Algorithm portfolio selection as a bandit problem with unbounded losses

DOI10.1007/S10472-011-9228-ZMaRDI QIDQ408989zbMATHOpenAlexFDO

Authors Matteo Gagliolo, Jürgen Schmidhuber

Publication date 12 April 2012

Published in Annals of Mathematics and Artificial Intelligence (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/s10472-011-9228-z

zbMATH Keywords

computational complexitycombinatorial optimizationsurvival analysisconstraint programmingonline learningsatisfiabilityalgorithm selectionmulti-armed bandit problemalgorithm portfoliosLas Vegas algorithmsmeta learning

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.) (68T20) Online algorithms; streaming algorithms (68W27) Analysis of algorithms and problem complexity (68Q25) Nonparametric inference (62G99) Survival analysis and censored data (62N99)

Recommendations

• Learning dynamic algorithm portfolios
• Algorithm portfolios
• Online algorithms for the portfolio selection problem. With a foreword by Günter Schmidt
• Real-time solving of computationally hard problems using optimal algorithm portfolios
• A PAC Approach to Application-Specific Algorithm Selection

Cites work

• scientific article; zbMATH DE number 700091 (Why is no real title available?)
• scientific article; zbMATH DE number 2000819 (Why is no real title available?)
• Algorithm portfolios
• Algorithm survival analysis
• Combining Multiple Heuristics
• Hannan Consistency in On-Line Learning in Case of Unbounded Losses Under Partial Monitoring
• Heavy-tailed phenomena in satisfiability and constraint satisfaction problems
• Learning Theory
• Learning dynamic algorithm portfolios
• Learning parallel portfolios of algorithms
• Machine Learning: ECML 2004
• Minimizing Regret With Label Efficient Prediction
• Nonparametric Estimation from Incomplete Observations
• Optimal speedup of Las Vegas algorithms
• Principles and Practice of Constraint Programming – CP 2004
• Reactive search and intelligent optimization
• SATzilla: portfolio-based algorithm selection for SAT
• Some aspects of the sequential design of experiments
• Stochastic local search. Foundations and applications.
• Survival analysis. Techniques for censored and truncated data.
• The Nonstochastic Multiarmed Bandit Problem
• The weighted majority algorithm

Cited in

(9)

• Variable ordering for decision diagrams: a portfolio approach
• Analyzing bandit-based adaptive operator selection mechanisms
• Real-time solving of computationally hard problems using optimal algorithm portfolios
• Gorthaur-EXP3: bandit-based selection from a portfolio of recommendation algorithms balancing the accuracy-diversity dilemma
• Learning dynamic algorithm portfolios
• Machine Learning: ECML 2004
• \textsc{Alors}: an algorithm recommender system
• Multi-armed bandits with censored consumption of resources
• Parallel algorithm portfolio with market trading-based time allocation

Describes a project that uses

Uses Software

• SATzilla
• SPOT

This page was built for publication: Algorithm portfolio selection as a bandit problem with unbounded losses

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q408989)