Comparison Tests for Positive Term Series    ∑ an                                                                                                      

 

 

In a Nut Shell:  One strategy to evaluate the convergence or divergence of an infinite

series is to compare it with a known series that either converges or diverges.  i.e. The

harmonic series is known to diverge so it can be used to test divergence.  The p-series

and geometric series are candidates to check for either convergence or divergence.

Click here for a reminder of these series.

 

 

Two comparison tests, the Comparison Test and the Limit Comparison Test, are

available to test positive term series for convergence or divergence. They provide

no information about the sum of the series.

 

 

Strategy:  Compare one series, ∑ bn , where it is known to either converge or

diverge with the series to be tested, ∑ an  .

 

Suggestion:  For the test pick ∑ bn that converges if you think  ∑ an  converges. 

Likewise, for the test pick ∑ bn that diverges if you think  ∑ an  diverges.

 

 

Comparison Test

 

Suppose     0     an         bn      for all n;   If     bn   converges,  then   ∑ an    converges

 

Suppose      0     bn         an      for all n;   If     bn   diverges,  then   ∑ an    diverges

 

 

Limit Comparison Test (Interested in large values of n)

 

Suppose    an      0,    bn  >  0   for all n   and

            lim ( an / bn  )  =  L       (an   and  bn    are sequences )  and  L  is a finite number
          n → ∞     
                                                                    
     If    L  > 0,  then the series      ∑ an   and   bn    either
                                                  
n = 1               n = 1

     both series converge or both series diverge.

    If    L  =  0 ,   an ‘s  are much smaller than  bn ‘s  for large  n

                                                               
    If    L  =  0,    bn    converges,  then   ∑ an    converges.
                         n =1                                n = 1

 

 

Click here for examples.

 




Copyright © 2019 Richard C. Coddington

All rights reserved.