トラス等の離散構造物の位相最適化では、Dornらが考案した グランドストラクチャ法(以降、GS法)が利用されることが多い。 GS法は断面積最適化手法を拡張した単純な手法であり、逐次2次 計画法等の非線形計画法(以降、NLP)を利用することで容易に最 適位相を求めることができる。%A new approach integrating heuristics and nonlinear programming is proposed for topology optimization of discrete structures under stress and local buckling constraints. In the approach, the optimal ground structures are searched by heuristics, and topology optimization is performed for each ground structure based on nonlinear programming. Two heuristics are applied; namely, tabu search, and a new heuristic method called topology mining, which is based on the apriori algorithm. Three numerical examples are solved to demonstrate the effectiveness of the proposed approach. Moreover, the efficiency in view of computational cost of the two heuristics is discussed by comparing the optimization results.
展开▼