Spyros Angelopoulos

Biography

Spyros Angelopoulos studies how algorithms should act when decisions cannot wait for all the facts. A prediction about future requests can help a system allocate resources or choose where to search, but a poor forecast can make it perform worse. He designs online algorithms and proves how much they can gain from good advice while limiting the cost of bad advice.

Research topics

• Online algorithms
• Algorithms with predictions
• Competitive analysis