Multi-objective optimization of grillages using adaptive genetic algorithm

Abstract

The paper proposes a technology for multi-objective global optimization of grillage-type foundations applying adaptive genetic algorithm (AGA) seeking for the minimal consumption of material in the foundations; this is achieved by minimizing the number of supporting piles and cross-section dimensions of connecting beams. Compromise objective function to be minimized involves the maximum reactive force in piles and maximum bending moment in the connecting beams. Feasibility of a simple weighting technique for the objective function is proved by numerical investigation of objective function domain for several different topologies of foundations. Sizing problem of connecting beams is solved together with the optimization problem. Original finite element program was employed for solution of direct problem. The effectiveness of AGA is shown by optimization of two small-scale grillages possessing up to 17 design parameters. Though the AGA does not yield qualitatively better solutions, it obviously saves the analysis time to obtain the same or even better results due to the faster convergence.