Optimal solution of engineering design problems through differential gradient evolution plus algorithm: a hybrid approach

Abstract

It is very necessary and applicable to optimize all disciplines. In practical engineering problems the optimization has been a significant component. This article presents the hybrid approach named as differential gradient evolution plus (DGE+) algorithm which is the combination of differential evolution (DE) algorithm and gradient evolution (GE) algorithm. DE was used to diversify and GE was used for intensification with a perfect equilibrium between exploration and exploitation with an improvised distribution of dynamic probability and offers a new shake-off approach to prevent premature convergence to local optimum. To describe the success, the proposed algorithm is compared to modern meta-heuristics. To see the accuracy, robustness, and reliability of DGE+ it has been implemented on eight complex practical engineering problems named as: pressure vessel, belleville spring, tension/compression spring, three-bar truss, welded beam, speed reducer, gear train and rolling element bearing design problem, the results revealed that DGE+ algorithm can deliver highly efficient, competitive and promising results.

Complementary transient thermal models and metaheuristics to simultaneously identify linearly temperature-dependent thermal properties of austenitic stainless steels

“…Computing sensitivity derivatives is laborious when multidimensional, nonlinear, multimodal, discontinuous, and transient cases are considered. Gradient methods are not free from the 'local minima trap' problem, which means searching for a local optimum solution instead of a global optimum one [11]. They can exploit a limited search region around a given starting value, but they tend to fail when the starting value is far from the global solution [12].…”