STRACTURAL MORPHGENESIS AND OPTIMIZATION BY AUTONOMOUS DECENTRALIZED FINITE ELEMENT METHOD

近年来，结构工程在找形分析、形状分析和优化设计等方面取得了显著的进展。解决这些问题的方法和程序多种多样。近年来，进化计算(如遗传算法和免疫系统)等仿生方法在许多领域的应用迅速增加，本研究项目讨论了仿生方法在形状分析和弹性结构优化设计中的适用性。我们的主要求解方法是用有限元方法结合自主分散的概念来发展的。这个概念是生命或生物系统的特征之一。自主性的概念是在单元自动机中用局部规则的思想来实现的。为了有效地求解代数方程组，在求解过程中引入了离散化的概念和松弛技术，并给出了弹性结构的形变算例，以说明该方法的有效性和有效性。它们是平面桁架的最小重量设计问题，以及二维弹性体中的空间桁架结构和自组织问题。

Recent advances on form finding analysis, shape analysis and optimum design of structures in structural engineering are remarkable. There exist various solution methods and procedures to solve these problems. In recent years, a rapid increase in the use of life-like or biological approach such as the evolutional computation (for example, the genetic algorithms )GA) and the immune system) have seen in many fields.In this research project the applicability of life-like approach to shape analysis, and optimum design of elastic structures is discussed. Our main solution method is developed with the finite element method in conjunction with the concept of autonomous decentralization. This concept is the one of characteristics in life or biological system. The concept of autonomousness is implemented with the idea of local rule in the cell automaton. And the concept of decentralization is introduced with the relaxation technique in solution procedure in order to solve effectively the system of algebric equations.As numerical examples, the morphgenesis of elastic structure is given in order to show the effectiveness and validity of our method. They are minimum weight design problems of plane truss, and spatial truss structures and self-organization problem in two-dimensional elastic body.