Showing 6 results for Pareto
A. Khalkhali, S. Samareh Mousavi,
Volume 2, Issue 3 (7-2012)
Abstract
In order to reduce both the weight of vehicles and the damage of occupants in a crash event simultaneously, it is necessary to perform a multi-objective optimization of the automotive energy absorbing components. In this paper, axial impact crushing behavior of the aluminum foam-filled thin-walled tubes are studied by the finite element method using commercial software ABAQUS. Comparison of the present simulation results with the results of the experiments reported in the previous works indicated the validity of the numerical analyses. A meta-model based on the feed-forward artificial neural networks are then obtained for modeling of both the absorbed energy (E) and the peak crushing force (Fmax) with respect to design variables using those data obtained from the finite element modeling. Using such obtained neural network models, a modified multi-objective GA is used for the Pareto-based optimization of the aluminum foam-filled thinwalled tubes considering three conflicting objectives such as energy absorption, weight of structure, and peak crushing force.
J. Reza Pour, B. Bahrami Joo, A. Jamali, N. Nariman-Zadeh,
Volume 4, Issue 4 (12-2014)
Abstract
Robust control design of vehicles addresses the effect of uncertainties on the vehicle’s performance. In present study, the robust optimal multi-objective controller design on a non-linear full vehicle dynamic model with 8-degrees of freedom having parameter with probabilistic uncertainty considering two simultaneous conflicting objective functions has been made to prevent the rollover. The objective functions that have been simultaneously considered in this work are, namely, mean of control effort (MCE) and variance of control effort (VCE).The nonlinear control scheme based on sliding mode has been investigated so that applied braking torques on the four wheels are adopted as actuators. It is tried to achieve optimum and robust design against uncertainties existing in reality with including probabilistic analysis through a Monte Carlo simulation (MCS) approach in multi-objective optimization using the genetic algorithms. Finally, the comparison between the results of deterministic and probabilistic design has been presented. The comparison of the obtained robust results with those of deterministic approach shows the superiority robustness of probabilistic method.
M. Salehpour, A. Jamali, A. Bagheri, N. Nariman-Zadeh,
Volume 7, Issue 4 (12-2017)
Abstract
In this paper a new type of multi-objective differential evolution employing dynamically tunable mutation factor is used to optimally design non-linear vehicle model. In this way, non-dominated sorting algorithm with crowding distance criterion are combined to fuziified mutation differential evolution to construct multi-objective algorithm to solve the problem. In order to achieve fuzzified mutation factor, two inputs as generation number and population diversity and one output as the mutation factor are used in the fuzzy inference system. The objective functions optimized simultaneously are namely, vertical acceleration of sprung mass, relative displacement between sprung mass and unsprung mass and control force. Optimization processes have been done in two bi- and three objective areas. Comparison of the obtained results with those in the literature has shown the superiority of the proposed method of this work. Further, it has been shown that the results of 3-objective optimization include those of bi-objective one, and therefore it gives more optimum options to the designer
Mohammad Salehpour, Ali Jamali, Ahmad Bagheri, Nader N. Nariman-Zadeh,
Volume 8, Issue 4 (12-2018)
Abstract
In this paper, a new version of multi-objective differential evolution with dynamically adaptable mutation factor is used for Pareto optimization of a 5-degree of freedom vehicle vibration model excited by non-stationary random road profile. In this way, non-dominated sorting algorithm and crowding distance criterion have been combined to differential evolution with fuzzified mutation in order to achieve multi-objective meta-heuristic algorithm. To dynamically tune the mutation factor, two parameters, named, number of generation and population diversity are considered as inputs and, one parameter, named, the mutation factor as output of the fuzzy logic inference system. Conflicting objective functions that have been observed to be optimally designed simultaneously are, namely, vertical seat acceleration, vertical forward tire velocity, vertical rear tire velocity, relative displacement between sprung mass and forward tire and relative displacement between sprung mass and rear tire. Furthermore, different pairs of these objective functions have also been chosen for bi-objective optimization processes. The comparison of the obtained results with those in the literature unveils the superiority of the results of this work. It is displayed that the results of 5-objective optimization subsume those of bi-objective optimization and, consequently, this achievement can offer more optimal choices to designers.
Mrs Ghazal Etesami, Dr Mohammad Ebrahim Felezi, Prof Nader Nariman-Zadeh,
Volume 9, Issue 3 (9-2019)
Abstract
The present paper aims to improve the dynamical balancing of a slider-crank mechanism. This mechanism has been widely used in internal combustion engines, especially vehicle engines; hence, its dynamical balancing is important significantly. To have a full balance mechanism, the shaking forces and shaking moment of foundations should be eliminated completely. However, this elimination is usually impossible. Hence, in the current study, a multi-objective optimization is carried out to maintain the optimal balance of mechanism. The vertical and horizontal components of shaking forces and shaking moment are considered as objective functions. Also, the design variables are included the mass, the moment of inertia and the mass center location of mechanism links. The length of mechanism links is also considered constant for achieving a fixed slider course. The four-objective optimization is applied using a differential evolution algorithm. The optimization results are presented in Pareto diagrams as suitable tools for selecting a mechanism with desired characteristics according to the importance of each objective function. The optimal mechanism is finally introduced by the mapping method. The comparison of optimized mechanisms and the original one indicates a significant reduction of shaking forces and shaking moment as well as the reduction of energy consumption.
Dr. Mohammad Salehpour, Dr. Ahmad Bagheri,
Volume 11, Issue 3 (9-2021)
Abstract
In this study, a multi-objective differential evolution with fuzzy inference-based dynamic adaptable mutation factor with hybrid usage of non-dominated sorting and crowding distance (MODE-FM) is utilized for Pareto optimization of a 5-degree of freedom nonlinear vehicle vibration model considering the five conflicting functions simultaneously, under different road inputs. The significant conflicting objective functions that have been observed here are, namely, vertical seat acceleration, vertical forward tire velocity, vertical rear tire velocity, relative displacement between sprung mass and forward tire and relative displacement between sprung mass and rear tire. Different road inputs are, namely, double-bump, stationary random road and non-stationary random road. It is exhibited that the optimum solutions of 5-objective optimization contain those of 2-objective optimization and, as a result, this important matter creates more options for optimal design of nonlinear vehicle vibration model.