A FAST MARGINAL FEASIBILITY SEARCH METHOD IN SIZE OPTIMIZATION OF TRUSS STRUCTURES

Maybe it is reasonable that in optimization problems based on sensitivity analysis one could reach the vicinity of the optimum point with a minor number of analyzes. Besides, it is also fair to accept the fact that the optimum point activates at least one constraint in constrained optimization problems. Based on these concepts a new method is proposed in the present study. It utilizes four well-organized operators to reach the global optimum solution; ordered by first a Subspace Search (SS) operator that transforms the whole design space into a series of subspaces in order to rapidly reach the Feasible-Non-Feasible (FNF) margin at the early stages by doing a few number of analyzes. It is then followed by a Marginal Sensitivity Analysis (MSA) operator that determines the sensitivity degree of each design variable to constraints violation, near the margin of FNF region. Next, the Marginal Search (MS) operator is used to determine a local optimum point near the FNF border in the feasible region. Finally, the roulette wheel (RW) operator is employed to select, in a random manner, only one variable for updating in each iteration. The robustness and effectiveness of the proposed method is verified on several well-known benchmark truss examples. The results show that the proposed method not only speeds up the optimization procedure, but also it ensures the non-violated global optimum design point without a need for multiple runs.