Association Math Modeling Competition (AoCMM) is a math modeling competition, held from 9.27.2017 – 10.10.2017. Our school sent a number of teams (namely Team 690 to 694). I, with other three team members (Bendit, Celeste and Chi Hang) is one of them. In one sentence simply, I will describe this competition is a disaster, hence this reflection that I must note them down, and hopefully learn from it.
Problems: (see http://aocmm.org/problems-2017)
1. It has been shown typing patterns can be used to identify a person. To confirm this idea, we collected typing patterns from eleven of our officers using two different typing methods: fourteen short English quotes and six paragraphs composed of random characters. The quotes and paragraphs are grouped into three categories: eight quotes and three paragraphs where the officer’s identity is known, six quotes where the officer’s identity is unknown, and another three paragraphs where the officer’s identity is also unknown. Based on the first category, how would one match the second and third categories to the officers?
2. There will always be times when taxis are vacant. Some drivers say that you should head to the city center to find more customers, but is that always true?
1) Suppose you are a taxi driver in NYC, what should you do when your car is vacant?
2) If you are the head of a taxi company, what would you advise your drivers do?
(Beforehand + to 2.10.2017)
Before the competition, I start preparing the competition: (i) Plan the schedule (ii) Ask Chi Hang and Bendit to learn statistics (iii) Ask Bendit to type the table of contents (as AoCMM has already a format that requires us to follow). Turn out the preparation is quite worthless: (i) The schedule we planned cannot match the opening time of school computer room, which reduce our working hours a lot. (ii) Chi Hang just ignored my request, which is quite expected. (iii) I planned the table of contents, and the first template is out 29.9.2017, two days after the competition start, GREAT!!!
We cannot gather to discuss the problems until 1.10.2017, Bendit and I have discussed (and decided) the big picture of the two problems in telephone (I think that is natural due to past experience, the model is designed mainly by me, with help with Bendit, as we (especially me), have a lot of math skills. However, One of school teachers Mr. So points out that although we have difficulties communicating with Chi Hang, the potential contribution of Chi Hang shouldn’t be overlooked). In the first problem, the obvious way is that we can extract the possible tying characteristic in reference template (More precisely, typing characteristic should not randomly varies for a fixed person, which should be justified by academic (biometric) references). And then compare with the typing characteristic with the test template. Find the most similar one. For problem two, Bendit inspired me to think in a game theoretic way. Hence by some searching, I found this reference as the cornerstone of our model (https://www.researchgate.net/publication/260026333_A_novel_approach_to_independent_taxi_scheduling_problem_based_on_stable_matching).
After the discussion in the telephone, 1/10 comes, when I came back to the working place carrying the lunch of my teammates. I discovered that my teammates just quite ignored me, (obviously) falsely believing themselves have built up the model, based on our telephone discussed and paper. It quite brings up my anger and sadness that anyone ignore my request that to discuss the model together. And anyone just try to do things (which thinks that it is useful) by themselves. At last, I am able to force themselves to discover their mistake. It lasts only a short discussion because we need too pack things and go back home. After the competition, I asked Mr. So that the portion of discussing the model and typing the paper, Mr. So thinks that only when the model is left with details, the paper should be begin typing. Hence had I forced them to discuss the model, the discussion would be more efficient. One of our big mistakes is that we failed to discuss much of our model. The holes of model are badly fixed by me in the last minutes.
Back to our model, we basically done a half of problem 1, thinks out the procedure to solve problem 1, which I lately written the notes of problem 1(AoCMM 2017 Problem 1). And dividing the two sub-problems into half, considering the case of taxi drivers knows about the location of waiting passengers and taxi drivers and the case of street hailing. The paper have already made us solve two out of four cases (the case that taxi drivers knows about the location of waiting passengers, for both part (a) and (b)). That’s all.
In the next day, which I will describe as the most useless day, but with the highest opportunity cost. I discovered that Chi Hang has been hugging Bendit (For outside readers, it is some sort of showing friendship, but not showing homosexuality…), but that makes Bendit cannot work efficiently and gets more frustrated towards Chi Hang and Bendit. I gave up training of Cheering team, but only exchanges that we typing the first little part. To make the matter “worse”, Mr. Leong at last found us and accused us violating the school rules of using school classrooms. I am not going to say that how the stupid school rules are, nor the miscommunication of teachers and us. In fact, this incident may benefit us to make us closer to each other. At last, this is 3.10.2017.
(3.10.2017 to 10.10.2017)
In the mean time, Bendit discovers that the data given by AoCMM in problem 1 didn’t give the typing characteristic that commonly used in academic. In 1/10/2017 I see the problem which I thought just a small detail which requires a little time to solves. However, I need to point out even I preform badly this time, neglecting this mistake, just passing to Bendit to solve. I cannot escape this fault, letting Bendit himself to finish this task. Although Bendit has acquired the prefect skills to solve, I know Bendit is a person that don’t want to solve the problem himself but always search help from me. After all, my school teacher Mr. Lee comes up and suggest us for other typing characteristic. And then Bendit and I finish up the issue. After that, we have gathered for typing in 5.10.2017, 8.10.2017, 9.10.2017 and the last day 10.10.2017.
I thinks that our performance is rubbish because
1. For our paper, I think we fail to write out in a clear way that the judge could understand, in other words, I think our paper is quite incomprehensible.
That’s because (i) I discover our team members don’t quite understand what we are doing, which make my anger and devastation as I have explained repeatedly and no one ask me. (ii) The writer cannot write things out clearly? (iii) No one are able to check except me, which are extreme busy thinking solutions.
2. I feel like our model is bad.
I think the blames should go to the school, which fails us to teach us the formal way of statistics, the correct way to select the “best” model. I think my team members didn’t train to write assumptions and justifications, sensitivity analysis… And not knowing the philosophy of model, even I have kept in touch with Physics and Economics (two so call celebrity mathematical models). I didn’t feel like I am doing right.
3. Rubbish team work.
This is no laughing matter and we should not turn a blind eye on it: We fail to communicate (expected in Mr. So, pointing out we just have 1+1+1+1<1, but not an excuse), this would be a wake up call for me to think up solution. I am tired to explain the solution that is discussed more than three times. Separately to Chi Hang and Celeste for the use of letting them slapping quite incomprehensible words into latex! We need to fix this.
Yes, We need to fix this.