A simulated annealing approach to the economic optimization of reinforced concrete cantilever retaining walls

Abstract

This paper presents an algorithm based on simulated annealing applied to the economical optimization of retaining RC walls. Solutions, based on 26 design variables, are checked against failure due to overturning, sliding, excessive stresses on the foundation and material strength compliance. Results show that the artificial intelligence algorithm finds relations between the footing dimensions and the permissible bearing stress. It also leads to solutions that use the minimum concrete strength and the maximum steel strength. However, excessive slenderness of the optimized walls indicates the necessity of including in the analysis limit states that are usually ignored, e.g. the limit state of deflections.