Numerical Series
A series is the sum of infinitely many terms. When does it converge? Criteria to determine this.
Complete Theory
2A numerical series is the sum of infinitely many terms: . The partial sum is the sum of the first terms. The series converges to if (exists and is finite). If the limit of partial sums does not exist or is infinite, the series diverges.
Necessary (but not sufficient) condition for convergence: . Warning: it is not sufficient! The harmonic series diverges even though .
Geometric series: for ; diverges for . The partial sum is .
Riemann -series: converges if and only if . Proof via the integral test: converges iff .
Absolute convergence: if converges, then converges (absolutely). Absolute convergence implies ordinary convergence, but not vice versa (e.g., alternating harmonic series converges conditionally but not absolutely).
Ratio test (D'Alembert): . Then:
- : series converges absolutely
- (or ): series diverges
- : test is inconclusive
Especially useful for series with factorials () or exponentials ().
Root test (Cauchy): . Same three conclusions as the ratio test. Often easier when involves -th powers.
Comparison test: if for all sufficiently large , then: converges converges; diverges diverges.
Limit comparison test: if and , then and have the same behaviour (both converge or both diverge).
Leibniz alternating series test: converges if is monotonically decreasing and . Convergence is conditional (not absolute if diverges). Moreover (error bound).
Worked Examples
1Exercises with Solutions
1Keep studying
Guided exercises on this topic
Recommended Books
As an Amazon Associate I earn from qualifying purchases.