Artificial Immune system based Bio-computational Algorithm for Optimum Structural Design

• Main Authors: Tzu-Lun Kuan, 管姿倫
• Other Authors: C.J.Shih
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/48355207305078892146

碩士 === 淡江大學 === 機械與機電工程學系 === 91 === Immune system is an intelligent biologically reacting system that contains extremely strong matching and memory ability. The main function is to identify the antigens first and then through a prompt and complicate process to form antibodies. This thesis use the theory of biologically immune system to build an artificial immune algorithm (IA) of engineering optimization for solving general optimum design problems. A modified expression operator has been developed for the treatment of design constraint functions. The basic idea is on the basis of natural random search combined with the gene exchange gradually to improve the feasibility of infeasible cells. There is no need to select certain un-predetermined parameters and create a certain transformation function merging the constraint in the objective function. It shows the modified expression operation is much reliable and efficient than that in generaltransformation approach. Artificial immune system has been applied to constrained single-objective optimization problem as well as to constrained multi-objective optimization problem. Especially in IA based multi-objective optimization approach can obtain a set of Pareto solution design in one phase that is particularly convenient for a designer to select a final suitable design. Global criterion method has been combined with IA based multi-objective optimization approach to obtain the unique design that has the shortest distance to the theoretically non-existing ideal design. Several design examples illustrate the usability and advantages of presenting IA based single- and multi-objective optimization method that is good to apply general engineering design problems.