Genetic Algorithms For Optimal Design Of The Combinative Cylinders Paperhanging Assembled Structure

On the background of selecting an optimal combination of elastic pads' height between the two adjacent cylinders, which carry out the design requirement of combinative cylinders assembled structure, in this dissertation, the optimal design model is established ground on geometrical and mechanical model of the combinative cylinders assembled structure. The algorithm that integrates multi-objective's decision-making method with minimal distance between two points in Euclid's space into the modified genetic algorithm to solute the optimal model is presented.Firstly, the structure's geometrical design parameters are expressed with the discrete design variables. Secondly, the structure's mechanical model that based on the elastic material mechanical characteristic and a certain enclasped state and enclasped thinking of multi-layers' shell is established. The elastic pad's heights comprise of some basic discrete height. At the same time, there are some design targets during the assemble process. So the problem belongs to structural optimization with discrete variables and multiple objectives.Based on the genetic algorithm's global searching capability with probability regulation and Euclid's space distance metric to settle multi-objective, the algorithm that integrates multi-objective's decision-making into the modified genetic algorithm to solute the optimal model with discrete variables and multi-objective is proposed. During the algorithm's design, the Euclid's space distance metric is proposed to transform the multi-objective problem into single objective problem. And some modified measure to fitness function and crossover probability and mutation probability are used to improve the performance of the algorithm and avoid premature convergence.The modified genetic algorithm that is analyzed with the schema theorem is feasible. The results obtained from the modified genetic algorithm indicate the algorithm is feasible and effective by some representative examples. Finally, the procedure named GA-MOSOD is programmed with Visual C++6.0 software based on the above-mentioned design thinking.