14. Calculus
- Syllabus
- 0606–2028–2029
- Topic
- 14
- Level
- —
The derived function f′(x) gives the gradient of y=f(x) at each value of x. It turns a curve into a new function describing how quickly and in which direction the original function is changing.
Between x and x+Δx, the average gradient is Δy/Δx. As the second point moves closer and Δx tends to 0, the secant gradient approaches the tangent gradient. That limiting gradient is the derivative at x.
For f(x)=x2, the average gradient from x=3 to x=3.1 is (3.12−32)/0.1=6.1. Moving the second point still closer makes this value approach 6, which is f′(3) because the derived function is f′(x)=2x.
A derivative is not one fixed gradient unless the original graph is a straight line. Only an informal limit idea is required here; deriving rules from first principles is outside this syllabus.
Differentiation notation tells you both the order of change and the variables involved. Several symbols can name the same first derivative, but finite changes and limiting change must not be confused.
| Notation | Meaning |
|---|---|
| f′(x) or dxdy | first derivative; gradient/rate of change |
| f′′(x) or dx2d2y | derivative of the first derivative |
| ΔxΔy | average change over a non-zero interval |
| Δx→0limΔxΔy | instantaneous derivative dxdy |
If y depends on x, dy/dx means differentiate y with respect to x. Differentiating again gives d/dx(dy/dx)=d2y/dx2; the square belongs to the derivative order, not to the value dy/dx.
Δy/Δx uses actual finite increments and is generally only an average. The symbols dy and dx belong to derivative notation and should not be replaced casually by finite changes.
Differentiate each standard outer function, then multiply by the derivative of its inner expression. This chain-rule factor is what adjusts the rate for a composite function.
| f(x) | f′(x) |
|---|---|
| xn | nxn−1 for rational n |
| sinx | cosx |
| cosx | −sinx |
| tanx | sec2x |
| ex | ex |
| lnx | 1/x |
\frac{d}{dx}f(g(x))=f'(g(x))g'(x)
For y=(3x2+4)1/3, differentiate the outer power and multiply by 6x: dy/dx=31(3x2+4)−2/3(6x)=2x(3x2+4)−2/3.
Trigonometric derivative rules require angles in radians. Preserve the domain of the original function, especially for rational powers and lnx, and do not omit the inner derivative.
When two changing functions are multiplied or divided, differentiating each part separately is not enough. Keep the product or quotient structure visible until every derivative is in place.
(uv)'=u'v+uv',\qquad \left(\frac uv\right)'=\frac{u'v-uv'}{v^2}
For y=x2e3x, take u=x2 and v=e3x. Then dy/dx=2xe3x+x2(3e3x)=xe3x(2+3x). The chain rule is still needed inside v′.
For a quotient, write the numerator in the order u′v−uv′ before simplifying. Reversing those terms changes the sign. A quotient may instead be rewritten as a product with a negative power when that makes the structure clearer.
The derivative of uv is not u′v′, and the derivative of u/v is not u′/v′. Factor only after the full rule has been applied so that terms are not lost.
At x=a, the derivative f′(a) is the tangent gradient. A normal is perpendicular to the tangent, so its gradient is the negative reciprocal when the tangent gradient is non-zero.
m_{\text{tan}}=f'(a),\qquad m_{\text{normal}}=-\frac1{m_{\text{tan}}},\qquad y-f(a)=m(x-a)
For y=ln(4x−3) at x=2, the point is (2,ln5) and dy/dx=4/(4x−3), so mtan=4/5. The tangent is y−ln5=54(x−2) and the normal is y−ln5=−45(x−2).
Use the coordinates of the point on the curve, not just the derivative. If the tangent gradient is 0, the tangent is horizontal and the normal is vertical, so the negative-reciprocal formula is not written as a finite number.
A stationary point has a horizontal tangent, so solve dy/dx=0 for its x-coordinate and then substitute into the original curve to find its y-coordinate.
Differentiate, factor or otherwise solve the derivative equation, reject values outside the stated domain, and recover every corresponding point from the original function. A derivative equation gives candidate x-values, not complete coordinates.
For y=x3−2x2−4x+8, dy/dx=3x2−4x−4=(3x+2)(x−2). Hence x=−2/3 or x=2, giving stationary points (−2/3,256/27) and (2,0).
This step locates stationary points but does not by itself determine whether each is a maximum or minimum. Points of inflexion are outside this syllabus, so do not add them as a third required category.
A derivative converts a small change or rate in one variable into the corresponding change or rate in another. The same chain relationship controls connected rates and small-increment approximations.
| Situation | Relationship |
|---|---|
| connected rates | dtdy=dxdydtdx |
| small increment | Δy≈dxdyΔx at the starting value of x |
If V=34πr3 and dr/dt=0.5cms−1, then dV/dt=4πr2(dr/dt). At r=0.25cm this is π/8cm3s−1. Units confirm that a volume rate was found.
For y=tanx, dy/dx=sec2x. Increasing x from −π/4 by a small amount h gives Δy≈sec2(−π/4)h=2h.
Evaluate the derivative at the relevant instant or starting value. The small-change relation is an approximation, whereas a connected-rate equation is an instantaneous equality; do not discard signs or units.
A practical maximum or minimum is found only after the constraints have reduced the target quantity to a function of one variable on a realistic domain.
Write the constraint, eliminate extra variables, state the physical domain, differentiate the target function, solve f′(x)=0, and verify that the candidate gives the required maximum or minimum. Finish by reporting the requested dimensions or quantity, not just x.
An open tank with square base x and volume 5 has height h=5/x2 and metal area A=x2+4xh=x2+20/x for x>0. Since A′=2x−20/x2, the stationary value satisfies x3=10. Also A′′=2+40/x3>0, so the tank has minimum area at x=310 and h=5/(310)2.
A stationary value outside the physical domain is invalid. A calculus candidate is not automatically the requested optimum: justify its nature and check any relevant endpoint or feasibility condition.
A maximum or minimum conclusion needs evidence about how the gradient changes, not merely the equation f′(x)=0.
| Test at x=a | Maximum | Minimum |
|---|---|---|
| first derivative | f′ changes +→− | f′ changes −→+ |
| second derivative, given f′(a)=0 | f′′(a)<0 | f′′(a)>0 |
If stationary points occur at x=0.5 and x=3, with f′′(x)=4x−7, then f′′(0.5)=−5<0, so the first is a maximum; f′′(3)=5>0, so the second is a minimum.
If f′′(a)=0, the second derivative test is inconclusive, so use the first derivative test or another valid argument. State the derivative value or sign change and the resulting conclusion; points of inflexion are not included.
Integration reverses differentiation. If F′(x)=f(x), then every function F(x)+C has derivative f(x) because the derivative of a constant is zero.
\int f(x),dx=F(x)+C\quad\Longleftrightarrow\quad F'(x)=f(x)
Because d(2x3)/dx=6x2, ∫6x2dx=2x3+C. The constant is essential: 2x3+1 and 2x3−7 have the same derivative. If a point on the curve is given, substitute it after integrating to determine C.
Sometimes a known derivative exposes a less obvious integral. For example, differentiating xe4x gives e4x+4xe4x, so the relationship can be rearranged to recover an antiderivative involving xe4x.
An indefinite integral is a family of functions and must include C. A definite integral has numerical limits and therefore does not carry an arbitrary constant in its final value.
For powers of x, increase the exponent by 1 and divide by the new exponent. The exponent −1 is the exception and produces a logarithm.
\int x^n,dx=\frac{x^{n+1}}{n+1}+C\ (n\ne-1),\qquad \int\frac1x,dx=\ln|x|+C
\int\frac1{ax+b},dx=\frac1a\ln|ax+b|+C\quad(a\ne0)
Rewrite before integrating. For example, \int\left(4x+5-\frac1{2x+3} ight)dx=2x^2+5x-\frac12\ln|2x+3|+C. Differentiating the result recovers the integrand.
Do not apply the power formula with n=−1, because it would divide by zero. The factor 1/a compensates for the derivative of ax+b; absolute-value bars may be omitted only when the domain guarantees ax+b>0.
For a standard function of ax+b, use its usual antiderivative and divide by the inner derivative a. This is the reverse chain rule.
| Integrand | Antiderivative |
|---|---|
| (ax+b)n, n=−1 | a(n+1)(ax+b)n+1+C |
| 1/(ax+b) | a1ln∣ax+b∣+C |
| sin(ax+b) | −a1cos(ax+b)+C |
| cos(ax+b) | a1sin(ax+b)+C |
| sec2(ax+b) | a1tan(ax+b)+C |
| eax+b | a1eax+b+C |
∫sin((ϕ+π)/3)dϕ=−3cos((ϕ+π)/3)+C, because differentiating the cosine contributes an extra factor 1/3 as well as the negative sign.
Trigonometric integration assumes radians. Check the result by differentiating: a missing reciprocal inner factor is revealed immediately. The case n=−1 uses the logarithmic row, not the power row.
A definite integral gives signed accumulation. A geometric area must be non-negative, so identify intersections and integrate upper function minus lower function on every interval where their order is fixed.
\int_a^b f(x),dx=F(b)-F(a),\qquad A=\int_a^b(\text{upper}-\text{lower}),dx
Solve for all relevant intersection or axis-crossing values, sketch the ordering mentally from the equations, split the interval when the upper curve changes or when a curve crosses the axis, then add the positive area pieces.
Between y=6x−x2 and y=x from x=0 to x=5, the curve is above the line. The area is ∫05(5x−x2)dx=[25x2−31x3]05=125/6.
A negative definite integral is signed area, not a negative physical area. For a region made from two pieces or two curves that exchange order, split at the crossing and add magnitudes rather than allowing cancellation.
For straight-line motion, differentiation moves from displacement to velocity to acceleration; integration moves back and introduces constants fixed by the particle's stated conditions.
v=\frac{ds}{dt},\qquad a=\frac{dv}{dt}=\frac{d^2s}{dt^2},\qquad s=\int v,dt,\qquad v=\int a,dt
If v=6t−12 and s=5 at t=0, then a=6 and s=3t2−12t+C. The initial condition gives C=5, so s=3t2−12t+5. The particle is instantaneously at rest when v=0, at t=2.
The sign of v gives direction. To find total distance over an interval, find every time at which v=0, split the interval there, and add the magnitudes of the displacement changes. Directly integrating velocity gives net displacement, not necessarily distance.
Keep displacement, distance, velocity, speed and acceleration distinct and include units. A zero acceleration means constant velocity, not necessarily zero velocity; a zero velocity at one instant does not mean the particle remains at rest.
A kinematics graph carries information through its slope and signed area. Identify the vertical quantity first: displacement and velocity may be negative, while distance and speed are non-negative.
| Graph | Slope | Area under graph |
|---|---|---|
| displacement-time | velocity | not used here as a standard kinematics quantity |
| distance-time | speed; never negative | not used here as a standard kinematics quantity |
| velocity-time | acceleration | displacement change |
| speed-time | rate of change of speed | distance travelled |
| acceleration-time | not required here | velocity change |
A horizontal displacement-time segment means rest; a straight segment means constant velocity. On a velocity-time graph, crossing the time axis marks an instant of rest and usually a direction change; the gradient gives acceleration, so a horizontal velocity segment has zero acceleration.
Areas below the time axis on a velocity-time graph count negatively towards displacement but positively towards total distance after taking their magnitudes. A distance-time graph cannot decrease, even when the particle reverses direction.
Do not confuse a graph's height with its slope. For example, a velocity of zero is read from the height of a velocity-time graph, whereas zero acceleration is shown by a horizontal tangent or segment on that graph.