Stochastic Multi-armed Bandit With Knapsack and Its Application in Wireless Edge Computing
#wireless
#networks
#communications
Multi-armed bandits (MAB) is a popular sequential optimization technique for distributed decision making under uncertainty given no prior knowledge of the environment. It uses the history of previous decisions and observations as well as side information (if available) to arrive at the current decision. Different from traditional bandits, bandits with knapsacks (BwK) also consider global constraints in the sequential optimization process. In this talk, I will discuss the BwK model in general and its application to the server selection problem for computation offloading in a wireless network. Time permitting, extension of the model to linear contextual bandits will be also discussed.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
- 99111 Ring Road
- Victoria, British Columbia
- Canada V8P5C2
- Building: Engineering and Computer Science Building
- Room Number: ECE 660
- Click here for Map