Problem Solving with Differentiation

Problem solving with differentiation uses derivatives to locate maximum or minimum values of a quantity that depends on a variable. The key idea is to model the quantity algebraically, reduce it to a function of one variable, differentiate, and solve where the derivative is zero to find critical points. This method is central to optimisation because it converts practical constraints into mathematics and then uses gradient behavior to identify the best possible value.

• Optimisation is the process of finding the greatest or smallest possible value of a quantity, such as area, volume, profit, or time, subject to given constraints. In calculus, this is done by expressing the quantity as a function and using its derivative to locate points where the function may change from increasing to decreasing or vice versa.
• A variable quantity is the value you want to optimise, while a constraint is the condition that links the variables in the problem. This matters because most real problems begin with several variables, and the constraint lets you rewrite the target quantity in terms of just one variable so differentiation can be applied directly.
• A critical point occurs when the derivative is zero or undefined, and these are the main candidates for maxima or minima. In typical algebraic optimisation at this level, the derivative is set equal to zero because a turning point has a horizontal tangent, so the instantaneous rate of change is zero.
• If a function is written as $Q(x)$, then its derivative $Q'(x)$ or $\frac{dQ}{dx}$ measures how the quantity changes as $x$ changes. This is why the equation $\frac{dQ}{dx}=0$ is so important: it identifies where the quantity stops increasing and starts decreasing, or the reverse.

• The derivative tells you whether a function is increasing or decreasing at a particular value of the variable. If $\frac{dQ}{dx}>0$, the quantity is rising as $x$ increases, and if $\frac{dQ}{dx}<0$, it is falling, so a change in sign of the derivative signals a local extremum.
• At a local maximum or minimum, the graph is often flat for an instant, which is why the derivative is set to zero. In symbols, if $x=a$ is a turning point of $Q(x)$, then typically $Q'(a)=0$, meaning the tangent there is horizontal.
• Optimisation problems usually require more than differentiation alone because the quantity of interest may depend on several variables. The mathematical strategy is to use the given conditions to eliminate extra variables first, since differentiating a multivariable expression without reducing it to one variable will not directly identify the required extremum in this context.
• The shape of the resulting function gives useful insight about whether the stationary point is a maximum or a minimum. For example, a quadratic with negative leading coefficient opens downward and therefore has a maximum turning point, while a quadratic with positive leading coefficient opens upward and therefore has a minimum.

General optimisation workflow

• Define the variable clearly before doing any algebra, and state what it represents in the context of the problem. This is essential because the meaning of the derivative depends on the variable, and unclear notation often leads to differentiating the wrong expression or using the wrong units.
• Write the target quantity as a formula using geometric, algebraic, or contextual relationships. For example, if the quantity to optimise is area, volume, or cost, start from the basic formula for that quantity rather than guessing a simplified expression too early.
• Use the constraint to remove extra variables so the target quantity becomes a function of one variable only. This step is the heart of most optimisation problems, because differentiation of a one-variable function such as $Q(x)$ is straightforward, while a formula like $Q(x,y)$ does not yet match the intended method.
• Differentiate and solve $\frac{dQ}{dx}=0$ to locate candidate extrema. This works because turning points occur where the instantaneous rate of change is zero, so the graph temporarily levels out.
• Substitute back into the original quantity formula to find the actual maximum or minimum value. The solution to $\frac{dQ}{dx}=0$ usually gives the value of the variable, not the optimised quantity itself, so this final substitution is necessary to answer the question asked.

Core optimisation rule: Model first, reduce to one variable, differentiate, solve $\frac{dQ}{dx}=0$, then substitute back.

Worked generic model structure

• Suppose a rectangle has sides $x$ and $y$ and a condition links them, such as $2x+y=K$ for some constant $K$. You would rearrange to get $y=K-2x$, substitute into $A=xy$, and then differentiate the resulting one-variable expression for area.
• This pattern appears across many contexts, even when the symbols change from $A$ or $V$ to other letters. The mathematics stays the same because the derivative only needs a single independent variable and a differentiable function built from the constraint.

Quantity to optimise vs variable used

• The optimised quantity and the variable you solve for first are often not the same thing. For example, solving $\frac{dQ}{dx}=0$ usually gives a value of $x$, but the final answer may require the corresponding value of $Q$, so you must distinguish between the input value and the quantity being maximised or minimised.

Stationary point vs final answer

• A stationary point is a candidate location for an extremum, not automatically the requested answer. The problem may ask for the dimensions that produce the best value, or it may ask for the best value itself, and these are different outputs from the same calculation.

Maximum vs minimum

• A maximum occurs when the function changes from increasing to decreasing, while a minimum occurs when it changes from decreasing to increasing. Recognising the function's overall shape or checking the sign of the derivative around the critical point helps decide which one has been found.

• In many school-level problems, the resulting expression is a quadratic, so classification can often be done from the sign of the leading coefficient. That shortcut is efficient, but it should only be used when you are certain of the graph type and the algebra has been simplified correctly.

• Start by naming the quantity to optimise and the variable you will use, because this makes the algebra purposeful and reduces confusion later. Examiners often award method marks for setting up the model correctly, even before differentiation is completed.
• Check that your formula is in one variable before differentiating. If you still have two variables in the expression, that usually means you have not yet used the constraint fully, and any derivative you compute will not match the intended optimisation method.
• After solving $\frac{dQ}{dx}=0$, read the question again to decide whether it wants the variable value, the dimensions, or the maximum/minimum quantity. Many lost marks come from stopping at the stationary value of $x$ and forgetting to evaluate the actual quantity.
• Use a reason to justify maximum or minimum, such as the sign change of the derivative, the quadratic shape, or comparison of candidate values. A correct numerical answer with no classification may be incomplete if the question explicitly asks why the result is a maximum or minimum.
• Do a reasonableness check by considering the context and domain. If a dimension becomes negative, if the quantity is impossible for the given setting, or if the stationary value lies outside the allowed interval, then the model has either been misused or the candidate point must be rejected.

Always check: variable defined, constraint used, derivative set to zero, value substituted back, and extremum justified.

• A common misconception is thinking that solving $\frac{dQ}{dx}=0$ directly gives the maximum or minimum quantity. In fact, it usually gives the value of the variable where the extremum occurs, and you still need to substitute this back into $Q(x)$ to find the actual optimised value.
• Students often differentiate too early, before rewriting the target expression in one variable. This creates an incomplete method because the derivative of an expression containing extra dependent variables does not yet reflect the single-variable relationship imposed by the constraint.
• Another frequent error is confusing the derivative with the original function. If $Q'(x)=0$ is solved correctly but then the resulting value is substituted into the derivative again instead of the original quantity formula, the final answer will not represent the required maximum or minimum.
• Classification is sometimes assumed rather than shown. Even if the turning point looks like a maximum from intuition, you should support that conclusion with function shape, sign change, or another accepted method so the argument is mathematically complete.

• Problem solving with differentiation is closely connected to stationary points, because optimisation uses the same idea of locating where the gradient is zero. The difference is that the turning point now represents a best possible value in a real or geometric context rather than just a feature of a graph.
• It also connects to algebraic modelling, since the hardest part of many optimisation problems is often building the correct expression before calculus begins. This means success depends not only on differentiation rules but also on translating words, diagrams, and conditions into equations.
• In more advanced mathematics, optimisation extends beyond quadratics to trigonometric, exponential, and rational functions, and may involve endpoints as well as stationary points. The core logic remains the same: define the objective, apply constraints, analyze critical points, and interpret the result in context.
• Outside pure mathematics, differentiation-based optimisation appears in engineering design, economics, physics, and data science. The reason is universal: whenever a system can be described by a changing quantity, derivatives provide a principled way to identify efficient, safe, or optimal choices.