Reduction Techniques for Model Checking and Learning in MDPs

Authors: Suda Bharadwaj, Stephane Le Roux, Guillermo Perez, Ufuk Topcu

IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We present some results of running the reduction techniques on two model checking case studies from PRISM [Kwiatkowska et al., 2011], a well-known probabilistic model checking tool, as well as a model-learning task in a gridworld.
Researcher Affiliation Academia University of Texas at Austin Universit e libre de Bruxelles {suda.b, utopcu}@utexas.edu, {stephane.le.roux, gperezme}@ulb.ac.be
Pseudocode Yes Algorithm 1 MDP reductions based on irrelevant distributions
Open Source Code No No concrete access to source code is provided. The paper states: "Our algorithm has been implemented in a prototype tool and empirical results suggest that our algorithm achieves a high reduction rate." but no link or explicit release statement.
Open Datasets Yes The protocol is modelled in the PRISM language by [Kwiatkowska et al., 2001].1 We implement our reduction techniques, as well as the currently-used ones, and the results are given in Tab. 1. 1Model details can be found on the PRISM case studies website. We use model parameters: 2 coins and K = 2, and reset = TRUE and K {1, 2}, respectively.
Dataset Splits No No specific dataset split information (percentages, sample counts, or explicit splitting methodology) is provided. The paper describes experiments on existing PRISM models and generated gridworlds.
Hardware Specification No No specific hardware details (like GPU/CPU models, memory amounts, or specific computer specifications) are provided.
Software Dependencies No The paper mentions "PRISM [Kwiatkowska et al., 2011]" as a tool used, but does not provide a specific version number. No other specific software dependencies with versions are listed.
Experiment Setup Yes For standardization, 20% of each gridworld was populated with obstacles." and "Explicitly, we aim to learn a policy that with probability at least 1 δ will be ε-optimal in maximizing reachability probability. We measure the decrease in sampling required before ε-optimality is achieved for ε = 0.01 and δ = 0.05.