Evolving Novel Gene Regulatory Networks for Structural Engineering Designs

TY - JOUR

T1 - Evolving Novel Gene Regulatory Networks for Structural Engineering Designs

AU - Dubey, Rahul

AU - Hickinbotham, Simon John

AU - Colligan, Andrew

AU - Friel, Imelda

AU - Buchanan Berumen, Edgar

AU - Price, Mark

AU - Tyrrell, Andy

N1 - © 2024 Massachusetts Institute of Technology. This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

PY - 2024/8/15

Y1 - 2024/8/15

N2 - Engineering design optimization poses a significant challenge, usually requiring human expertise to discover superior solutions. While various search techniques have been employed to generate diverse designs, their effectiveness is often limited by problem-specific parameter tuning, making them less generalizable and scalable. This paper introduces a framework inspired by evolutionary and developmental (Evo-Devo) concepts, aiming to automate the evolution of structural engineering designs. In biological systems, Evo-Devo governs the growth of single-cell organisms into multi-cellular organisms through the use of Gene Regulatory Networks (GRNs). GRNs are inherently complex andhighly nonlinear, and this paper explores the use of neural networks and genetic programming as artificial representations of GRNs to emulate such behaviors. In order to evolve a wide range of Pareto fronts for artificial GRNs, this paper introduces a new technique, a real-value encoded neuro-evolutionary method termed “real-encoded NEAT” (RNEAT). The performance of RNEAT is compared with two well-known evolutionary search techniques across different 2D and 3D problems. The experimental results demonstrate two key findings: Firstly, the proposed framework effectively generates a population of GRNs that can produce diverse structures for both 2D and 3D problems. Secondly, the proposed RNEAT algorithm outperforms its competitors on more than 50% of the problems examined.

AB - Engineering design optimization poses a significant challenge, usually requiring human expertise to discover superior solutions. While various search techniques have been employed to generate diverse designs, their effectiveness is often limited by problem-specific parameter tuning, making them less generalizable and scalable. This paper introduces a framework inspired by evolutionary and developmental (Evo-Devo) concepts, aiming to automate the evolution of structural engineering designs. In biological systems, Evo-Devo governs the growth of single-cell organisms into multi-cellular organisms through the use of Gene Regulatory Networks (GRNs). GRNs are inherently complex andhighly nonlinear, and this paper explores the use of neural networks and genetic programming as artificial representations of GRNs to emulate such behaviors. In order to evolve a wide range of Pareto fronts for artificial GRNs, this paper introduces a new technique, a real-value encoded neuro-evolutionary method termed “real-encoded NEAT” (RNEAT). The performance of RNEAT is compared with two well-known evolutionary search techniques across different 2D and 3D problems. The experimental results demonstrate two key findings: Firstly, the proposed framework effectively generates a population of GRNs that can produce diverse structures for both 2D and 3D problems. Secondly, the proposed RNEAT algorithm outperforms its competitors on more than 50% of the problems examined.

KW - Evolutionary Search

KW - Gene Regulatory Networks

KW - NEAT

KW - CGP

KW - Design Optimization

U2 - 10.1162/artl_a_00448

DO - 10.1162/artl_a_00448

M3 - Article

SN - 1064-5462

JO - Artificial Life

JF - Artificial Life

ER -