ON THE USE OF GENETIC ALGORITHMS TO ASSESS THE SEISMIC RESISTANCE OF PLANAR FRAME STRUCTURES

Abstract. Limit analysis represents a widely adopted strategy for the assessment of the bearing capacity and collapse mechanism of frame structures. Within this framework, in this paper an original strategy for the effective evaluation of the ultimate resistance and the corresponding failure mechanism of planar frame structures subjected to lateral loads is presented. The methodology is based on the generation of the elementary collapse mechanisms to be combined following a collapse load factor minimization criterion. When a large number of possible combinations has to be investigated a prompt procedure able to quickly converge to the actual collapse load factor and to jump out of local minima is needed. For this reason a procedure based on genetic algorithms is here adopted and a dedicated user-friendly software was developed in the NetLogo programming environment. Validations of the proposed procedure with respect to nonlinear static analysis are reported together with some significant sensitivity analyses with respect to load distribution parameters. The results demonstrate the reliability of the procedure and can provide useful information also in view of seismic design optimization strategies.