Morphogenetic Designof Truss Structures Based On Genetic Algorithm

Topology optimization is the hot topics of structural optimization, truss structures are the first choice of studying structural optimization because of its widely application in projects and easily calculation in mechanics. Traditional mathematical induction method requires variables to be continuous and objective function to be differentiable. However, the variables in truss structures are discrete. The traditional structure optimization design method with discrete variables has great limitations, because it is difficult to converge and obtain the optimum solution. These problems, especially the discreteness and global optimization, encountered in the traditional algorithm can be solved by the genetic algorithm in truss optimization design, because, the genetic algorithm has low requirement of variables and high global search capability.Based on the genetic algorithm, the optimization design of truss structures has been accomplished in FORTRAN. At present, the computing efficiency is reduced when rod elements are used as chromosome genes in the optimization design of truss, because lots of mobile systems are produced during the analyzing. The triangles consisting of rod elements are adopted as chromosome genes to reduced the mobile system and improve the computation efficiency. For improve the efficiency, the topology form of truss structures, cross-sectional area of members and structural shapes are taken as the three variables of genetic algorithm and are optimized separately. Moreover, the parallel designs for two variables (truss structure topology form and cross-sectional area of members) and three variables (topology form of truss structures, cross-sectional area of members and structural shape) are also carried out in parallel. The results of the examples of plane trusses and spatial trusses show that:(I) the optimization design based on GA is easy to implement; (II) the optimum solution can be obtained efficiently by GA.For meeting the requirements of structural functions, for example, the top surface having same displacement under the weight of precision instruments, it is a problem of truss structure optimization to seek truss structure form meeting such requirements. The optimization design on the constraints of same displacement in plane trusses and spatial trusses is carried out by genetic algorithm. The results of the examples indicate that it is feasible and effective to do the optimization design on the constraints of same displacement in plane trusses and special trusses based on the genetic algorithm.