TY - JOUR

T1 - EngiO – Object-oriented framework for engineering optimization

AU - Berger, Ricarda

AU - Bruns, Marlene

AU - Ehrmann, Andreas

AU - Haldar, Ayan

AU - Häfele, Jan

AU - Hofmeister, Benedikt

AU - Hübler, Clemens

AU - Rolfes, Raimund

PY - 2021/3

Y1 - 2021/3

N2 - This paper presents an object-oriented optimization framework for engineering optimization using the Matlab programming syntax. The novelty of the developed framework lies in its approach to remove redundancies by providing an interface for central routines of the optimization processes. Object-oriented programming is used to implement an abstract optimizer class, which controls the optimization process and provides unified interfaces for optimization. The software architecture reduces code complexity and allows concise implementation of derivative-free algorithms and optimization tasks. Therefore, it is ideal for students and researchers to improve and develop algorithms or to solve optimization problems. The proposed framework features parallel evaluation of objective functions and handles global optimization including single-objective, multi-objective and constrained problems. In this paper, the versatility of the framework is demonstrated by using analytic test problems as well as practical engineering problems.

AB - This paper presents an object-oriented optimization framework for engineering optimization using the Matlab programming syntax. The novelty of the developed framework lies in its approach to remove redundancies by providing an interface for central routines of the optimization processes. Object-oriented programming is used to implement an abstract optimizer class, which controls the optimization process and provides unified interfaces for optimization. The software architecture reduces code complexity and allows concise implementation of derivative-free algorithms and optimization tasks. Therefore, it is ideal for students and researchers to improve and develop algorithms or to solve optimization problems. The proposed framework features parallel evaluation of objective functions and handles global optimization including single-objective, multi-objective and constrained problems. In this paper, the versatility of the framework is demonstrated by using analytic test problems as well as practical engineering problems.

KW - Engineering optimization

KW - Global derivative-free optimization

KW - Optimization framework

UR - http://www.scopus.com/inward/record.url?scp=85099688044&partnerID=8YFLogxK

U2 - 10.1016/j.advengsoft.2020.102959

DO - 10.1016/j.advengsoft.2020.102959

M3 - Article

AN - SCOPUS:85099688044

VL - 153

JO - Advances in engineering software

JF - Advances in engineering software

SN - 0965-9978

M1 - 102959

ER -