A hybrid operator-based multifactorial evolutionary algorithm for inverse-engineering design of soft network materials

Abstract

Soft network materials (SNMs) are a class of network materials with periodic thin curved filaments in lattice patterns, offering excellent physical properties of high stretchability, highly tunable mechanical properties, and lightweight. Several theoretical models have been established for the inverse-engineering design of SNMs for reproducing the nonlinear J-shaped stress–strain curves for mechanical bionics. However, the existing design models required the semi-subjective decision in the choice of key geometric parameters, critical strain, etc., under the limited material libraries. In this study, we developed a brand-new hybrid operators-based multifactorial evolutionary algorithm (HOMFEA) for the inverse-engineering design of SNMs to reproduce the J-shaped stress–strain curves of various target soft biological tissues. The core algorithms in HOMFEA include a hybrid operator proposed based on the memetic algorithm theory, a vertical cultural transmission, and population formation coupled with modeling methods for SNMs. The HOMFEA-based design strategy allows more accurate reproduction of target stress–strain curves compared to the conventional EA-based framework and a phenomenological-based mechanical model. The high-performance computing capability suggests that it can serve as a reliable tool for more inverse-engineering design scenarios.