Business calculus guide and calculator

1. Functions, units and the business model

A function assigns one output to each permitted input. If x is units sold per month, C(x) is monthly cost, p(x) is price per unit and R(x)=x·p(x) is monthly revenue. Profit is P(x)=R(x)−C(x). A model's domain is part of its meaning: negative sales or production above capacity may be mathematically possible but commercially irrelevant.

Keep dollars, units and time periods consistent. Fixed cost is the part independent of x within the model's relevant range; it is not necessarily unavoidable forever. Average cost C(x)/x requires x>0. Solve P(x)=0 for break-even candidates, then check the feasible domain.

Worked example: building a model

Suppose C(x)=600+8x and p(x)=20. Then R(x)=20x and P(x)=12x−600. At 80 units, revenue is $1,600, cost is $1,240 and profit is $360. Break-even is x=50 units. This assumes all units sell at the stated price and no costs were omitted.

Before calculus, review equations, fractions, powers, logarithms and function composition. ln is the natural logarithm; log in this workspace means base 10. Use * for multiplication, ^ for powers and parentheses for grouping.

Try this topic in the calculator → · Back to contents

2. Read graphs before trusting a calculation

Intercepts show where a function meets an axis. A vertical asymptote can signal an excluded input; a horizontal asymptote describes long-run behavior when the relevant limit exists. Increasing, decreasing and concave behavior describe different aspects of a curve.

A graph is a picture of samples, not proof. A line across a gap does not establish continuity, and a zoomed view can hide roots or turning points. Check the formula's domain and the numerical scale. A cost curve can be increasing while its marginal cost decreases.

Worked example: average fixed cost

With C(x)=600+8x, average cost is A(x)=600/x+8 for x>0. As production grows, A(x) approaches $8 per unit from above. It never equals $8 at a finite positive x. The model is undefined at x=0; the vertical behavior does not mean producing zero units has infinite total cost.

Explore explicit functions with the graphing calculator. Plot 600/x+8 and inspect positive x. Its function graphing does not automate symbolic asymptotes or implicit curves.

Try this topic in the calculator → · Back to contents

3. Limits and continuity

A limit describes the value a function approaches near an input, not necessarily its value at that input. A two-sided finite limit needs agreement from both sides. Continuity at a requires f(a) to exist, the limit to exist, and both values to agree.

For rational functions, factor and cancel only for inputs where the original denominator is nonzero. Cancellation may reveal a limit while leaving a hole in the original domain. Infinite limits, oscillation and one-sided domains need separate reasoning.

Worked example: a removable hole

For f(x)=(x^2−9)/(x−3), factor the numerator to obtain (x−3)(x+3). When x≠3, f(x)=x+3. Therefore the limit as x approaches 3 is 6, but the original function is undefined at 3. Defining f(3)=6 would fill the hole and make that extension continuous.

Use Limit exploration to inspect nearby values. The sample table can support intuition but cannot prove a limit; finite samples can miss rapid oscillation or a narrow singularity.

Try this topic in the calculator → · Back to contents

4. Derivatives and tangent lines

The derivative f′(x) is the limit of [f(x+h)−f(x)]/h as h approaches zero, if that limit exists. It measures instantaneous change. The secant slope over a finite interval is an average rate; a derivative is a local rate.

Rules: a constant differentiates to zero; (x^n)′=n·x^(n−1); sums differentiate term by term. For products use (uv)′=u′v+uv′; for quotients use (u/v)′=(u′v−uv′)/v^2 wherever v≠0. The chain rule gives [g(u(x))]′=g′(u(x))·u′(x).

Worked example: tangent and local estimate

C(x)=250+12x+0.05x^2 has derivative C′(x)=12+0.1x. At x=20, C=510 and C′=14. The tangent is L(x)=510+14(x−20), or 14x+230. It predicts C(21)≈524; the exact model gives 524.05. The 0.05 difference is the approximation error for this one-unit step.

The on-page Derivatives method displays supported symbolic derivatives plus their point values. It preserves the original function's domain check at the selected point. Symbolic expressions are not necessarily reduced to the shortest form.

Try this topic in the calculator → · Back to contents

5. Marginal cost, revenue, profit and average cost

If C, R and P are totals in dollars and x measures units, C′, R′ and P′ have units dollars per additional unit. Marginal profit satisfies P′=R′−C′. Marginal cost approximates C(x+1)−C(x), but the two are not generally identical.

If price depends on quantity, revenue is x·p(x), so R′(x)=p(x)+x·p′(x). Forgetting the second term overstates marginal revenue for a downward-sloping price curve. Average cost C(x)/x differs from marginal cost; its derivative is [xC′(x)−C(x)]/x^2.

Worked example: marginal revenue

Let p(x)=50−0.1x. Then R(x)=50x−0.1x^2 and R′(x)=50−0.2x. At x=100, the price is $40 and marginal revenue is $30 per unit. Selling another unit reduces the assumed price on the modeled sales, so marginal revenue is below price.

Use Derivatives with a cost, revenue or profit formula, then inspect the exact change using h=1. Numerical examples illustrate a supplied model; they do not establish market demand.

Try this topic in the calculator → · Back to contents

6. Optimization, curvature and feasible boundaries

For a differentiable function, interior stationary candidates satisfy f′=0. A positive second derivative at a stationary point indicates a strict local minimum; a negative value indicates a strict local maximum. A zero second derivative is inconclusive. Changes in the sign of f′ can classify behavior; changes in concavity help identify inflection points.

For a continuous function on a closed bounded interval, compare feasible stationary points, nondifferentiable candidates and endpoints. A local optimum need not be global. If x is a whole-number production quantity, compare neighboring feasible integers rather than automatically rounding the continuous optimum.

Worked example: a concave profit function

P(x)=180x−0.9x^2−1,500 on 0≤x≤180 has P′=180−1.8x and P″=−1.8. The stationary point is x=100. Profit there is $7,500, compared with −$1,500 at zero and $1,740 at 180. Because this quadratic is concave throughout the interval, x=100 is the global maximum within that interval.

Our bounded numerical scan finds candidates and compares endpoints; it does not prove it found every stationary point. The analytic concavity argument above supplies the stronger conclusion for this particular example.

Try this topic in the calculator → · Back to contents

7. Exponential and logarithmic models

For continuous growth, A(t)=A0·exp(kt). Its derivative is kA(t), so k is the proportional growth rate per unit time. Decay uses negative k. This is different from discrete growth A0(1+r)^t, though the two can be matched with k=ln(1+r).

Rules: [exp(u)]′=exp(u)·u′; [ln(u)]′=u′/u on u>0; [log10(u)]′=u′/[u·ln(10)]. For a positive base a, [a^x]′=a^x·ln(a). The chain rule is needed when the exponent or log argument is itself a function.

Worked example: continuous growth

A(t)=2,000·exp(0.04t) yields approximately $2,983.65 after ten years. Its instantaneous growth at t=10 is 0.04·A(10), approximately $119.35 per year. Four percent continuous growth corresponds to an effective one-year increase of exp(0.04)−1≈4.0811%.

The differential equation A′=kA has this exponential solution. A logistic model A′=kA(1−A/K) adds a limiting scale K, but the suitability of either model must be justified by evidence. This workspace evaluates supplied formulas; it does not estimate parameters or solve arbitrary differential equations.

Try this topic in the calculator → · Back to contents

8. Price elasticity of demand

When Q(p) is demand as a function of positive price, the nonnegative convention for ordinary downward demand is E(p)=−pQ′(p)/Q(p). E>1 is elastic, E<1 is inelastic, and E=1 is unit elastic. Some texts instead report the signed elasticity pQ′/Q; check the convention before comparing answers.

For revenue R(p)=pQ(p), R′(p)=Q(p)[1−E(p)]. Under the stated differentiable demand model, a small price increase raises revenue when demand is inelastic and lowers it when demand is elastic. This does not establish a profit optimum because costs may also change.

Worked example: linear demand

Let Q(p)=1,200−20p. At p=30, Q=600 and Q′=−20, giving E=1. At p=40, Q=400 and E=2, which is elastic. The meaningful domain requires positive price and quantity; the line's negative-demand extension is not a market prediction.

For the Elasticity method, use x for price and enter Q(x), not inverse demand p(x). The surplus method below uses the opposite convention.

Try this topic in the calculator → · Back to contents

9. Implicit differentiation and related rates

An implicit relation F(x,y)=0 may define y locally as a function of x. Where Fy≠0 and the relation is differentiable, dy/dx=−Fx/Fy. First check that the supplied point lies on the intended curve.

For a related-rates problem, differentiate with respect to time. If x and y both vary, Fx·dx/dt+Fy·dy/dt=0, so dy/dt=−(Fx/Fy)·dx/dt. Describe units and whether a rate is increasing or decreasing. A missing rate is not zero.

Worked example: a level curve

For x^2+y^2=25 at (3,4), Fx=6 and Fy=8. Thus dy/dx=−3/4. If dx/dt=2 units per minute, dy/dt=−1.5 units per minute. In the tool enter F=x^2+y^2−25 and verify its displayed level is zero.

The workspace computes derivatives at a supplied point and displays the relation's level. It does not solve for y or prove an implicit branch exists where Fy=0. For applications involving several independent changing quantities, write and differentiate the full relation before substitution.

Try this topic in the calculator → · Back to contents

10. Antiderivatives and the Fundamental Theorem

An antiderivative F satisfies F′=f on an interval. The indefinite integral is F(x)+C, where C is an arbitrary constant. If f is continuous on [a,b], the Fundamental Theorem gives ∫a^b f(x)dx=F(b)−F(a). The definite integral is signed accumulation, not always geometric area.

Basic rules: ∫x^n dx=x^(n+1)/(n+1)+C for n≠−1; ∫1/x dx=ln|x|+C on intervals excluding zero; ∫exp(kx)dx=exp(kx)/k+C for k≠0. Integrate sums term by term. Always differentiate a proposed antiderivative to check it.

Worked example: recover cost from marginal cost

If C′(x)=8+0.2x, then C(x)=8x+0.1x^2+K. Given C(0)=300, K=300. Cost at 50 units is 400+250+300=$950. The integral from zero to 50 of marginal cost is $650; it excludes the given fixed cost.

The symbolic antiderivative mode supports a defined rule set, not arbitrary algebraic integration. The definite-integral mode uses numerical Simpson integration and requires a finite continuous integrand on the entire interval.

Try this topic in the calculator → · Back to contents

11. Substitution, integration by parts and improper cases

Substitution reverses the chain rule. Choose u=g(x) and include du=g′(x)dx. For a definite integral, transform the bounds or return to the original variable before evaluating; do not mix conventions.

Worked example: substitution

For ∫2x·exp(x^2)dx, take u=x^2 and du=2x dx. The result is exp(x^2)+C. From x=0 to x=1 the integral is e−1≈1.71828. Differentiating exp(x^2) confirms the integrand.

Integration by parts reverses the product rule: ∫u dv=uv−∫v du. For ∫x·exp(x)dx choose u=x, dv=exp(x)dx; the result is (x−1)exp(x)+C. From zero to one, the value is 1.

An improper integral needs a limiting argument at an infinite bound or singular point. Do not integrate across a pole as if the function were continuous. ∫−1^1 1/x dx is not an ordinary convergent improper integral, even though a symmetric principal-value procedure can produce zero. The calculator rejects detected domain failures and does not support improper integrals. It cannot detect every narrow singularity from samples.

Numerical integration helps check a definite answer on a valid finite interval; it does not supply a symbolic proof or automate all substitution and by-parts reasoning.

Try this topic in the calculator → · Back to contents

12. Accumulation, average value and area between curves

If f(t) is a rate in units per hour, its integral over hours gives accumulated units. The average value of a continuous function on [a,b] is (1/(b−a))∫a^b f(x)dx. This is an interval average, not necessarily the arithmetic mean of endpoint values.

Worked example: changing production rate

A rate r(t)=40+6t over a five-hour period produces ∫0^5(40+6t)dt=200+75=275 units. The average rate is 275/5=55 units per hour. The endpoint average happens to agree because this rate is linear; that shortcut is not general.

For area between curves, integrate upper minus lower on each interval where their order is known. If they cross, split at intersections and reverse the subtraction where needed. Integrating a signed difference across the whole interval can cancel positive and negative regions.

Example: between y=x and y=x^2 on [0,1], x is the upper curve. Area is ∫0^1(x−x^2)dx=1/2−1/3=1/6 square units. Use Definite integral with x-x^2 to check it.

Try this topic in the calculator → · Back to contents

13. Consumer and producer surplus

Use inverse demand D(q) and supply S(q): both return price per unit from quantity q. At a competitive intersection D(q*)=S(q*)=p*, the standard model defines consumer surplus ∫0^q* D(q)dq−p*q* and producer surplus p*q*−∫0^q* S(q)dq.

Worked example: linear curves

Let D(q)=80−q and S(q)=20+0.5q. Their intersection has 80−q=20+0.5q, giving q*=40 and p*=40. Consumer surplus is (80+40)·40/2−40·40=$800. Producer surplus is 40·40−(20+40)·40/2=$400.

Enter x for quantity, demand as 80-x, supply as 20+0.5*x, bounds zero and 40, selected quantity 40 and price 40. The tool checks sampled curve ordering but does not automatically solve equilibrium. If the supplied quantity and price are not equilibrium, interpret the output only as conditional surplus areas under the supplied model.

These areas rely on the model and allocation assumptions. They are not a complete welfare analysis, evidence of willingness to pay, or a forecast of realized business profit.

Try this topic in the calculator → · Back to contents

14. Continuous cash flows and present value

A continuously paid income rate r(t), measured in dollars per year, discounted at continuous annual rate k, has present value PV=∫0^T r(t)exp(−kt)dt. If the income rate is constant r and k≠0, PV=r(1−exp(−kT))/k. When k=0, PV=rT.

Worked example: a continuous income stream

A hypothetical $1,000 per year received continuously for three years at continuous discount rate 5% has PV=1,000(1−exp(−0.15))/0.05≈$2,785.84. If an initial investment is $2,500, subtract it separately to obtain approximately $285.84 net present value under these assumptions.

In the Continuous present value method, enter f(x)=1000, bounds zero and three, and rate 5. The tool integrates the discounted rate; it does not automatically deduct an initial investment or infer a tax treatment.

Annual year-end payments of $1,000 have different timing and therefore a different present value. For discrete cash flows, use the Capital Budgeting & NPV Calculator. Do not mix a continuous discount rate with an effective annual rate without conversion: k=ln(1+i).

Try this topic in the calculator → · Back to contents

15. Two-variable models and constrained decisions

For f(x,y), a partial derivative changes one variable while holding the other fixed. The gradient is (fx,fy). The differential df≈fx dx+fy dy estimates a small combined change. Second partials form the Hessian.

Worked example: an unconstrained local maximum

Let P(x,y)=100x+80y−x^2−y^2. Then Px=100−2x and Py=80−2y. At (50,40), both vanish. Pxx=Pyy=−2 and Pxy=0, so the Hessian determinant is 4 with Pxx<0. The second-derivative test gives a strict local maximum; this globally concave quadratic also has its global maximum there.

At a stationary point, determinant D=fxx fyy−fxy^2: D>0 with fxx>0 gives a local minimum; D>0 with fxx<0 gives a local maximum; D<0 gives a saddle. D=0 is inconclusive. A nonstationary point cannot be classified by that test.

For a smooth equality g(x,y)=c with nonzero constraint gradient, Lagrange candidates satisfy ∇f=λ∇g together with g=c. Example: maximize xy with x+y=20 and x,y≥0. Equations y=λ and x=λ yield x=y=10; xy=100, exceeding the boundary value zero. Our constrained mode checks residuals at a supplied candidate; it does not solve all candidates or certify the constrained optimum.

Additional constraints, inequality cases, corners, scaling and mixed units need explicit treatment. Use the algebra workspace for supported small linear systems; nonlinear constrained systems require additional analysis.

Try this topic in the calculator → · Back to contents

Choose the right BusCalc tool

• Graphing calculator: explicit function plots and value tables.
• Algebra and matrix calculator: quadratics, small linear systems and matrices.
• Scientific calculator: expression arithmetic and scientific functions.
• Financial calculator: periodic time value of money.
• Capital Budgeting & NPV: discrete project cash-flow comparisons.
• Business finance calculators: practical pricing, break-even and cash-flow worksheets.
• Decision Feasibility Calculator: evaluate supplied business constraints and uncertainty after choosing an appropriate model.

References, coverage and independence

Coverage was checked against publicly accessible topic outlines and open educational references. BusCalc's lessons, numerical examples, quizzes and software are original; no textbook exercises, answer keys, restricted course material or publisher artwork were reproduced. Links are for further study and do not imply endorsement, affiliation, accreditation or syllabus equivalence. Follow your own course's notation and permitted-calculator rules.

• OpenTextBookStore: Business Calculus topic overview
• Coast Community College District: published business-calculus course outline
• Open Textbook Library: Business Calculus with Excel catalog and topic list
• OpenStax Calculus Volume 1: derivatives and integrals
• OpenStax Calculus Volume 3: multivariable calculus

Prepared November 2, 2026. Broad introductory coverage is not every possible business-calculus problem. Verify consequential work and ask an instructor or qualified reviewer when a method or assumption is uncertain.