Spectrum-Based Fault Localisation for Multi-Agent Systems
Authors: Lúcio S. Passos, Rui Abreu, Rosaldo J. F. Rossetti
IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experimental evaluation shows that 10 out of 42 heuristics yield the best diagnostic accuracy (96.26% on average).We present an experimental study on the impact of 42 heuristics in the ESFL-MAS diagnostic accuracy using the well-known and real-world representative Pickup and Delivery Problem as test suite; |
| Researcher Affiliation | Academia | L ucio S. Passos , and Rui Abreu , , and Rosaldo J. F. Rossetti , DEI/FEUP Artificial Intelligence and HASLab Palo Alto University of Porto Computer Science Lab. INESC TEC Research Center Porto, Portugal University of Porto Campus de Gualtar 3333 Coyote Hill Rd Porto, Portugal Braga, Portugal Palo Alto, CA, U.S.A. lucio.san.passos@gmail.com, rui@computer.org, and rossetti@fe.up.pt |
| Pseudocode | Yes | Algorithm 1: ESFL-MAS Algorithm |
| Open Source Code | No | No explicit statement or link was provided indicating that the source code for the described methodology is publicly available. |
| Open Datasets | Yes | We use an instance of the Pickup and Delivery Problem (PDP) [Savelsbergh and Sol, 1995] to test our approach because (i) it is well-known and (ii) it is a real-world representative problem. ... The Second Edition of the Multi-Agent Programming Contest (MAPC)1 [Dastani et al., 2007] provides an instance of PDP known as the Gold Miners scenario. |
| Dataset Splits | No | The paper describes generating test cases and collecting data over time steps, but does not provide explicit training, validation, or test dataset splits in terms of percentages or counts for reproducing the data partitioning for a model. |
| Hardware Specification | No | No specific hardware details (like GPU/CPU models, memory, or cloud instance types) used for running the experiments are mentioned. |
| Software Dependencies | No | The paper mentions 'Agent Speak' and 'Jason [Bordini et al., 2007]' but does not provide specific version numbers for these software dependencies. |
| Experiment Setup | Yes | For the experimental setup, we randomly generated 5 test cases and each of them corresponded to a set of initial positions for: agents, the depot, and gold nuggets. To collect information to generate spectra, the MAS with 25 agents was executed 75 times for each test case recording 1000 time steps. and Seeking completeness of our test suite and knowing MAS organizations [Horling and Lesser, 2004], we implement modified MASs varying both the coordination and spatial organisation dimensions resulting in the following types of MAS: (1) Non-Coordinated and Non-Organised (NCNO), where agents work individually (not cooperatively) and do not receive a search quadrant (loose spatial organization); (2) Non Coordinated and Organised (NCO), where agents work individually but each of them has an assigned search quadrant; (3) Coordinated and Non-Organised (CNO), where agents coordinate the gold-nuggets search, yet there is no allocated quadrant; and (4) Coordinated and Organised (CO), where agents coordinate the gold-nuggets search as well as have an assigned search quadrant. |