Within my PhD Work using multi objective optimization, I am aiming to determine more efficient ways to go about performing the optimization when highly expensive problems are present. While I have looked into surrogate modelling before there is a large variety of work present within the field. This summer school that focuses on the synergy between parallel optimization and surrogate based methodologies seemed like a good opportunity to consider new methods that could be relevant to my work.
Department email
There was a simple application process in which we needed to indicate our current background within the field as well as propose a problem that could be discussed within group work stage of the summer school.
By attending, I needed to ensure that I came to all sessions and fully participated with the group work.
The summer school lasted five days, during the first two days all the sessions focussed on presentations from researchers within the field who presented the state of the art methodologies in addition to their personal projects as well as future work. The third day began with the last presentation on the synergy between the previously presented topics, before beginning group work. This was a half day and finished before lunch followed by a social activity. During the final two days we worked in groups of up to eight on the proposed problems. We finish by giving a short presentation about how we went about tackling the problems and the final solutions we came up with.
I enjoyed getting to know others within the field as well as the challenge of applying multi-objective optimization methods to real world applications. The presentations gave a clear overview of the topics as well as some innovative applications.
The main challenges occurred during the group work. They mostly stemmed from the problem having not been fully presented from an optimization point of view. Due to this it was necessary to first redefine the problem so that it was within the right framework before it was possible to go about solving. While normally this would just be a minor issue, due to the time restraints present this caused us to be rushed. The other challenge that we were faced with was that within our group we ended up with two different points of view of how this problem should be solved. This lead to us implementing both methods so that we could compare and contrast the benefits of each.
Yes, the knowledge of different way to approach multi objective problems has been useful. While I have not currently yet had the chance to fully utilize the option of incorporating parallelization for HPC (high performance computing) in order to increase efficiency, it could well be beneficial in future work.
I would recommend that when going on a summer school it can be beneficial to ensure you have a basic understanding of the field before hand as even though they are likely to go through the main topics it can be a lot of information to take in. It is important that you also try to be sociable and ask others about what they do in order to expand you network and gain the most from the experience.
Form completed: 09 Oct 2018