Parameter control for polyglot evolutionary algorithms

The goal of this project is to build a cloud-ready solution for the manipulation of evolutionary algorithms.

Such infrastructure should aim at promoting the scalability of the evaluation systems through parallelisation, thereby speeding up the computation of such optimisation routines, as supported by research evidence coming from the work of Dr. Pasquale Salza.

Furthermore, the final deliverable is expected to support the definition of evolutionary algorithms in multiple programming languages, as great is the popularity of such optimisation strategies for these to have been implemented and utilised across a number of languages and indeed libraries.

Finally, in order to aid the human supervision and control over evolutionary algorithms being run in such cloud infrastructure, a user interface should be developed. This ought to allow the user to not only monitor the progress of the optimisation strategy, but also control its parameters, which may enable control over the evolutionary process.

This project will be supervised by Prof. Dr. Harald C. Gall, Dr. Pasquale Salza, and Marco Edoardo Palma.

Please reach out to Marco Edoardo Palma or Dr. Pasquale Salza if you have further questions about the project.

1. Karafotias, G., Hoogendoorn, M. and Eiben, Á.E., 2014. Parameter control in evolutionary algorithms: Trends and challenges. IEEE Transactions on Evolutionary Computation, 19(2), pp.167-187.Proksch et al., “Enriching In-IDE Process Information with Fine-Grained Source Code History”, SANER’17.
2. Salza, P. and Ferrucci, F., 2019. Speed up genetic algorithms in the cloud using software containers. Future Generation Computer Systems, 92, pp.276-289.
3. Salza, P., Hemberg, E., Ferrucci, F. and O'Reilly, U.M., 2017, July. Towards evolutionary machine learning comparison, competition, and collaboration with a multi-cloud platform. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (pp. 1263-1270).