Optimum Design of Larg Scale Truss Structure by Levy Flight-Based Improved Coot Optimization Algorithm Considering Discrete Variables

One of the main concerns of the structural engineers is the necessity of using discrete variables in optimum design of structural systems in which a predefined set of design sections are considered to be used during the optimization process. Most of the recently developed metaheuristic algorithms are capable of dealing with continuous design variables; however, the capability of these algorithms in dealing with a problem of discrete variables is one of the major concerns in optimization field. For this purpose, the applicability of the recently developed Coot optimization algorithm is evaluated in this paper in dealing with optimum design of 4 truss structures with 10, 25, 52 and 72 bars in which the design sections are predefined through discrete variables. Meanwhile, the improved Coot optimization algorithm is also proposed to increase the overall performance of the standard Coot algorithm in which the random movements in the main loop of the algorithm are replaced by Levy flight which denotes on a stochastic process with step length determined by levy distribution. The capability of the standard Coot and the proposed improved Coot optimization algorithms is investigated regarding the considered truss problems while the best and statistical results demonstrate the overall capability of the improved algorithm in dealing with discrete design scheme in truss optimization problems.