Formulate Research Question

On 16 Oct 2022, at 19:28, Pereira Carvalhido, Flávia Carvalho Gavinha [fp48756@student.uni-lj.si](mailto:fp48756@student.uni-lj.si) wrote:

Dear Professor Iztok,

My name is Flávia Carvalhido and I'm a part of group C. For this week's milestone, we have been brainstorming about some projects and research experiments that we believe would be interesting for this subject and for the proposed research paper we need to prepare. This email serves as a way of validating our ideas. We are mostly interested in understanding if you believe our ideas would be achievable during the suggested assignment's timeline, if they fit the theme of the course, and if they would meet your grading requirements (as in, if done properly, would it be considered a project worthy of a grade 9/10).

Our ideas started from 2 papers that the whole group found interesting: (1) A review on crowd simulation and modeling and (2) Fuzzy Logic-Based Model That Incorporates Personality Traits for Heterogeneous Pedestrians.

As a starting point, we all considered Social Psychological Crowds interesting (stemming from the theme of paper 2), moreover the concepts of Emotion Contagion and the psychological dimensions which might affect group behavior. Given the big recent interest in these types of crowds and how to model them, there have been some recent developments in this area that aren't further explored in paper 1's systematic review.

Therefore, we propose to make a Review of Psychological Crowds Simulation and Modeling, further exploring more recent state-of-the-art techniques and others already mentioned in paper 1, have them all perform the same task, and evaluate and compare the different approaches, both on a technical and conceptual level.

We are still uncertain about which paper's task to replicate. Whist paper 1 includes a task capable of comparing all types of models in a quantitative way, paper 2 aims to evaluate the developed model in a qualitative way, by generating simulations and asking users to identify which personality type was being modeled in the crowd, based on the OCEAN personality categories. Any suggestions on which would be a good experiment for our proposed idea: paper 1's or paper 2's?

Would this idea fit the course's assignment? Do you believe it could have some added scientific value? Does it comply with the aspects asked above? If not, what do you suggest we change in order to make it a better project concept, which would be achievable and proper for this course's goals?

Thank you for your attention to this. We await your response.

Best regards, Flávia Carvalhido

pon. 17. 10. 2022 20:59

Flávia,

To answer your questions, yes, the ideas fit the theme and from that perspective they meet the criteria for grading. The end results obviously depend on your commitment and how persistent you will be in devoting time to this project. With a wise distribution of tasks and a continuous effort you will probably reach your final goal. If you intend to do a review of various models, I suggest you limit their number, and start with just a few adding additional ones if you have time. Giving yourselves too big a challenge may lead to you getting overwhelmed. Qualitative assessments can get tricky, because they require you to have everything setup in such a way that you can then invite others to view, test and answer polls. Since the poll requires time too, you have to have everything finished well ahead of time. In addition, it can become tricky to get enough people involved to make the results valid. What I’m trying to say is, a quantitative comparison may turn out to be easier to accomplish.

If you intend to go the fuzzy path, there are numerous papers discussing the use of fuzzy logic in crowd simulation (https://scholar.google.si/scholar?start=0&q=fuzzy+logic+crowd+simulation). A couple of examples:

In addition perhaps have a look at a fuzzy toolkit for Unity that is part of a masters’ thesis currently being prepared by a student of mine (I’m giving you the code as is, there is no documentation, but the toolkit can be added to any project via Window->Package Manager->Plus->Add package from git URL)

Iztok

Troedelmaus / Collective-Behaviour-22-23-Group-C

Formulate Research Question #3