Review ol ~ Online Computation and Competitive Analysis Authors" Allan Borodin and R a n EI-Yaniv Publisher: Cambridge University Press Hardcover: ISBN 0-521-56392-5 414 pages Reviewer: Neal E. Young Overview Since roughly 1985, theoretical research in the competitive analysis of online algorithms (a subfield of theoretical computer science) has grown tremendously. This book by Borodin and EI-Yaniv is the first comprehensive guide to this body of research. The book covers some of the fund~.mental problems that have been considered during this period: list accessing, paging, the k-server problem, metrical task systems, as well as some more application-oriented topics: in scheduling (bin packing, virtual circuit routing, load balancing) and in investment (portfolio selection). The authors use these problems to summarize the important results in the field (including proofs), as well as to introduce general methods, including the use of potential functions, randomized algorithms (very important in competitive analysis!), and g~.rne theory. The final chapter considers competitive analysis in the general context of decision making under uncertainty and formally compares competitive analysis to other theoretical approaches. Each of the 15 chapters contains a few (typically 5-10) exercises, usually having the reader expand on, or prove some detail of, the analysis
/lp/association-for-computing-machinery/review-of-online-computation-and-competitive-analysis-by-allan-borodin-0br8cqCbBF