Risk-Aware Stochastic Shortest Path
Authors: Tobias Meggendorfer9858-9867
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Evaluation of our prototype implementation shows that risk-aware control is feasible on several moderately sized models.Our results for the first experiment are summarized in Table 1.The results of the second experiment, evaluating the influence of the threshold, are depicted in Fig. 6. |
| Researcher Affiliation | Academia | Tobias Meggendorfer Institute of Science and Technology Austria, Am Campus 1, 3400 Klosterneuburg, Austria tobias.meggendorfer@ist.ac.at |
| Pseudocode | Yes | Algorithm 1: Value Iteration to compute CVa R |
| Open Source Code | Yes | Our implementation, all models, and instructions to reproduce the experiments can be found at https://doi.org/10.5281/zenodo.5764140. |
| Open Datasets | Yes | Furthermore, we consider two models from the literature, namely Fire Wire (Kwiatkowska, Norman, and Sproston 2003), the IEEE 1394 Fire Wire root contention protocol, and WLAN (Kwiatkowska, Norman, and Sproston 2002), the CSMA/CA mechanism of the 802.11 Wireless LAN protocol.Our implementation, all models, and instructions to reproduce the experiments can be found at https://doi.org/10.5281/zenodo.5764140. |
| Dataset Splits | No | The paper evaluates its algorithms on various "models" (e.g., Grid, Fire Wire, WLAN, Walk) which represent the systems under analysis, rather than conventional machine learning datasets that are typically partitioned into training, validation, and test sets. There is no mention of dataset splits like percentages or sample counts for these models. |
| Hardware Specification | Yes | running on consumer-grade hardware (AMD Ryzen 5 3600, 3.60 Ghz, 16 GB RAM). The JVM is limited to 10 GB of RAM through -Xmx10G. |
| Software Dependencies | Yes | We implemented prototypes of our algorithms in Java (Oracle JVM 17.0.1), delegating LP calls to Gurobi 9.1.2 |
| Experiment Setup | No | The paper describes the characteristics of the models used for evaluation (e.g., Grid dimensions, janitor behavior, Walk model rules) but does not provide specific experimental setup details such as hyperparameters (learning rates, batch sizes, number of epochs) or other system-level training configurations for its algorithms. |