研究用于结构拓扑优化的基于材料添加策略的进化算法(AESO方法)。基于进化算法的思想,利用单元材料相对密度的变化描述材料的添加(从0到1)或删除(从1到0)。当某些单元满足进化准则时,单元的相对密度进行0—1变化。研究发现.基于一步变化策略的AESO方法往往不能获得正确的拓扑形式,其原因可能是,进化后的响应量是基于密度为0或很小时的敏度经线性近似获得的,与实际相差很大。这种敏度的计算误差问题在ESO、BESO等硬杀算法中都存在。本文提出将进化过程分成多步,以软杀的思想进行硬杀优化,即使材料密度逐渐由0变化到1,实现材料的逐步添加。基于该策略,提出了渐进密度AESO方法.并比较分析了这种逐步添加的做法对结果的影响。算例验证了该方法的正确性和有效性。渐进密度AESO方法为双向进化算法(BESO)提供了有效的进化(材料添加)策略。
This work presents a new version of the additive evolutionary structural optimization (AESO) procedure based on sensitivity analysis for topology optimization of continuum structures. It has been proved by illustrative examples that the one-step AESO algorithm based on sensitivity analysis can't obtain good optimal result in the optimization of continuum structures. The reason of failure was pointed out as.. the sensitivity of design variables can't accurately describe the relationship of objective function and design variables when the design variables are changing too much (from 0 to 1). And a new version so called progressive AESO algorithm based on sensitivity analysis was proposed and demonstrated by illustrative examples of heat conduction structure optimization. By which the strategy of adding material in the biirectional ESO algorithm based on sensitivity analysis was provided.