I've participated in two of them, and I've been the team leader. The first thing you need to do is to look at the specialties of the members of your team, such as Physics, Chemistry, Biology, Architecture, and so on, and prioritize topics related to the subject matter. If you don't have people who are good at statistical data processing, you can consider avoiding topics with large data volumes, and vice versa. If you don't have anyone who knows Matlab, LINGO, or C, avoid optimization schemes, scheduling schemes, and topics that are computationally complex and impossible to implement by hand (but honestly, no one on the team knows how to program, so it's basically impossible to get anywhere). If you don't have anyone on your team with any specialties, it's best to do the most open-ended questions that leave you unsure of what you're being asked to do (some novice teams tend to go for the easiest and most straightforward questions, and end up not winning any prizes).

The above is the principle, in addition to the specific selection of the time to remember to spend a couple of hours on all the questions can be selected to collect information, we discuss, and then determine, do not get started on the beginning of the boredom drilling. And do not change in the middle of the question, do not split into two ways at the same time to do two topics.