A sequential surrogate-based multiobjective optimization method: effect of initial data set
- PDF / 3,386,457 Bytes
- 24 Pages / 595.276 x 790.866 pts Page_size
- 1 Downloads / 169 Views
(0123456789().,-volV)(0123456789(). ,- volV)
A sequential surrogate-based multiobjective optimization method: effect of initial data set Maria Guadalupe Villarreal-Marroquin1 Jose M. Castro4
•
Jose Daniel Mosquera-Artamonov2
•
Celso E. Cruz3
•
Ó Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract Process optimization based on high-fidelity computer simulations or real experimentation is commonly expensive. Therefore, surrogate models are frequently used to reduce the computational or experimental cost. However, surrogate models need to achieve a maximum accuracy with a limited number of sampled points. Sequential sampling is a procedure in which sequentially surrogates are fitted and each surrogate defines the points that need to be sampled and used to fit the next model. For optimization purposes, points are sampled on regions of high potential for the optimal solutions. In this work, we first compared the effect of using different initial sets of points (experimental designs) in a sequential surrogatebased multiobjective optimization method. The optimization method is tested on five benchmark problems and the performance is quantified based on the total number of function evaluations and the quality of the final Pareto Front. Then an industrial applications on titanium welding is presented to show the use of the method. The case study is based on real experimental data. Keywords Sequential design optimization Surrogate models Multiobjective optimization Design of experiments Manufacturing
1 Introduction For manufacturing businesses to be successful in the global market, they must strive to deliver high quality products at the lowest possible cost. One approach to select the processing conditions to achieve these goals is to run experiments on the manufacturing floor. Such experimentation is & Jose Daniel Mosquera-Artamonov [email protected] Maria Guadalupe Villarreal-Marroquin [email protected] 1
Modeling Optimization and Computing Technology SAS de CV, Monterrey, NL, Mexico
2
Posgrado en Ingenieria de Sistemas, Universidad Autonoma de Nuevo Leon, Monterrey, NL, Mexico
3
Gerencia de Manufactura y Procesos Especiales, Centro de Ingenieria y Desarrollo Industrial, Estado de Mexico, Edomex, Mexico
4
Integrated Systems Engineering Department, The Ohio State University, Columbus, OH, USA
usually costly and requires considerable amount of time and effort, which may not be feasible during production [1]. Alternatively, companies use advance computer simulations to represent their processes. Such computer simulations along side with optimization methods are used to identify the values of the processing conditions (variables) that optimize the relevant performance measures (objectives). Joining simulation and optimization in a single framework for defining the best possible process parameters is an actual need in current engineering practice [2–6]. However, a major difficulty of optimizing engineering problems based on simulations is that each function evaluati
Data Loading...
