Improved Features for Runtime Prediction of Domain-Independent Planners
Explore this paper's citation graph
Summary
This work proposes a new, extensive set of instance features for planning, and investigates its effectiveness across a range of model families, and concludes that its models predict runtime much more accurately than the previous state of the art.
- Type
- article
- Published
- 2014-05-11
- Cited by
- 56
- References
- 28
- Access
- Open access
- OpenAlex
- https://openalex.org/W49698117
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:8533013
Keywords
Benchmark (surveying), Planner, Computer science, Set (abstract data type), Domain (mathematical analysis)
References
- In defense of PDDL axioms
- The FF Planning System: Fast Plan Generation Through Heuristic Search
- Planning Through Stochastic Local Search and Temporal Action Graphs in LPG
- Analyzing Search Topology Without Running Any Search: On the Connection Between Causal Graphs and h+
- Algorithm runtime prediction: Methods & evaluation
- Learning from planner performance
- Empirical hardness models: Methodology and a case study on combinatorial auctions
- How to Solve It Automatically: Selection Among Problem Solving Methods
- SATzilla: Portfolio-based Algorithm Selection for SAT
- The Metric-FF Planning System: Translating ''Ignoring Delete Lists'' to Numeric State Variables
- Portable high-performance superconducting: high-level platform-dependent optimization
- The Fast Downward Planning System
- What Makes Planners Predictable?
- Learning Portfolios of Automatically Tuned Planners
- Mining IPC-2011 Results
- Searching for Plans with Carefully Designed Probes
- Understanding Random SAT: Beyond the Clauses-to-Variables Ratio
- Learning the Empirical Hardness of Optimization Problems: The Case of Combinatorial Auctions
- Arvand : the Art of Random Walks
- Exploiting Competitive Planner Performance
Cited by
- Planning through Automatic Portfolio Configuration: The PbP Approach
- On the Online Generation of Effective Macro-Operators
- Portfolio-based planning: State of the art, common practice and open challenges
- Identifying and Exploiting Features for Effective Plan Retrieval in Case-Based Planning
- ASlib: A benchmark library for algorithm selection
- Multiagent plan repair by combined prefix and suffix reuse
- Feature Selection for SUNNY: A Study on the Algorithm Selection Library
- Statistical Regimes and Runtime Prediction
- Learning and Tuning Meta-heuristics in Plan Space Planning
- The IBaCoP Planning System: Instance-Based Configured Portfolios
- Static and Dynamic Portfolio Methods for Optimal Planning: An Empirical Analysis
- Algorithm Selection for Combinatorial Search Problems: A Survey
- Model-lite planning: Case-based vs. model-based approaches
- Efficient Parameter Importance Analysis via Ablation with Surrogates
- Efficient benchmarking of algorithm configurators via model-based surrogates
- Beyond Forks: Finding and Ranking Star Factorings for Decoupled Search
- Warmstarting of Model-based Algorithm Configuration
- What You Always Wanted to Know about the Deterministic Part of IPC 2014 (But Were too Afraid to Ask)
- Performance Modelling of Planners from Homogeneous Problem Sets
- Predictive models and abstract argumentation: the case of high-complexity semantics
Related papers
- Classification Control of Project Variation
- An Analysis of Language Variation in Cross-talks and Sketches and its Humor Effect
- Theoretical Analysis of the Benchmark for Choosing Manipulative Instruments of Monetary Policies
- The Professional Urban Planner
- Playing Angry Birds with a Domain-Independent PDDL+ Planner