Multiobjective Optimization of Auxetic Aerospace Cellular Structures Using a Python Implementation of Direct Multisearch

Abstract

Multiobjective optimization problems are common in aerospace structural design, where improving one performance criterion may degrade another. In this context, auxetic cellular structures are relevant due to their potential for lightweight design, energy absorption, and deformation control. Since their effective response depends strongly on geometry and often requires numerical simulation, these structures are suitable candidates for derivative-free, geometry-based optimization. This work presents a Python implementation of the Direct Multisearch (DMS) algorithm, developed from an existing MATLAB reference code and assessed through MATLAB-Python equivalence tests. The resulting implementation was coupled with an external finite-element and homogenization pipeline to optimize a parametrized re-entrant honeycomb unit cell. The final results produced a structured Pareto front of trade-off solutions between auxetic behaviour and stiffness-related performance, demonstrating the suitability of the Python DMS implementation for aerospace-oriented, simulation-based multiobjective optimization.