Introduction
Materials science as a field is being disrupted with advances in machine learning and automation,\cite{Lookman_2015,Liu_2017,Correa_Baena_2018} where high-throughput experimentation (HTE) capabilities accelerate discovery of materials in more complex search spaces. Users not only save time on experimentation by virtue of automated workflows with faster processing, but also leveraging on equipment with larger batches of experiments to increase throughput and thus minimise experimental time. \cite{Zhang_2009,Mennen_2019} There have been many successful applications of HTE, particularly in the single objective problem space alongside machine learning-assisted optimisation strategies. \cite{buonassisi,Sun_2019,Burger_2020,Dave_2020,Gongora_2020,Langner_2020,Li_2020,Shimizu_2020,Wang_2020,Bash_2022,Deneault_2021,Mekki_Berrada_2020} However, many real-world problems are more complex, specifically with multiple conflicting properties to be optimized, for example: strength vs ductility in metal alloys, \cite{Li_2016} device thickness vs fill factor in photovoltaics, \cite{Ramirez_2018} or selectivity vs current density in catalysts. \cite{Ren_2019} In addition, such problems may include constraints that restrict the space of feasible solution. This motivates the need for multi-objective optimisation strategies with constraint handling capabilities to be integrated in HTE setups. \cite{Alsharif_2020,Bash_2020,Grizou_2020,Abdel_Latif_2021} The first step could consist of formulating complex material science problems as constrained multi-objective optimisation problems (CMOPs).