Mathematical Modelling: Strategies for Translating Real-World Problems into Mathematical Models | 数学建模：实际问题转化为数学模型的策略

📚 Mathematical Modelling: Strategies for Translating Real-World Problems into Mathematical Models | 数学建模：实际问题转化为数学模型的策略

Mathematical modelling is the art and science of turning a messy, real-world situation into a clean set of mathematical symbols, equations and assumptions that can be analysed, solved and interpreted. It is not a single step but a cycle of understanding, simplifying, representing, solving, validating and refining. This article sets out the core strategies that students need in order to make this translation reliably and efficiently, using the classic example of car stopping distance as a running case study.

数学建模是将繁杂真实世界情境转化为一组简洁数学符号、方程和假设的艺术与科学，使人们能够对其进行分析、求解和解读。它不是单一一步，而是一个循环往复的过程：理解、简化、表示、求解、验证与改进。本文以汽车停车距离这一经典案例为主线，系统介绍学生需要掌握的核心策略，以便可靠而高效地完成这种转化。

1. Understanding the Problem | 理解问题

Before any symbols appear, the modeller must know what the problem is actually asking. This means identifying the objective — the quantity to predict or optimise — and the constraints imposed by the real world. For the stopping-distance example, the objective is to predict how far a car travels from the moment the driver sees an obstacle to the moment the car stops.

在出现任何符号之前，建模者必须弄清问题究竟在问什么。这意味着要明确目标——需要预测或优化的量——以及现实世界所施加的约束。以停车距离为例，目标是预测从驾驶员看到障碍物到汽车完全停下这段时间内汽车总共行驶的距离。

Key questions at this stage include: What is the input and what is the output? Is the relationship deterministic or random? What time scale and spatial scale are relevant? Writing a one-sentence problem statement in plain language forces the modeller to be precise about scope. For instance: “Find the total stopping distance of a car travelling at a given initial speed on a dry, level road when the driver applies the brakes after a fixed reaction time.”

此阶段的关键问题包括：输入是什么？输出是什么？关系是确定性的还是随机的？相关的时间尺度和空间尺度是什么？用简明语言写出一句话问题陈述，有助于建模者准确界定范围。例如：“求汽车在干燥水平路面上以给定初速行驶、驾驶员经固定反应时间后踩下刹车时的总停车距离。”

2. Making Assumptions | 做出假设

No real-world problem can be modelled exactly in all its detail, so the modeller must deliberately simplify. Assumptions are the bridge between reality and mathematics, and every assumption should be stated explicitly so that its effect can later be tested.

没有哪个现实问题能在所有细节上被精确建模，因此建模者必须有意简化。假设是连接现实与数学的桥梁，每条假设都应明确陈述，以便日后检验其影响。

For the stopping-distance model, the standard simplifying assumptions are: the car moves in a straight line; the driver’s reaction time is constant; the braking deceleration is constant; air resistance and gradient are negligible; the road surface is uniform and dry; and the brakes are in good condition. These assumptions turn a fuzzy physical process into a tractable mathematical system.

对于停车距离模型，标准简化假设包括：汽车做直线运动；驾驶员反应时间恒定；刹车减速度为常数；空气阻力和坡度可忽略；路面均匀干燥；刹车状态良好。这些假设将模糊的物理过程转化为易于处理的数学系统。

A useful habit is to classify assumptions into three groups: structural assumptions about shape and form, quantitative assumptions about magnitudes, and boundary assumptions about the limits of validity. Writing assumptions down before equations forces clarity and honesty about what the model can and cannot capture.

一个有用的习惯是将假设分为三类：关于形态和结构的结构性假设，关于量级的定量假设，以及关于有效范围的边界性假设。在列方程之前写下假设，能迫使建模者对模型能捕捉什么、不能捕捉什么保持清晰和诚实。

3. Identifying Variables and Parameters | 识别变量与参数

After simplification, the next strategy is to distinguish variables from parameters. Variables are quantities that change within the problem; parameters are fixed constants that characterise the specific situation. In the stopping-distance model, the initial speed is usually a variable, while the reaction time and the braking deceleration are parameters fixed for a given driver, car and road.

简化之后，下一步策略是区分变量与参数。变量是问题中会变化的量；参数则是刻画具体情境的固定常数。在停车距离模型中，初速通常是变量，而反应时间和刹车减速度则是针对特定驾驶员、汽车和路面而固定的参数。

Symbols must be chosen carefully and defined with units. The table below summarises the quantities for the running example:

符号必须仔细选择并注明单位。下表总结了本案例中的各个量：

Quantity	Symbol	Unit	Type
Initial speed	u	m/s	variable
Reaction time	t_r	s	parameter
Braking deceleration	a	m/s²	parameter
Reaction distance	s_r	m	output
Braking distance	s_b	m	output
Total stopping distance	S	m	output

Defining this structure in words and symbols before writing any equation is what separates a disciplined modeller from a guesser. It also makes it much easier to check dimensions and to communicate the model to others.

在写任何方程之前，用文字和符号界定这一结构，正是训练有素的建模者与随意猜测者的分水岭。这样做还能更容易地检查量纲并向他人传达模型。

4. Choosing a Mathematical Structure | 选择数学模型结构

Once variables and parameters are identified, the modeller must choose the mathematical form that links them. This choice is driven by the assumptions made earlier. Constant speed during reaction time suggests a linear relation — distance equals speed multiplied by time. Constant deceleration during braking suggests a quadratic relation, since speed changes linearly with time and distance depends on the square of speed.

变量和参数确定之后，建模者必须选择将它们联系起来的数学形式。这一选择由先前做出的假设驱动。反应时间内匀速运动提示线性关系——距离等于速度乘以时间。刹车阶段匀减速运动提示二次关系，因为速度随时间线性变化，而距离取决于速度的平方。

The SUVAT equations of kinematics are the natural toolbox here. They give:

在此，运动学中的 SUVAT 方程组是自然的选择工具，它给出：

s_r = u × t_r

s_b = u² ÷ (2a)

So the total stopping distance is a mixed model: a linear term representing the reaction phase plus a quadratic term representing the braking phase. This illustrates a key strategy — simple models can often be assembled by combining smaller, well-understood pieces.

因此总停车距离是混合模型：表示反应阶段的线性项加上表示刹车阶段的二次项。这体现了关键策略——简单模型往往可以通过组合更小、更熟悉的模块来构建。

The same logic applies more broadly. If growth is proportional to current size, choose an exponential model. If growth slows as a limit is approached, choose a logistic model. If two quantities move together with a roughly constant ratio, choose a linear model. The modeller learns to match the mathematical behavior to the observed behavior.

同样的逻辑适用于更广范围。如果增长与当前规模成正比，则选择指数模型。如果增长在接近极限时放缓，则选择逻辑斯蒂模型。如果两个量以大致恒定的比率共同变化，则选择线性模型。建模者要学会让数学行为匹配观测行为。

5. Formulating Equations | 构建方程

With the mathematical structure chosen, the next step is to write the equations precisely. This includes stating the domain of each variable, for example u ≥ 0, t_r > 0, a > 0. The model combines the two distance components:

选定数学结构后，下一步是精确地写出方程。这包括注明每个变量的定义域，例如 u ≥ 0、t_r > 0、a > 0。模型将两个距离分量组合起来：

S = u·t_r + u²/(2a)

Where S is the total stopping distance in metres, u is the initial speed in m/s, t_r is the reaction time in seconds, and a is the magnitude of the deceleration in m/s².

其中 S 为总停车距离（米），u 为初速度（m/s），t_r 为反应时间（秒），a 为减速度大小（m/s²）。

Formulating the equation is not merely writing a symbol string; it requires checking that every term has consistent dimensions. The first term has units (m/s) × (s) = m; the second term has units (m/s)² ÷ (m/s²) = m. Both terms, therefore, qualify as distances, which is a quick and powerful sanity check.

构建方程不仅仅是写下符号串，还必须检查每一项的量纲是否一致。第一项的单位为 (m/s) × (s) = m；第二项的单位为 (m/s)² ÷ (m/s²) = m。因此，两项都可作为距离，这是一个快速而强有力的合理性检验。

Another useful discipline is to express a word problem as a conditional statement: “Given u and the parameters t_r, a, find S.” This explicit input-output view lets the modeller decide whether the problem is a direct computation, an inverse problem, or an optimisation problem, and it guides the later solution strategy.

另一个有用的训练是用条件语句表达应用题：“已知 u 以及参数 t_r、a，求 S。”这种明确的输入输出视角，使建模者能够判断问题是直接计算、反问题还是优化问题，并指导后续求解策略。

6. Solving the Model | 求解模型

Once the model is formulated, the modeller chooses a solution method. For simple algebraic models, direct substitution is enough. Suppose u = 30 m/s, t_r = 0.8 s and a = 6 m/s². Then the reaction distance is 30 × 0.8 = 24 m, and the braking distance is 30² ÷ (2 × 6) = 900 ÷ 12 = 75 m, giving a total stopping distance of 99 m.

模型建立后，建模者要选择求解方法。对简单代数模型，直接代入即可。设 u = 30 m/s、t_r = 0.8 s、a = 6 m/s²，则反应距离为 30 × 0.8 = 24 m，刹车距离为 30² ÷ (2 × 6) = 900 ÷ 12 = 75 m，总停车距离为 99 m。

In many problems the model must be rearranged. For example, if the driver wants to stop within 60 m, what is the maximum safe speed? Rearranging the model gives a quadratic equation in u:

在许多问题中，模型需要重排。例如，若驾驶员希望在 60 m 内停下，最大安全速度是多少？重排模型得到关于 u 的二次方程：

u² + (2a·t_r)u − (2a·S) = 0

Substituting t_r = 0.8, a = 6, S = 60 gives u² + 9.6u − 720 = 0. Using the quadratic formula:

代入 t_r = 0.8、a = 6、S = 60，得 u² + 9.6u − 720 = 0。利用求根公式：

u = [−9.6 + √(9.6² + 4 × 720)] ÷ 2 ≈ 22.4 m/s

Only the positive root has physical meaning, demonstrating another modelling skill: applying common sense to filter mathematical solutions. The negative root is ignored because speed cannot be negative, and the answer is rounded to an appropriate degree of accuracy.

只有正根具有物理意义，这展示了另一项建模技能：运用常识过滤数学解。负根被舍弃，因为速度不可能为负，答案按适当精度取整。

7. Verification and Validation | 验证与检验

A solution is only meaningful if the model itself is sound. Verification asks whether the equations were solved correctly; validation asks whether the equations describe reality well enough. Both are essential steps before the model can be trusted.

解只有在模型本身可靠时才有意义。验证（verification）问的是方程是否求解正确；检验（validation）问的是方程是否足够好地描述现实。在进行模型之前，两者都是必不可少的步骤。

One verification tool is dimensional analysis, already discussed. Another is testing extreme cases. If u = 0, the model gives S = 0, which is correct — a stationary car needs no stopping distance. If a is very large, the braking distance becomes very small, which matches intuition. If t_r = 0, only the braking term remains, which again makes sense.

验证工具之一是前面已讨论过的量纲分析；另一个是检验极端情形。若 u = 0，模型给出 S = 0，这是正确的——静止的汽车不需要停车距离。若 a 很大，刹车距离变得很小，这符合直觉。若 t_r = 0，则只余刹车项，这也合理。

Validation compares the model output with real data. Road-safety studies often show that a typical reaction time is between 0.7 and 1.5 seconds, and braking deceleration on dry asphalt is roughly 6 to 8 m/s². If the model predicts 99 m for u = 30 m/s and measured stopping distances at that speed are around 95–110 m, the model agrees well. If data were collected on a wet road, the model would overestimate safety because the assumption of a dry road has been violated.

检验则是将模型输出与真实数据比较。道路安全研究常表明，典型反应时间在 0.7 到 1.5 秒之间，干燥沥青路面上的刹车减速度约为 6 到 8 m/s²。若模型在 u = 30 m/s 时预测 99 m，而实测值约为 95–110 m，则模型吻合良好。如果在湿滑路面上采集数据，模型会高估安全性，因为它违背了“干燥路面”这一假设。

This step often reveals that the model is good in some regimes and poor in others, prompting the modeller to return to earlier stages — a reminder that modelling is not linear but cyclical.

这一步常常揭示模型在某些范围内表现良好、在另一些范围内表现不佳，促使建模者回到较早阶段——这提醒我们，建模不是线性的，而是循环的。

8. Sensitivity Analysis | 灵敏度分析

Sensitivity analysis examines how changes in parameters affect the output, revealing which parameters matter most. For the stopping-distance model, the derivative of the braking distance with respect to a is negative, because stronger braking reduces distance. The derivative with respect to t_r is simply u, meaning each extra second of reaction time adds exactly u metres to the stopping distance.

灵敏度分析考察参数变化如何影响输出，揭示哪些参数最重要。对于停车距离模型，刹车距离对 a 的导数为负，因为更强的刹车会缩短距离。S 对 t_r 的导数恰为 u，意味着反应时间每增加一秒，停车距离就增加恰好 u 米。

More dramatically, the braking distance is proportional to u², so doubling the speed quadruples the braking distance. This nonlinear amplification is one of the most important insights the model provides — it explains why speed limits matter so much for road safety.

更显著的是，刹车距离与 u² 成正比，因此速度加倍会使刹车距离变为四倍。这种非线性放大是模型提供的最重要洞见之一——它解释了为什么限速对道路安全如此重要。

In general, the modeller can compute partial derivatives or rerun the model with perturbed parameters. If a small change in a parameter causes a huge change in the output, extra effort should go into measuring that parameter accurately. Sensitivity analysis thus directs both data collection and further refinement.

一般而言，建模者可以计算偏导数，或用摄动参数重新运行模型。如果某个参数的微小变化导致输出巨大变化，就应投入额外精力精确测量该参数。因此，灵敏度分析既指导数据收集，也指导进一步改进。

9. Iteration and Refinement | 迭代与改进

Seldom is the first model the final model. Iteration means going around the modelling cycle again, using what was learned from validation to improve the assumptions. For the stopping-distance example, a first refinement might add the effect of air resistance by treating deceleration as a function of speed rather than a constant.

第一版模型很少是最终版本。迭代意味着再次走完建模循环，利用从检验中学到的东西改进假设。以停车距离为例，第一次改进可将减速度视为速度的函数而非常数，从而计入空气阻力效应。

A second refinement might separate total reaction time into perception time and movement time. A third might model the road gradient, where the effective deceleration becomes a·cos θ ± g·sin θ depending on whether the car is going uphill or downhill. Each refinement increases accuracy but also increases complexity, so the modeller must weigh both.

第二次改进可能将总反应时间细分为感知时间和动作时间。第三次改进可能建模道路坡度，此时有效减速度变成 a·cos θ ± g·sin θ，具体取决于汽车上坡还是下坡。每次改进都提高精度，但也增加复杂度，建模者必须权衡二者。

A key principle here is parsimony — a model should be no more complex than necessary. Adding parameters reduces clarity and makes estimation harder. The modeller should only add complexity when validation shows that the current model misses an important pattern and when the added terms have a clear physical or contextual justification.

这里的关键原则是简约性——模型不应比所需更复杂。增加参数会降低清晰度并使估计更加困难。只有当检验表明当前模型遗漏了重要规律，且新增项有明确物理或情境依据时，建模者才应增加复杂度。

10. Communicating Results | 沟通结果

The final strategy is communication. A model is useless if its results cannot be understood by others — engineers, policymakers or fellow students. The modeller must state the assumptions clearly, present the equations with defined symbols, show the numerical results with appropriate rounding, and honestly report the model’s limitations.

最后一项策略是沟通。如果结果无法被工程师、决策者或同学理解，模型就毫无用处。建模者必须清晰陈述假设，用已定义符号呈现方程，以适当精度展示数值结果，并诚实地说明模型的局限性。

Useful presentation devices include graphs of S against u, tables comparing model predictions with measured data, and a summary of sensitivity findings. For the stopping-distance model, a graph of S versus u shows a gentle upward curve for the reaction component and a much steeper curve for the braking component, instantly communicating why high-speed driving is dangerous in braking terms.

有用的展示方式包括：S 随 u 变化的图像、比较模型预测与实测数据的表格，以及灵敏度结果摘要。对于停车距离模型，S 对 u 的图像显示反应分量是缓和的上升曲线，而刹车分量是陡峭得多的曲线，立刻能传达出高速行驶为何在刹车距离上如此危险。

Finally, communication includes acknowledging uncertainty. Every model has error bars, hidden assumptions and edge cases. A responsible conclusion explicitly states the range of conditions under which the model is reliable — for example, “valid for dry roads, passenger cars, and speeds between 10 and 60 m/s” — and recommends how the model might be extended in future work.

最后，沟通还包括承认不确定性。每个模型都有误差范围、隐含假设和边界情况。负责任的结论应明确指出模型可靠的条件范围——例如“适用于干燥路面、乘用车、10 至 60 m/s 的速度区间”——并建议未来如何扩展模型。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Mathematical Modelling: Strategies for Translating Real-World Problems into Mathematical Models | 数学建模：实际问题转化为数学模型的策略