Discrete Math › Section E · Sequences › Topic E.2

Arithmetic and geometric sequences

The simplest way for a sequence to change is by the same amount each step, or by the same factor each step. These two patterns, arithmetic and geometric, are the first sequences whose closed formulas we can derive rather than guess, and they mark the line between linear and exponential growth.

Two ways to grow

A gym adds 6 new members every week; a rumour reaches twice as many people every hour. The first story adds a fixed amount each step, the second multiplies by a fixed factor. In the topic "Describing sequences and sums" we saw that a sequence can be pinned down by a recurrence and initial conditions. Here we take the two simplest recurrences, \(a_n = a_{n-1} + d\) and \(a_n = r\,a_{n-1}\), and turn each into a closed formula. Telling which pattern you are looking at, from data or from a description, is the first question to ask of any new sequence.

Arithmetic sequences

A sequence is arithmetic when consecutive terms differ by a constant \(d\), the common difference: \(a_n - a_{n-1} = d\) for every \(n \ge 1\). With first term \(a_0 = a\),

a_n = a_{n-1} + d \quad\text{and}\quad a_n = a + dn

are its recursive definition and closed formula. The list 4, 11, 18, 25 has \(d = 7\), so \(a_n = 4 + 7n\). The difference may be negative: 30, 26, 22 has \(d = -4\). The algebra topic "Arithmetic sequences" wrote \(a_n = a_1 + (n - 1)d\) because it started counting at \(a_1\). Both are correct; they differ only in where the index starts.

Why the formula works: telescoping

Write the defining equation once for each step from 1 to \(n\): \(a_1 - a_0 = d\), \(a_2 - a_1 = d\), and so on up to \(a_n - a_{n-1} = d\). Add all \(n\) equations. The right-hand side is \(dn\). On the left, every term except the first and last appears once with a plus sign and once with a minus sign, so the sum collapses, or telescopes, to \(a_n - a_0\). Hence \(a_n = a_0 + dn\). The argument needs nothing but the constant difference, so it covers every arithmetic sequence at once.

Worked example: an arithmetic sequence

The sequence 4, 11, 18, 25, … starts at \(a_0\). Find a recursive definition, a closed formula, and decide whether 200 and 300 are terms. 1. Differences: \(11 - 4 = 7\), \(18 - 11 = 7\), \(25 - 18 = 7\). Constant, so the sequence is arithmetic with \(d = 7\). 2. Recursive: \(a_n = a_{n-1} + 7\) with \(a_0 = 4\). Closed: \(a_n = 4 + 7n\). 3. Solve \(4 + 7n = 200\): \(n = 196/7 = 28\), a natural number, so \(a_{28} = 200\). 4. Solve \(4 + 7n = 300\): \(n = 296/7\), not a whole number, so 300 is not a term. 5. Check: \(a_3 = 4 + 21 = 25\), as listed.

Geometric sequences

A sequence is geometric when each term is a constant multiple \(r\) of the one before, so the ratio of consecutive terms never changes. With first term \(a_0 = a\),

a_n = r\,a_{n-1} \quad\text{and}\quad a_n = a\,r^n

To see the closed formula, iterate: \(a_1 = ar\), \(a_2 = ar \cdot r = ar^2\), \(a_3 = ar^3\). Reaching \(a_n\) takes exactly \(n\) multiplications by \(r\). The algebra topic "Geometric sequences and series" wrote \(a_1 r^{n-1}\) for the same reason as before. A ratio between 0 and 1 gives decay, and a negative ratio makes the signs alternate.

Worked example: a geometric sequence

Find closed formulas for 5, −15, 45, −135, … and for 96, 48, 24, 12, …, both starting at \(a_0\). 1. First list: \(-15/5 = -3\) and \(45/(-15) = -3\). The ratio is constant, \(r = -3\), so \(a_n = 5(-3)^n\). 2. Check a later term: \(a_6 = 5 \cdot 729 = 3645\), and the sign is positive because 6 is even. 3. Second list: \(48/96 = 24/48 = \frac{1}{2}\), so \(b_n = 96\left(\frac{1}{2}\right)^n\). 4. Check: \(b_3 = 96/8 = 12\), as listed. The terms halve forever and never reach 0.

Compound growth is geometric

Money earning compound interest is the standard real geometric sequence. Put $2,000 into an account paying 3% a year, compounded once a year, and make no further deposits or withdrawals. Each year the balance is multiplied by 1.03, so the balance after \(n\) years is \(a_n = 2000(1.03)^n\). After 10 years it is about $2,687.83, and after 20 years about $3,612.22. The second decade adds more than the first, because 3% is taken of a larger balance each year. Simple interest, which pays 3% of the original $2,000 every year, gives an arithmetic sequence instead: $60 more each year.

Linear against exponential growth

An arithmetic sequence has a linear closed formula, so its graph of points lies on a straight line with slope \(d\). A geometric sequence with \(r > 1\) and a positive first term grows exponentially, and its differences, \(a r^{n+1} - a r^n = a r^n (r - 1)\), grow too. Exponential growth always overtakes linear growth eventually. Compare \(100 + 50n\) with \(2^n\): at \(n = 9\) they are 550 and 512, but at \(n = 10\) they are 600 and 1,024, and the gap only widens. Starting values decide who leads at first; the type of growth decides who wins.

Measuring growth with big-O

Computer scientists compare growth with big-O notation. For positive sequences, \(f(n) = O(g(n))\) means that some constant multiple of \(g(n)\) is at least \(f(n)\) for all large enough \(n\). So \(100 + 50n = O(n)\), since \(150n \ge 100 + 50n\) once \(n \ge 1\). Constant factors and starting values are ignored; only the rate of growth counts. The common classes, from slowest to fastest growing, are \(O(1)\), \(O(\log n)\), \(O(n)\), \(O(n \log n)\), \(O(n^2)\) and \(O(2^n)\). An algorithm whose step count is geometric in \(n\) becomes hopeless on large inputs.

Neither arithmetic nor geometric

Plenty of sequences fit neither pattern. The triangular numbers 1, 3, 6, 10, 15 have differences 2, 3, 4, 5, which are not constant, and ratios 3, 2, about 1.67, which are not constant either. But the differences themselves form an arithmetic sequence. That observation is the doorway to the next topic, "Polynomial sequences", which shows that a sequence whose differences are arithmetic, with a non-zero common difference, has a quadratic closed formula. The test order is always the same: compute differences, then ratios, and if neither is constant, look at the differences of the differences.

Common mistakes to avoid

First, mixing the two index conventions: if the list starts at \(a_0\), the closed formula is \(a + dn\), not \(a + (n-1)d\). Second, testing for a geometric sequence by subtracting; the test is division. Third, dropping the sign of a negative ratio, which hides the alternating signs. Fourth, calling a sequence arithmetic after checking only one difference: check several. Fifth, judging growth by the first few terms: \(100 + 50n\) leads \(2^n\) for a while, yet the exponential wins from \(n = 10\) on.

Key takeaways

Arithmetic: constant difference \(d\), recurrence \(a_n = a_{n-1} + d\), closed formula \(a_n = a_0 + dn\), proved by telescoping. Geometric: constant ratio \(r\), recurrence \(a_n = r\,a_{n-1}\), closed formula \(a_n = a_0 r^n\), found by iterating. Formulas starting at \(a_1\) shift \(n\) to \(n - 1\). Arithmetic growth is linear, geometric growth with \(r > 1\) and a positive start is exponential and always wins in the end, and big-O notation compares such rates while ignoring constants.

Flashcards

Flashcard

Arithmetic sequence

Consecutive terms differ by a constant d

Test by subtracting neighbours.

Test your understanding

Question 1 of 5

The sequence 7, 3, −1, −5, … starts at a₀. What is its closed formula?

Question 2 of 5

A geometric sequence has a₀ = 3 and common ratio 2. What is a₅?

Question 3 of 5

The list 2, 5, 8, 11, … is indexed so that its first term is a₁ = 2. Which closed formula fits?

Question 4 of 5

Which of these sequences is geometric?

Question 5 of 5

A geometric sequence has a₀ = 5 and a₃ = 40. What is its common ratio?

← Previous topicNext topic →