大概是去年做的E题，和别人做了对比，又重新阅读了美赛的官方说明。
也包括阅读评委评审意见的收获，整理如下。
如果对本文有疑问或者想找男朋友，可以联系我，点击此处有我联系方式。
2019 Contest changes/updates can be found in below:
MCM/ICM is an all electronic submission!
You are not required to mail a print copy of your signed Control Sheet.
We have added 3 hours to the contest start time to allow for teams to download, read, discuss, and choose a problem.
The contest problems will become available precisely at 4:50PM EST on Thursday January 24, 2019 on the following mirror sites:
In competition, some tips:
Partial solutions are acceptable. There is no passing or failing cut-off score, and numerical scores will not be assigned. The MCM/ICM contest judges are primarily interested in the team’s approach and methods.
The summary is an essential part of your MCM/ICM paper. The judges place considerable weight on the summary, and winning papers are often distinguished from other papers based on the quality of the summary.
Restatement and clarification of the problem: State in your own words what you are going to do.
Explain assumptions and rationale/justification: Emphasize the assumptions that bear on the problem. Clearly list all variables used in your model.
Include your model design and justification for type model used or developed.
Describe model testing and sensitivity analysis, including error analysis, etc.
Discuss the strengths and weaknesses of your model or approach.
The judges will evaluate the quality of your writing in the Solution Paper:
- Conciseness and organization are extremely important.
- Key statements should present major ideas and results.
- Present a clarification or restatement of the problem, as appropriate.
- Present a clear exposition of all variables, assumptions, and hypotheses.
- Present an analysis of the problem, including the motivation or justification for the model that is used.
- Include a design of the model.
- Discuss how the model could be tested, including error analysis and stability (conditioning, sensitivity, etc.).
- Discuss any apparent strengths or weaknesses in your model or approach.
About MCM competition template:
Papers must be typed in English, with a readable font of at least 12 point type.
The solution must consist entirely of written text, and possibly figures, charts, or other written material only. No non-paper support materials such as computer files or software will be accepted.
The Solution Paper must display the team control number and the page number at the top of every page; for example, use the following page header on each page:
The names of the students, advisor, or institution should NOT appear on any page of the electronic solution. The solution should not contain any identifying information other than the team control number.
Print/Download one copy of the Control Sheet.
Download or copy the team Summary Sheet. (This should be used as the first page of your electronic email submission.
COMAP will accept only a Adobe PDF of your solution. DO NOT include your Control Sheet, programs or software with your email as they will not be used in the judging process.
Limit one solution per email.
The names of the students, advisor, or institution should NOT appear on any page of the electronic solution. Your team’s summary should be included as the first page of your file. Note: The attachment must be less than 17MB.
About MCM competition time:
Each team is required to submit an electronic copy of its solution paper by email to solutions@comap.com. Any team member or the advisor may submit this email.
Send the signed Control Sheet by email to COMAP:After the signed control sheet is prepared, email it to: forms@comap.com.
In the subject line of your email write: your team’s control number. For example: 1900000.
Some Thinking About Reading UMAP
论文中的某些部分可以简写，但是考虑因素要全面。符号说明简明清晰的包含所有要素。摘要写一页，包括使用的方法和对结果的总结,传达模型的意义(有所限制的展示基本的模型和潜在的假设)。包含你的结果，结果做到量化。
有关模型(必须有优缺点)：网络发展的今天，很多参赛队都能在网上搜到差不多的模型进行改进，然后解决问题。但是更希望建立连续(具有凝聚力)的模型，并且在足够的细节上进行描述推导来解释结果之间潜在的联系。
论文简洁有组织。在符号说明中定义所有变量，但是论文中不能定义太多，及解释每个变量的意义甚至单位，以变量的单位保证方程的相等性。官方没说层次分析法遗传算法不好，放开了去使用，但是不要写推导过程，可以提供附件代码。
对于因子、常数、系数，确保有详细的解释说明。记录搜集数据的工作，列出数据缺失而产生的假设和简化问题而做出的假设，以数学方程表达变量之间的关系，使用合理的方法分析变量之间的关系来支持你的结论。
做只与模型相关的假设，包括理由以及如何影响模型。重点是阐述假设如何应用到模型中，甚至做模型假设的灵敏性分析检验假设，模型假设可以啊，但这种假设会对模型的精确行产生多少影响呢？需要灵敏行分析。并进行敏感性分析以解决所做的任何假设，或估计所用相关变量的预期变化。只对最最最重要的变量进行灵敏性分析。
描述模型建立的原因和动机。(数据分析详细的描述如何使用的数据)，如果行业内有相关的制定标准，一定要引用进来。搜寻可靠的数据，并记录数据来源。不希望有广阔的数据来源和精细复杂的模型，只希望看到在模型建立过程中的分析，不强行规定解决问题的方向。
成功的团队提供了解释模拟结果的统计显着性。然而最成功的团队详细解释了他们的模拟过程；一些甚至包括流程图。结果是不重要的，重点是对结果的分析。为了问题而建模，而不是为了建模而建模(最好的论文考虑通过模型而测量指标，而其他论文为了测量指标而建立模型)。
然而，最成功的团队将他们的敏感性分析集中在他们在建模过程中所做的假设上。这提供了敏感性分析和建模过程之间的一致性，并且提高了团队建议的可信度（如果模型结果对所做的假设不敏感）或建议的进一步分析区域。对模型假设的参数的检验，懂得灵敏性分析的目的，为什么做灵敏性分析。
明确设定的相关变量的单位（保证方程两边单位相等），做出的假设要么是为了弥补数据的不足，要么是为了解释相关变量之间的数学关系。引用别人的方法并不可耻(shame)，如果引用请做好标注，会详细的检查参考文献。适当的数学技术或方法来分析数学关系并形成建议，写出来的信件包含所有因素，使用简单的英语描述非技术性元素。
