I have an alternating sum and am looking for its asymptotic behaviour. Due to the change of signs there occure a lot of cancellations between the different terms. However the terms are all different, so no terms cancel exactly. Is there still a method that can be used to find the dominating behaviour of the sum?
Thanks