Combined Size and Shape Optimization of Truss Structures Using the Heap-Based Optimizer (HBO) Algorithm

The simultaneous size and shape optimization of truss structures is a notoriously challenging optimization problem due to its NP-hard nature, highly nonlinear design constraints, and conflicting design variables. In this study, a novel human-based metaheuristic algorithm, called the Heap-based Optimizer (HBO), is presented for the configuration optimization of truss structures. Inspired by the Corporate Rank Hierarchy (CRH) concept, search agents are logically arranged into a hierarchy based on their fitness values. Solutions then compete with one another for promotions within the hierarchy by efficiently interacting with each other in a formal way. Overall, the proposed method is characterized by its robust exploration and ability to escape local optima. The effectiveness of the proposed technique was tested using five challenging truss benchmarks. The overall weight of the structure was minimized as an objective function, and the finite element procedure was coded in MATLAB. The selected variables for the problem included both member cross-sectional areas and nodal coordinates of the joints. Normal stresses, Euler buckling stresses, and nodal displacement limitations were moreover considered as design constraints. For statistical significance, the Friedman ranking test was implemented to extract clear winners from the comparison process. Results point to the clear superiority of HBO compared to all other state-of-the-art methods. Specifically, HBO designs were (on average) 7.3% lighter than those previously reported in literature. A comparison of computational speed furthermore demonstrated HBO’s rapid convergence capabilities, with an estimated 35.5% reduction in computational effort. Further research into HBO as an efficient structural optimizer is therefore highly recommended.