One of the primary goals of this course is to prepare you to understand and apply state-of-the-art optimization algorithms to applications in research and industry. The final project is an opportunity for you to explore an area of interest, and to develop and demonstrate deeper understanding. The milestones for the project are listed below (check syllabus for specific dates).
The project is open-ended as long as it is related to the subject material of the class. If you need help brainstorming ideas please talk to Dr. Ning, the TA, or your advisor. Your project may involve applying existing optimization algorithms to an application of interest, developing novel optimization strategies or algorithms, or extending studies from past projects.
Explore previously published research papers in your field of interest for inspiration. If you are involved in research, you may consider a project applying optimization to a part of your existing research efforts. These often make some of the best projects as you have more background understanding and can progress in your research. Tying this project in with your research is encouraged but certainly not necessary. The most important criteria is to pick a topic that you are excited about.
You will work in teams of size two or three. Collaboration skills are critically important in engineering. You may work individually only if it is supporting your ongoing university research project and the project does not lend itself well to additional collaboration. To be clear, working on an ongoing research project does not necessitate working individually.
When forming teams make sure your project is properly scoped for the team size. Larger teams need a larger (usually deeper) scope. It’s usually best if each person has an equal but different role that they can take ownership on. This is especially important if tying in with one of the team member’s existing research. One challenge some 3-person teams have faced is not identifying separate contribution areas up front leading to one or two of the members doing most of the work. Three person teams usually increase scope by adding disciplines (e.g., adding acoustic considerations to an aero/structural problem) or adding depth (e.g., quantifying the sensitivity of the results to various parameters, exploring multi-objective trade-offs, performing optimization under uncertainty, etc.)
The final deliverable should be made publicly available unless extenuating circumstances apply (like working with proprietary data, in which case you need to obtain prior approval). Why? You tend to produce higher quality work with a public audience, and you will be able to make a more significant impact for good.
Some of you may want to submit your work to a conference or journal down the road, and may be concerned about prior publication. Almost all publishers do not consider the posting of preprints or research outputs online as prior publication. However, if you have any concerns, the archives we use provide an embargo option. This allows you to submit your work, but select a user-defined future time on when the content should publicly appear.