Enhanced Hybrid Metaheuristic Algorithms for Optimal Sizing of Steel Truss Structures With Numerous Discrete Variables

Abstract

The advent of modern computing technologies paved the way for development of numerous efficient structural design optimization tools in the recent decades. In the present study sizing optimization problem of steel truss structures having numerous discrete variables is tackled using combined forms of recently proposed metaheuristic techniques. Three guided, and three guided hybrid metaheuristic algorithms are developed by integrating a design oriented strategy to the stochastic search properties of three recently proposed metaheuristic optimization techniques, namely adaptive dimensional search, modified big bang-big crunch, and exponential big bang-big crunch algorithms. The performances of the proposed guided, and guided hybrid metaheuristic algorithms are compared to those of standard variants through optimum design of real-size steel truss structures with up to 728 design variables according to AISC-LRFD specification. The numerical results reveal that the hybrid form of adaptive dimensional search and exponential big bang-big crunch algorithm is the most promising algorithm amongst the other investigated techniques.