First, write at the beginning

If someone asks, what does mathematical modeling give me? I would like to say that mathematical modeling brings me a wonderful sense of belonging. Mathematical modeling makes my mind grow rapidly, and it points out the direction of the application of mathematics to me.

Since I first participated in the simulation of mathematical Modeling contest held by UESTC at the end of 2019, I have also participated in the following university-level official contest in May 2020 and the simulation contest of MATHEMATICAL Modeling contest at the end of 2020. In these three competitions, there were awards, some won the highest award, some won the second highest award. Although I admit that the three competitions I participated in were only at the university level, which may not be as important as the provincial and national levels, I have always wanted to find an opportunity to summarize my experience and experience in the competitions. I have no training experience in mathematical modeling, and I may not have high mathematical talent, but I am proud of one thing — I treat every mathematical modeling competition with 200% sincerity and efforts, and I have always felt deeply that what mathematical modeling brings me is a sense of belonging.

This year I have a strong will is actually into electronic division senior colonel team training, to participate in the beauty of the game and the game in 2021, but because of some objective and subjective reasons, not very smoothly into the team (suffered a period of time, after all is love for the d/a), plus the finish from the final exam, training team started training, I think I missed a lot of should seize the opportunity. Although I often sigh, looking back on my experience in the competition and looking into my inner desire for mathematical modeling, I feel that I am full of expectations for the future. Maybe it is just my time. This is because of various factors, I made up my mind to summarize some of my own experience after the final exam, of course, just some of my own mental journey, including bad methods and detours and so on, and in view of the time relationship, pick some key points to write. At the same time because I didn’t accept training in mathematical modeling system, so this blog post, just record their comments on several competition and future place worthy of learning, so don’t have that kind of official or a biased conclusion, if readers think to have some help to oneself, can draw lessons from, if you think I am wrong method, welcome to point out!

Second, competition review

(I) Nov. 22-nov. 25, 2019

This is the first time for me to participate in the mathematical contest in modeling. To be direct, I have the mentality of trying. In fact, I have no idea about the word model when I was a freshman. Not to mention modeling, mathematical modeling is a completely undisciplined thing. However, with my love for mathematics, I signed up for this competition decisively. Unexpectedly, this competition gave me an indescribable love for the mathematical modeling contest and laid a foundation for my understanding of mathematical modeling.

Topic selection

• A) It is difficult to collect data from free charging facilities. B) It is difficult to collect data from free charging facilities. C) It is difficult to collect data from free charging facilities. There are no valid statistics for these different scenarios of devices and charging interfaces in different places. But in fact, I am more confident about such open questions. I believe in my ability of association and the ability to solve unknown problems.

• B) Star map recognition in celestial navigation. This problem is actually “very mathematical”, more research is in the three-dimensional graphics of geometric problems, including trigonometric functions, matrix in the three-dimensional graphics of the calculation. On the day of the competition we had just started learning integrals, and of course we didn’t learn much about linear algebra, so we chose to drop this problem. I gave up, but when I got this problem, I expanded the data package with Matlab to do a little shallow analysis, and my classmates and I have deduced 40 to 50 percent of the mathematical principle.

Time managementI don’t know how I learned about time planning, but I realized from the very beginning that before the competition, I should make a schedule of time planning for the whole team, so that the three first-time participants could have a number in mind.Probably content is above the text in the form, in fact, after the real match, only to find that in fact planning is difficult to implement real time, step by step to push very hard, but I’ve always thought a clear time planning can let the players know fairly well, can make to try to make progress to catch up with the plan, so in the later game, I always make a similar schedule based on the players. In fact, there were some problems when deciding the topic selection. Some members were unwilling to give up the B question and kept chewing on the B question. However, on the first night, we could not decide which question was really suitable for us, so the progress was very slow. I was worried too, but as the team leader, I had to make A decision immediately, so I decided to choose problem A to start.

The problem solvingOn the first night, I began to think about this topic in the early morning. In fact, I was the only one who stayed up late at night, so I thought about it in the all-night coffee shop of the school. As for the thinking content of the question, I recorded it in word in time, and also struggled with the topic repeatedly and updated the thinking direction repeatedly. That is, in the process of constantly struggling with the direction of thinking, I gradually became rightmodelI got some understanding. What impressed me most was that I understood the model as a function expression, which allowed me to better promote the modeling. Here are some of the thoughts (but many of them may be wrong, so readers are free to browse)In the following two days, I have been following my own thinking direction to improve the data analysis, and I used analogies for the data that could not be collected. I used analogies for similar or incomplete data sets to fit or reasonably assume the development trend of the required data. In the afternoon of the last day, I put my thoughts into English papers.

Thesis results feelingBecause this is the first time to participate in the mathematical contest in modeling, and his teammates tacit understanding degree is not high, so a lot of thinking cannot be synchronized to the other two teammates, how can lead to the final situation is not involved, so in terms of modeling, programming, and three paper have invested a lot of effort, also in the coffee shop stayed up the night was very thirsty. The process of this competition is very painful and painful, pacing alone in the small square outside the library, hours and hours of fidgeting just to get a little thought. At the last minute of the competition, I submitted a good paper, which filled me with a sense of achievement. The ups and downs of my mood in the three days were finally condensed into a paper, which was a very happy thing. The process of this competition was not smooth, but I learned how to face the seemingly impossible things, how to start from scratch, how to combine modeling programming and thesis. This experience laid the foundation for me to lead the team to participate in the competition. Specifically, I have better overall control, after all, I have been involved in all three aspects of the game.

(2) May 15- May 17, 2019 uESTC Mathematical Modeling Competition

In fact, this time the game, the heart is more bottomless, experienced the gains and losses of the last game, I have more confidence that I can calm down to solve the problem. This time the team is rebuilt, but I believe that the three people should be more tacit understanding. Because of the pandemic, the game will be played online, so the only way to be efficient is to make voice calls.

Topic selection

• A is also A typical data analysis question. It focuses on the economic development during the epidemic period and builds A model in line with the actual situation. This question is actually clearer to me than last year’s question, so first instinct: Answer A.
• B is an algorithm design problem to solve the problem of information matching when the search order is different. Because of the fact that the three people have not accumulated much algorithm in this aspect, I gave up the B question and chose A question.

scheduleIn fact, the schedule is similar to that of the previous year’s competition, and there are a lot of details that we know can’t be implemented, so it should be more operational. See a screenshot of some of the planning belowAs a captain, I have a distinctive characteristic, I will do my best to inspire the players, whether it is before the game, during the game or after the game. How to say, to encourage and comfort the team members is in fact to encourage and comfort themselves, the pressure of the captain is always greater than the pressure of the team members, in the pressure is already huge three days, the value of encouragement is mutual!The problem solvingIn determining which parameters and models are needed, I have listed some alternative algorithms and modelsIn the analysis of the overall economic situation, there is an obvious trend of first decline and then rise, while the recovery rate gradually slows down. I did not use the Logistic growth model, but the multi-parameter model constructed by myself, as shown in the figure below. If you want to ask me how to construct such a multi-parameter model, All I can say is that I took a fancy approach and then optimized it accordingly. Drawing tool: geometric sketchpad“The stock market is the barometer of the economy” — because of this “famous quote”, I put the focus of the next day to solve the problem on the stock, I downloaded “Hua CAI Life” haha, began to contact the first time to fry software.The thing to do, of course, is to use Matlab to fit the function to make reasonable predictions (BUT! This is very difficult because although there is a clear turning point in the outbreak, the volatility of the data before the turning point makes it difficult to fit the function, either before or after the turning point) in the decomposition of the problem that is later used to simplify the complexity of the problem to some extent by dividing it into medium – to long-term/short-term economies. In the short-term problem, I found that neural network can predict better, so I used BP neural network algorithm (with the help of Matlab, it can be simpler).Thesis results feelingThis competition gave me a much better feeling than the first time, whether it is the teamwork or the smoothness of solving problems. For the economic analysis of the epidemic, we have read a lot of economy-related materials, and modeled and analyzed the overall economy and local economy from various angles. Personally, I think it is a rather thorough analysis process, and of course, it is also the most substantial modeling process. However, just because of the thorough analysis, the content of the paper was actually messy. At that time, I wanted to pile up the quantity of the model, but there was no model body. Instead, I divided the problem into small details to build the model, which was of course a big drawback.

(3) 20.11.20.23, 2020.11.20.23, Electronic science and Technology University digital model Beauty competition simulation

The three people in the last competition had a tacit understanding, but due to some circumstances, one member of the team dropped out, so IN this competition, I found another student who wrote. I did not get along well, but on the whole, I paid more attention. In fact, I was very busy during this competition. I stayed up all night and took a comprehensive physical test (indoor test and outdoor test of 1km 50m). There were all kinds of things that disturbed my mood.

Topic selection

• A. College students choose the best vacation part-time job. In A nutshell, this question is given the personal information of college students (social time, expected working distance, expected working time…). And job information (working hours, working distance…) That decision! Which job should the college student choose is the best solution! The essence of this problem is the decision and matching problem of input and output. My first reaction when I saw this question — machine learning, with input and output, is to find a proper mapping relationship, but since all the information and data are unknown, machine learning is not a very direct way to train (but it must be a good way!). We finally chose the multi-attribute decision algorithm based on the mathematical modeling method tutorial.
• The typical mathematical derivation of question B is the uav cluster cooperative confrontation system based on swarm intelligence. At that time, there were some difficulties in understanding the meaning of the question, so the three people unified to choose another question.

The problem solvingThis problem is based on the mathematical modeling tutorial multi-attribute decision-making algorithm, on the basis of the joined the personalized elements, namely emphasis when users intend to filling out a questionnaire, the user has a different emphasis on each question and answer, at the same time every working range corresponds to the corresponding answer, and then through the multiple attribute decision making algorithm, Figure out how well the user fits in each job, the recommendation index!The picture shows that after figuring out the whole idea of the algorithm, I carried out a preliminary matrix operation, involving the transformation process of various matrices such as feature matrix, attribute matrix and decision matrix, and finally got the recommendation index J1-J8 of a user corresponding to eight different jobs, which can be seen in the calculation on the right
$8 J_i \ sum_1 ^ = 0.9999.$That means the whole operation is correct! The sum of all recommended exponentials is still close to 1.

Thesis results feelingThe modeling process was very difficult, because it was found from the very beginning that the problem path was very narrow, either the machine learning went from the training set to the test set, or the multi-attribute decision algorithm started from the matrix construction to the end. How to build a reasonable matrix is really difficult! In three days, the algorithm designed before was denied again and again, and the idea was updated again and again to find a more reasonable method. At the end of the second day, the idea of what to calculate the matrix was decided. In this process, the knowledge of engineering management and software engineering is learned and applied. We use Processon drawing software to draw a detailed algorithm flow chart, matrix construction flow chart, WBS decomposition diagram, etc., can give you a general look at the clear process.All in all, it was not easy these three days, and finally solved the problem of this narrow path. At that time, I was still confident of winning the prize. After all, every step of the calculation I did was verified by using papyri, and every step stood up to scrutiny.

Iii. Competition experience

Since I have not received professional training, my experience is accumulated from common sense, pre-race lectures, CSDN blog, and post-race summary. It can be said that I have listened to and recorded every pre-race lecture carefully. Here is a summary of some important points in my opinion.Most of these points are my own problems, so much of what follows is from my tournament memo.

(1) Modeling

• Familiar with the basic model of mathematical modeling textbooks, some questions based on the simple model to improve
• If you don’t have enough time to learn the models, be sure to see which models are commonly used before the game
• Open questions to learn to use an unconstrained mind to solve the problem
• Reasonable assumptions. Don’t get too obsessed
• It is recommended that you learn more about machine learning and have more experience in data processing
• Modeling is not about one person, it’s about three people working together to find the best solution

(2) Programming

• Proficient in at least one programming language, Matlab or Python is recommended for data analysis
• It is suggested that the students of programming have a good sense of programming, even if the problems will not be encountered, you can refer to the data to write a program to achieve
• If you’re interested in machine learning and have some level of understanding, that’s great
• In fact, in the early stage of the competition, only a small amount of programming will be used to do calculation, and in the later stage, programming may be involved to deal with specific problems such as fitting, prediction and planning

(3) Writing

• Build a framework early
• LaTex is, of course, the most ideal tool for writing papers, but if you are not skilled in using it, you should definitely choose Word
• Do not have a good English paper foundation, honestly write Chinese after modification and translation
• The paper is the most important part of the evaluation process! Typesetting and standardization determine the awards!
• The analysis section is in the future tense! No results!
• The annotation of the chart must be on the table below the chart. There are annotation tools in Word
• The paper catalogue is automatically generated, my first handwritten catalogue is really vomiting blood
• The standard format of the reference, the specific format self check, can now automatically generate the paper document format
• The abstract should not exceed one page of A4, and usually ends in 1/2-2/3 of the summary
• Multilevel heading layout (preset font size for multilevel heading)
• Remember the three words in bold
• Universal hypothesis (assuming all data is true…)
• When writing a paper in Word, there is a page number problem. Page breaks are really difficult to operate. It is faster to switch to WPS to add page numbers
• The specification of punctuation marks, the specification of formula glyphs are obvious concerns.

(4) Write memos frequently during the competition

This habit of writing memos is because I have a bad memory in my life. When I use memos in games, I find it is surprisingly useful, whether it is to record ideas in time or to modify things later.

(five) the captain should have a strong psychological

There is a lot of pressure on the captain if he is determined to do well in this competition. The first time I participated in the competition, the team members did not cooperate well with me, so my mentality broke down again and again in the three days, but I still survived. I spent a lot of effort in modeling, programming and thesis, and finally I finished a thesis at the last moment. Since I paid attention to the modeling and programming papers in the first competition, I led the team in the later competitions and had a better overall control view. After all, the three categories of labor division had been involved. Therefore, in the later games, although teammates have negative emotions, I also firmly believe that they are under control! No matter how messed up my mind is, or how influenced my mind is by the outside world, I have to show confidence in the big picture. Of course, confidence can not be blindly confident, confidence comes from the belief and determination to solve the problem and their own experience in the previous competition frustration, blow. Captain has a strong psychological ability to calm the mentality of others, the short-term depression is over is a positive face.

And last but not least! Hand in your paper within the time limit. No matter how hard or easy the game is, don’t lose by the rules!

I want to give you a few words that my math model told me when I failed:

• Loss is the best remedy for the brain
• The way of the world, good experience
• Good work is often done

This is to record my experience of participating in mathematical modeling in the past two years. Although I did not successfully enter the training team this time, these competitions have brought me much more than the prize itself, because — mathematical modeling is a kind of spirit.