We live in a world where information is highly abundant, for example, web pages, Wikipedia articles, Facebook contacts, etc. In order to be able to search for relevant information we need to be able to rank it according to its importance or popularity. Search engines such as Google have algorithms that rank webpages according to specific attributes, and it is this rank that determines the ordering of search results. Ranking algorithms are designed with specific goals in mind, for example, the ranking of news blogs should depend on the trustworthiness of the source. Nonetheless, algorithms do not necessarily behave as expected, and there is a need to analyze their large scale behavior. In this talk we use OR to model and analyze a special class of ranking algorithms closely related to Google's PageRank. We will set up a mathematical model to understand what makes highly ranked pages important, and we provide the tools to improve existing algorithms and design new algorithms with specific characteristics.
Copyright. All Rights Reserved.