Online Algorithms and Advice Complexity

posted Aug 10, 2015, 4:30 AM by Jhoirene Clemente   [ updated Aug 10, 2015, 5:15 AM ]
What: Lecture on Online Algorithms and Advice Complexity
When: 10:00 AM,  August 27, 2015
Where: Algorithms and Complexity Laboratory, Department of Computer Science, University of the Philippines Diliman

Many real world problems arise in so-called online environments. In these environments, information arrives in a piecewise manner and decisions are needed to be made without incomplete information about the future. Problems arising in such environments are called online problems and algorithms solving these problems are called online algorithms. In contrast, problems with complete information of the input are called offline algorithms.Not knowing the whole input is a huge disadvantage for online problems and so optimality is not always guaranteed even for problems that are polynomial-time solvable in the offline setting. Some examples of online problems are the paging problem, the ski-rental problem, and the time-series search problem. Algorithms for these problems are analyzed using competitive ratio, which is the worst ratio between the cost of the online  algorithm with respect to the offline counterpart. In this lecture, we will provide an introduction to online algorithms and competitive analysis. We will also provide an overview of advice complexity, which is used to measure the minimum amount of information needed by an online algorithm to be optimal. We will provide in this lecture some interesting questions about the advice complexity of online algorithms.