An Enhanced Symbiotic Organism Search Algorithm (ESOS) for the Sizing Design of Pin Connected Structures
- PDF / 2,427,995 Bytes
- 26 Pages / 595.276 x 790.866 pts Page_size
- 77 Downloads / 189 Views
RESEARCH PAPER
An Enhanced Symbiotic Organism Search Algorithm (ESOS) for the Sizing Design of Pin Connected Structures Mohammad H. Makiabadi1 · Mahmoud R. Maheri1 Received: 15 January 2020 / Accepted: 24 September 2020 © Shiraz University 2020
Abstract The symbiotic organism search (SOS) algorithm is a strong metaheuristic search engine with a great explorative capability which enables it to search for global optima efficiently. Nevertheless, its exploitive ability is less profound leading to a relatively low rate of convergence. In the current paper, an enhanced symbiotic organism search (ESOS) algorithm with the aim of improving the exploitive capability of the SOS algorithm is presented. The enhancements are done in the commensalism and mutualism phases of the algorithm. In the current study, a multi-ecosystem mechanism is also employed. The presented ESOS algorithm is used along with two different constraint handling methods, namely mapping strategy (MS) and penalty function (PF). This leads to two variants of the ESOS, termed as ESOS-MS and ESOS-PF. The two presented ESOS variants are utilized to optimize the weight of four benchmark truss structures. The results are compared with the optimal designs reported in the literature. The results demonstrate that the proposed ESOS variants not only can find better solutions compared to the other algorithms, but also are faster in doing so. Moreover, of the two presented ESOS variants, the ESOS-MS variant performs much better than the ESOS-PF variant. Keywords Symbiotic organism search algorithm (SOS) · Enhanced SOS · Weight optimization
1 Introduction Optimal design of truss structures has been a major area of research in the field of engineering optimization in the last three decades. Generally, optimum design of a structure aims at minimizing the total cost of that structure by reducing the sizes of structural elements, with a view to maintaining their adequate load carrying capacity. The presence of a high number of design variables, large size of the search space and the need to control a great number of design constraints, makes such a task computationally expensive. Gradientbased optimizers can be used in problems with continuous design variables (Allwood and Chung 1984; Fleury 1980). Though in general, these methods can obtain solutions faster and with a higher degree of accuracy compared to stochastic approaches in fulfilling local search tasks, the variables and cost function of the generators need to be continuous and a good starting point is imperative for these methods to * Mahmoud R. Maheri [email protected] 1
Department of Civil Engineering, Shiraz University, Shiraz, Iran
be successful. In this regard, the works of Schmit (1960), Vanderplaats and Moses (1973), Schmit and Farshi (1974), Schmit and Miura (1976), Arora and Haug (1976), Harless (1980), Belegundu and Arora (1985), Adeli and Kamal (1986), Ringertz (1985) and Joseph (1987) among others can be mentioned. Stochastic search methods based on natural phenomena, on the other hand, hav
Data Loading...
