Comparative Study of Particle Swarm Optimization Algorithms in Solving Size, Topology, and Shape Optimization

Abstract

This paper focuses on optimizing truss structures while propose best PSO variants. Truss optimization is one way to make the design efficient. There are three types of optimization, size optimization, shape optimization, and topology optimization. By combining size, shape and topology optimization, we can obtain the most efficient structure. Metaheuristics have the ability to solve this problem. Particle swarm optimization (PSO) is metaheuristic algorithm which is frequently used to solve many optimization problems. PSO mimics the behavior of flocking birds looking for food. But PSO has three parameters that can interfere with its performance, so this algorithm is not adaptive to diverse problems. Many PSO variants have been introduced to solve this problem, including linearly decreasing inertia weight particles swarm optimization (LDWPSO) and bare bones particles swarm optimization (BBPSO). The metaheuristic method is used to find the solution, while DSM s used to analyze the structure. A 10-bar truss structure and a 39-bar truss structure are considered as case studies. The result indicates that BBPSO beat other two algorithms in terms of best result, consistency, and convergence behaviour in both cases. LDWPSO took second place for the three categories, leaving PSO as the worst algorithm that tested.

A generative design-to-BIM workflow for minimum weight plane truss design

“…The recent studies have shown that GA (Dede, Bekirolu, & Ayvaz, 2011), TLBO (Dede, 2014), DE (Vu, 2015), Simulated Annealing (SA) (Millan-Paramo, 2018), Ant Lion Optimizer (ALO) (Salar & Dizangian, 2019), Crow Search Algorithm (CSA) (Javidi, Salajegheh, & Salajegheh, 2019) and Artificial Coronary Circulation System (ACCS) algorithm (Kooshkbaghi & Kaveh, 2020) have been used for optimal dimensioning of plane truss structures with minimum weight. Also, optimization of layout and geometry/shape with topology besides dimensioning was done using GA (Chen, Shui, & Huang, 2017;Neeraja, Kamireddy, Kumar, & Reddy, 2017;Rahami, Kaveh, & Gholipour, 2008), Firefly Algorithm (FA) (Miguel, Lopez, & Miguel, 2013), PSO (Harsono et al, 2020;Mortazavi & Toğan, 2016), JA (Degertekin, Lamberti, & Ugur, 2019). In addition, in recent studies, Total Potential Optimization (S. Ç.…”