Computer Science 789 Randomized Online Algorithms

Fall 2003

The material of this course is based on work supported by National Science Foundation Grant CCR-0312093.

• Instructor: Dr. Larmore Office, TBE B378B. Telephone, 702-895-1096
• Office Hours: to be announced
• Room and Time: TBE B365, Tuesdays and Thursdays, 5:30 PM to 6:45 PM.
• Days of Instruction: August 26, 2003 to December 4, 2003
• Textbook: ONLINE COMPUTATION AND COMPETITIVE ANALYSIS by Allan Borodin and Ran El-Yaniv.
• Lecture Topics and Assignments
• Written Homework: will be assigned, collected, and graded. You are permitted to discuss homework with each other or with others, but you must write the assignment by hand, not type or print from a computer file.

In an online problem, you are required to make decisions without knowing all the inputs.

In most computer problems, you are given all the inputs at the beginning. We call those offline problems.

Life is an online problem. For example, you must make decisions today without knowledge of what tomorrow will bring.

Algorithms are, by definition, deterministic. This means that, given certain inputs, the algorithm will make a predictable choice. A randomized algorithm, on the other hand, given certain inputs, will make a given choice with a predictable probability.

I will give an introduction to this fascinating new area.

• Old Lecture Topics and Assignments
• Old Handouts.

• Berkeley-1
• Berkeley-2
• Carnegie Mellon
• MIT
• Irvine
• Illinois
• Frankfurt
• Iowa State
• West Virginia University
• Dartmouth
• Johns Hopkins
• Aachen
• Technion
• Brics