1. IntroductionThe truss has been one of the most used designs throughout structural engineering history. Using such a design is advan-tageous in that it is simple and inexpensive to construct, modify, and maintain, especially in difficult-to-access areas. In theliterature, the design optimisation of trusses has seen a resurgence of interest recently. These design problems have usually involved minimising structural weight or cost while maintaining safety. These problems may have one or more design objectives such as dynamic stiffness (or natural frequency), compliance, frequency response function, force transmissibility, andbuckling factor [29].The optimisers used in truss design problems can be categorised as gradient-based methods (or local search), and meta-heuristics (MHs),more commonly known as evolutionary algorithms (EAs). Previous studies using gradient-based optimiserssuch as sequential linear programming [16,18], feasible direction method [42], and sequential quadratic programming[32,41], for truss design have been conducted. Some well-known EAs including genetic algorithms, have been implementedto solve structural optimisation problems [17,21,22,29,34,44,45]. Gradient-based methods have faster convergence rates andare more consistent in finding a local optimum, however, they require continuous design variables, and accurate derivativecalculations of design functions. This makes them difficult to use for most cases of structural optimisation, as inaccurateestimation of function derivatives may lead their search procedures to improper solutions. The EAs, on the other hand, haveemerged as strong candidates for this design task in the last few decades [10]. Compared to their gradient-based counter-parts, they are easier to use, more robust, and capable of dealing with all kinds of design variables since they do not requirefunction derivatives for searching. Moreover, their most outstanding feature is that the multiobjective versions of EAs cansearch for Pareto optimal sets within one optimisation run [10,15,49,55]. Nevertheless, they inevitably have slower
đang được dịch, vui lòng đợi..
