This repository contains the source code for the “Thompson sampling efficient multiobjective optimization” (TSEMO) algorithm outlined in (Bradford et al., 2018). The algorithm is written to optimize expensive, black-box functions involving multiple conflicting criteria by employing Gaussian process surrogates. It is often able to determine a good approximation of the true Pareto front in signficantly less iterations than genetic algorithms. To cite TSEMO use (Bradford et al., 2018).
To use TSEMO download all files contained in the repository and run the algorithm on the required test-function as shown in the example matlab file TSEMO_Example. To use the algorithm on your own functions simply copy the same format as the functions shown in the test-function folder. The algorithm can be applied to any number of inputs and objectives.
The algorithm has been successfully applied to several expensive multiobjective optimization problems:
- Determination of optimal conditions of a fully-automated chemical reactor system trading-off yield and environmental factors (Schweidtmann et al., 2018) including multi-step reactions and separation processes (Clayton et al., 2020)
-
Optimization of a chemical process using a life-cycle assessment and cost simulation (Helmdach et al., 2018)
-
Solvent selection for asymmetric catalysis using molecular descriptors (Amar et al., 2019)
E. Bradford, A. M. Schweidtmann, and A. A. Lapkin, Efficient multiobjective optimization employing Gaussian processes, spectral sampling and a genetic algorithm, Journal of Global Optimization, vol. 71, no. 2, pp. 407–438, 2018.
A. M. Schweidtmann, A. D. Clayton, N. Holmes, E. Bradford, R. A. Bourne, and A. A. Lapkin, Machine learning meets continuous flow chemistry: Automated optimization towards the Pareto front of multiple objectives, Chemical Engineering Journal, vol. 352, pp. 277-282, 2018.
D. Helmdach, P. Yaseneva, K. P. Heer, A. M. Schweidtmann, and A. A. Lapkin, A Multiobjective Optimization Including Results of Life Cycle Assessment in Developing Biorenewables-Based Processes, ChemSusChem, vol. 10, no. 18, pp. 3632-3643, 2017.
Y. Amar, A. M. Schweidtmann, P. Deutsch, L. Cao, and A. A. Lapkin, Machine learning and molecular descriptors enable rational solvent selection in asymmetric catalysis, Chemical Science, vol. 10, no. 27, pp. 6697-6706, 2019.
A. Clayton, A. M. Schweidtmann, G. Clemens, J. Manson, C. Taylor, C. Nino, T. Chamberlain, N. Kapur, A. Blacker, A. A. Lapkin, R. Bourne Automated self-optimisation of multi-step reaction and separation processes using machine learning, Chemical Engineering Journal, vol. 384, 123340, 2020.
