APPLICATION OF HYBRID RANDOM SEARCH METHOD TO OPTIMISATION OF ENGINEERING SYSTEMS’ PARAMETERS

This paper presents a modification of the Luus-Jaakola global optimization method, which belongs to the class of metaheuristic algorithms. A hybrid method is suggested, using a combination of random search methods: Luus-Jaakola method, adaptive random search method and best trial method. The obtained method is applied to the optimization of parameters of different engineering systems. This class of problems appears during the design of aerospace and aeronautical structures; its goal is the cost or weight minimization of the construction. These problems belong to the class of constrained global optimization problems, where the level surface of the objective function has uneven relief and there is a large number of variables. This means that the classical optimization methods prove to be inefficient and these problems should be solved using metaheuristic optimization methods, which provide sufficient accuracy at reasonable operating time. In this paper, the constrained global optimization problem is solved using the penalty method. Thus, the problem of exterior penalty function optimization is considered, where the penalty coefficients are chosen in such a way as to avoid the violation of the constraints. Two applied problems are considered in the paper: the determination of the high-pressure vessel parameters and the anti rattle spring parameters determination. Using the suggested algorithm, a software complex was developed, which allows us to solve engineering optimization problems. The results obtained using the suggested methods were compared with the results obtained using the non-modified Luus-Jaakola method in order to demonstrate the efficiency of the suggested hybrid random search method.