Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Non-Crossing Anonymous MAPF for Tethered Robots
Authors: Xiao Peng, Olivier Simonin, Christine Solnon
JAIR 2023 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We experimentally evaluate our approach on three different kinds of instances. |
| Researcher Affiliation | Academia | Xiao Peng EMAIL Olivier Simonin EMAIL Christine Solnon EMAIL CITI, INSA Lyon, Inria, F-69621 Villeurbanne, France |
| Pseudocode | Yes | Algorithm 1: Remove Cycle(Π, c) ... Algorithm 2: Build Solution(m, Π) ... Algorithm 3: new VNS(m, Π, kmax) ... Algorithm 4: lazy Approach(ub, Πl) ... Algorithm 5: dichotomous Approach(lb, ub, α) |
| Open Source Code | Yes | Our implementation is publicly available at https://gitlab.inria.fr/xipeng/tethered-amapf-jair2023.git. |
| Open Datasets | No | To study the sensibility of our algorithms to different configurations, we generate instances according to a random model that has three parameters o, n, and d which are described below. For all instances, the bounding polygon is the square B = [0, 200]2. |
| Dataset Splits | No | For each value of n, o, and d, we have randomly generated 30 instances (all instances with a same value of o share the same workspace). |
| Hardware Specification | Yes | All experiments reported in this paper are run on Grid5000 (Balouek, Carpen Amarie, Charrier, Desprez, Jeannot, Jeanvoine, L ebre, Margery, Niclausse, Nussbaum, L., Richard, O., P erez, C., Quesnel, F., Rohr, C., & Sarzyniec, L., 2013) with an AMD EPYC 7642 with 512GB of RAM. |
| Software Dependencies | No | Algorithms have been implemented in Python. In Fig. 8, we display the gap between the optimal makespan opt and the lower bound lb LBAP = maxπi ΠLBAP |πi|, the upper bound ub LSAP = makespan(s LSAP), and the upper bound ub LBAP = makespan(build Solution(s LBAP)). ... Algorithm 4 has been implemented in Java, using the Choco CP library (Prud homme, Fages, & Lorca, 2016). Algorithm 5 has been implemented in Python. |
| Experiment Setup | Yes | For all instances, the value of dt is set to 4 ... k is initialized to 2 and it is incremented each time the current solution is locally optimal; k is reset to 2 each time an improving move has been found; the search is stopped when k exceeds a given upper bound kmax or when a time limit is reached. ... The switching time of 60 seconds is chosen to find a compromise between the quality of the solution and the time required for resolution. ... For dicho, the rate α is set to 0.05 |