The Performance Comparison of Three Metaheuristic Algorithms on the Size, Layout and Topology Optimization of Truss Structures

Abstract

The structural optimization problem mostly deals with the weight minimization of the structural system. This issue can be assessed from the size, layout and topology aspects. No matter which aspect(s) is targeted, to solve the problem an optimization technique is required. In the last decades the metaheuristic techniques, as the non-gradient optimization algorithms, are widely applied on solving these classes of problems. In the structural optimization, the most time consuming part of the process is the objective function evaluation. Based on this fact, in the current work, metaheuristic techniques are divided into three main groups as single phase, double phase and multi-phase algorithms. Then based on the author knowledge, three representative methods are picked for each group and their search performance comparatively inspected on solving size, shape and topology optimization of truss structures. To meet this aim, Integrated Particle Swarm Optimization (iPSO), Teaching and Learning Based Optimization (TLBO) and Drosophila Food-Search Optimization (DSO) algorithms are selected, respectively. Different properties like accuracy, convergence rate and complexity of the algorithms are investigated. The outcomes are provided via illustrative diagrams and tables. Based on the achieved results, DSO shows the most complexity level among the other algorithm while the iPSO and TLBO can outperform it on both accuracy and convergence rate. Consequently, iPSO presents a higher accuracy level on finding optimal solutions and TLBO with the lowest standard deviation value through the process shows the highest level of stability on finding optimal solutions.

Enhanced quadratic approximation integrated with butterfly optimization: a new search algorithm tested on structural and mathematical problems

“…have better objective value). So, most of the time such an approach for local search does not provide a promising movement (Mortazavi, 2019a(Mortazavi, , 2019c. To tackle this problem, the current phase is replaced with the quadratic neighborhood search around the weighted particle scheme.…”

“…Despite of effective global search strategy of BOA, the local search of this method is not enough to handle the complex search domains with high number of local optima or engineering problems with different (convex or non-convex) boundaries (Mortazavi, 2019a(Mortazavi, , 2019c. To mitigate this shortcoming, the global search strategy of BOA is combined with EQA search scheme.…”

“…in which, indicates the problem's penalized objective function and ( ) is the value of non-penalized objective function and ( ) returns violation of ith constraint. To provide more adaptive scheme, 1 and 2 as the tuning terms are respectively taken as 1 and 1.5 at the start of the process while linearly increased up to 6 (Mortazavi, 2019c). Initially Two mechanical cases and finally a structural problem are investigated in this section.…”

“…butterflies): tracing of the best agent of the colony for global search and pairwise interacting of agents for local search. Current authors test the performance of the BOA on various constrained and unconstrained problems (Mortazavi, 2019a(Mortazavi, , 2019c and the results indicate that the BOA works considerable well on solving the unconstrained cases with smooth domains. However, for the problems with complex search domain (i.e.…”

Optimizing base isolation system parameters using a fuzzy reinforced butterfly optimization: A case study of the 2023 Kahramanmaras earthquake sequence

“…Ant Colony Optimization (ACO) is a method that models the food search behavior of ants (Dorigo and Blum, 2005); The Differential Evolution algorithm (DE) applies evaluation rules to different species (Storn and Price, 1997); The Gravitational Search Algorithm (GSA) replicates the interactions between masses (Rashedi et al, 2009); The Hunting Search Algorithm (HSA) gains insight from the hunting tactics employed by animals (Oftadeh et al, 2010); The Drosophila Food-search Algorithm (DFO) utilizes the survival instincts of a particular insect to establish a search pattern (Das and Singh, 2014); Ions Motion Algorithm (IMA) takes inspiration from the movements and interactions of positively and negatively charged ions (Javidy et al, 2015); The Teaching and Learning Based Optimization (TLBO) algorithm mimics the procedure of acquiring knowledge in a teaching space (Rao et al, 2011); Butterfly Optimization Algorithm (BOA) is based on the breeding and food search behavior of butterflies (Arora and Singh, 2019); Search and Rescue Optimization Algorithm (SRA) is designed to simulate search and rescue operations, including the assignment of tasks and responsibilities (Shabani et al, 2019). The findings from various studies indicate that the efficacy of these methods could vary depending on the type of problem being addressed (Mortazavi and Moloodpoor, 2021a; Mortazavi et al, 2018). The primary cause of this issue may be attributed to the fact that these methods utilize a pre-defined and constant search pattern(s) that makes them unable to adapt themselves to the prevailing conditions within the current search domain.…”