Memory Versus Randomization in On-line Algorithms (Extended Abstract)

  title={Memory Versus Randomization in On-line Algorithms (Extended Abstract)},
  author={Prabhakar Raghavan and Marc Snir},
The amortized analysis of on-line algorithms is of interest in a variety of situations [1, 6, 7]. An application of great practical importance is the caching problem [4, 7] the replacement algorithm for lines in a cache, or pages in a virtual memory. The standard algorithm suggested in the literature is Least Recently Used (LP~U). The analysis of LRU [7] shows that it achieves optimal worst-case performance (this will be made precise in section 2).