Combinatorial Network Optimization With Unknown Variables: Multi-Armed Bandits With Linear Rewards and Individual Observations

• Published in: IEEE/ACM transactions on networking Vol. 20; no. 5; pp. 1466 - 1478
• Main Authors: Yi Gai, Krishnamachari, B., Jain, R.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2012; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithm design and analysis; Algorithms; Combinatorial analysis; Combinatorial network optimization; Computation; Formulations; Indexes; Mathematical analysis; Mathematical models; multi-armed bandits (MABs); Networks; online learning; Optimization; Policies; Polynomials; Radiation detectors; Random variables; Studies; Upper bound

We formulate the following combinatorial multi-armed bandit (MAB) problem: There are N random variables with unknown mean that are each instantiated in an i.i.d. fashion over time. At each time multiple random variables can be selected, subject to an arbitrary constraint on weights associated with t...