Analysing deletions in competitive self-adjusting linear list algorithms

Abstract

There has been a great deal of work on self-adjusting algorithms for linear lists. Almost all of the prior work focused only on successful searches, but in [3] we designed and analysed self-adjusting algorithms for a linear list which were efficient for both successful and unsuccessful searches as well as insertions. Analysis of deletions is listed as an open question. This paper presents an improved version of MP which is also able to handle deletions efficiently, and proves that the new MP algorithm is competitive to offiine adversaries when considering successful searches, unsuccessful searches, insertions, and deletions.