Simulated Annealing for the Optimization of Trusses

Abstract

The Simulated Annealing (SA) Technique belongs to a Stochastic Optimization class of algorithms. This technique has been used as a soft computing technique in hard optimization tasks, such as, electronic components allocation, spatial representation of chemical compounds and Traveling Salesman type problems, for a long period. This technique is based on the mathematical description of the experimental cooling technique developed to design stronger crystals (like glass) and metals. In this paper this technique was implemented on a Matlab (Matlab, 2001) environment and applied to simple and difficult parametric truss optimization problems with constraints in displacements and stresses. The examples were selected in order to compare results with those presented by related literature. SA Technique performance is compared with those obtained with other heuristic methods like the Genetic Algorithm (GA) and with gradient based Mathematical Programming Algorithms, such as Sequential Quadratic Programming (SQP). The presented results show some disadvantages regarding the computational cost using the SA Technique, nevertheless the final results show better or similar accuracy than the ones obtained with the other methods.