Surrogate-assisted multi-objective optimization of the dynamic response of a freight wagon fitted with three-piece bogies

• Published in: International journal of rail transportation (Online) Vol. 9; no. 3; pp. 290 - 309
• Main Authors: Pandey, Manish, Regis, Rommel G., Datta, Rituparna, Bhattacharya, Bishakh
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 04.05.2021; Taylor & Francis Ltd
• Subjects: bolster suspension parameters; Dynamic response; evolutionary algorithm; Freight cars; Genetic algorithms; multi-objective optimization; Multiple objective analysis; Optimization; Pareto optimization; Radial basis function; rail vehicle dynamic simulation; Sorting algorithms; surrogate model; Three-piece bogie; Undercarriages; weighted optimization
• ISSN: 2324-8378; 2324-8386
• DOI: 10.1080/23248378.2020.1792808

The multi-objective optimization problem of the dynamic response of a freight wagon fitted with three-piece bogies is a challenging task due to interdependencies between the decision variables, conflicts between the objective functions and computationally expensive rail vehicle dynamic simulations. In this article, a novel approach of multi-objective optimization of the dynamic performance of a freight wagon fitted with three-piece bogies is presented. Surrogate modelling using radial basis function (RBF) ensembles have been used to model the five objective functions representing the dynamic performance of the freight wagon. Elitist Non-dominated Sorting Genetic Algorithm (NSGA-II) has been applied to optimize the objective functions formed using the surrogate models. Finally, to choose specific solutions from the identified Pareto fronts, the minimization of a weighted combination of the surrogate objective functions is solved for six combinations of weights. The weighted optimized solutions show significant improvements over the existing solution.