Something magical has happened; our team has been ranked as Honorable Mention (the third-highest award) in the IMMC!! Along with other prestigious schools such as the Hwa Chong Institute of Singapore, Gyunggi Science HS, we have made it to the third place in one of the biggest modelling competitions in the world!
This is undoubetldy one of the greatest achievements in my math career, and my next dream is to win the competition in 2021. But what is more amazing is that we used probably the most unconventional method, where we employed the programming language Scratch, to simulate consumer actions.
While C++ or Python could've been better in crunching down big amounts of data and we could've used path-finding algorithms by assigning a value to each aisle, we thought that Scratch would actually be better as it better models interactions between consumers. Before we came to this conclusion, we also thought of treating consumer flow just as fluid dynamics. The attraction that each product had could be considered as the bottom of the sink being lower than others, hence attracting more fluid than others. But we knew virtually nothing about fluid dynamics, and had to discard the idea.
Then, I encountered my "Eureka" moment; I remembered the book The Model Thinker saying that the more models the better. Hence, I decided to create two models--one in Scratch and the other in C++, where the former better modelled human interactions such as collisions or drag while the latter was better at modelling path-finding algorithms. I also created a flowchart as the one on the right.
Showing you the actual work would be much efficient than explaining it in mere words; I've added our final report so you could read it.
Download DOCX • 9KB
Download PDF • 19.60MB
The five days during the IMMC was, I have to admit, pure insanity. Researching and writing a report in mere five days was insanely difficult, and I think it's extremely important that we decide what to take out rather than what to put in. Every time IMMC comes up, I just know that I won't be able to sleep for 48 hours. But still, it is such great fun, and I will reply "yes" if anyone asks me if I will redo it in 2021, without even a tinge of second thought.
In the future, I hope I could research further on this topic. Also, as I am currently learning Mathematica and machine learning in Python, I would want to see whether either could be applied in this simulation.