Library design using genetic algorithms for catalyst discovery and optimization

Abstract

This study reports a detailed investigation of catalyst library design by genetic algorithm (GA). A methodology for assessing GA configurations is described. Operators, which promote the optimization speed while being robust to noise and outliers, are revealed through statistical studies. The genetic algorithms were implemented in GA platform software called OptiCat, which enables the construction of custom-made workflows using a tool box of operators. Two separate studies were carried out (i) on a virtual benchmark and (ii) on real surface response which is derived from HT screening. Additionally, we report a methodology to model a complex surface response by binning the search space in small zones that are then independently modeled by linear regression. In contrast to artificial neural networks, this approach allows one to obtain an explicit model in an analogical form that can be further used in Excel or entered in OptiCat to perform simulations. While speeding the implementation of a hybrid algorithm combining a GA with a knowledge-based extraction engine is described, while speeding up the optimization process by means of virtual prescreening the hybrid GA enables one to open the “black-box” by providing knowledge as a set of association rules.

Combinatorial and High‐Throughput Materials Science

“…The authors also presented a database linked to a powerful algorithm for the iterative discovery and optimization of catalytic samples. A more theoretical approach to data mining has been published by Clerc et al [110] The authors show how optimization problems can be solved by hybridizing a classical GA with a KD system (for example, a learning process using k nearest neighbors) that extracts information from a database. A schematic flow chart is shown in Figure 8.…”

“…[37,[48][49][50][51] Another evolutionary optimization approach combined with high-throughput synthesis and screening has been applied to the discovery of new catalysts for the oxidative dehydrogenation of ethane to ethylene. [52] Mirodatos and co-workers [53,54] used genetic algorithms in several ways: they developed a GA platform software called OptiCat that enables the user to construct custom-made workflows. They tested their approaches on a virtual benchmark test and on experimental response surfaces obtained from HT screening.…”

Selective Hydrogen Oxidation Catalysts via Genetic Algorithms

“…Baerns et al have studied the oxidative dehydrogenation of propane, using gaseous oxygen, [22][23][24] total propane oxidation, [25] and the production of hydrocyanic acid. [26] Yamada et al have investigated methanol synthesis, [27][28][29] and others have performed studies on (selective) oxidation, [30][31][32][33] reduction, [34] methane reforming, [35] and isomerisation. [36] Most of the catalysts used in these studies are mixed oxides containing four to five different metals.…”

Kombinatorische und Hochdurchsatz‐Techniken in der Materialforschung

“…[37,[48][49][50][51] Ein weiterer Ansatz zur evolutionären Optimierung kombiniert mit HT-Synthese und -Screening wurde bei der Entdeckung neuer Katalysatoren für die oxidative Dehydrierung von Ethan zu Ethylen angewendet. [52] Mirodatos et al [53,54] verwendeten genetische Algorithmen auf verschiedene Arten: Sie entwickelten eine GA-Softwareplattform mit dem Namen OptiCat, die dem Benutzer die Erstellung von maßgeschneiderten Arbeitsabläufen ermög-licht. Sie überprüften ihre Ansätze sowohl an virtuellen Benchmark-Tests als auch an experimentell gefundenen Antwortflächen aus dem HT-Screening.…”