Open Access

ARTICLE

Weight Optimization of Skeletal Structures with Multi-Point Simulated Annealing

L. Lamberti^1,2, C. Pappalettere^1,3

1 Dipartimento di Ingegneria Meccanica e Gestionale, Politecnico di Bari, Viale Japigia 182–70126–BARI–Italy.
2 Email: lamberti@poliba.it
3 Email: c.pappalettere@poliba.it

Computer Modeling in Engineering & Sciences 2007, 18(3), 183-222. https://doi.org/10.3970/cmes.2007.018.183

Abstract

This paper presents a novel optimization algorithm for minimizing weight of skeletal structures. The algorithm--denoted as MPISA (Multi Point Improved Simulated Annealing)--utilizes a multi-level simulated annealing scheme where different candidate designs are compared simultaneously. This is done in purpose to increase computational efficiency and to minimize the number of exact structural analyses.
MPISA is tested in three complicated design problems of skeletal structures: (i) sizing optimization of a planar bar truss under five independent loading conditions including 200 design variables; (ii) sizing-configuration optimization of a cantilevered bar truss including 81 design variables; (iii) sizing-configuration optimization of a frame structure including 84 design variables.
The new algorithm is compared to another multi-level simulated annealing algorithm and to gradient based optimizers recently presented in literature or included in commercial software.
Numerical results demonstrate the very high efficiency of MPISA which achieved weight savings ranging between 250 and 350 kg in all test problems.

---

Keywords

---