneling machine were optimized as an example by means of the MATLAB programs languages that applied the modified IGA after their mathematical models were constructed according to the ideas above.The results show that the robustness of the improved IGA not only decreases the total volume of shaft of planet reducer by 13.65%,but also its velocity and accuracy of convergence are superior to that of the unimproved IGA.