Topological Planning with Post-unique and Unary Actions
Authors: Guillaume Prévost, Stéphane Cardon, Tristan Cazenave, Christophe Guettier, Éric Jacopin
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We also experimentally show that our algorithm is capable of managing millions of NPCs per frame. We have carried out experiments on abstract settings to test Topo Plan. With the following configuration: AMD Ryzen 7 2700X (8-Core) CPU (3.7GHz), 32Gb of RAM and Windows 10 (64 bits), Topo Plan was able to provide more than 3.000.000 plans with up to 10 actions in it in less than 1.67ms, which is 10% of the time between two frames in a 60FPS video game. |
| Researcher Affiliation | Collaboration | 1Acad emie Militaire de Saint-Cyr Co etquidan, CRe C Saint-Cyr, France 2Universit e Paris Dauphine PSL, LAMSADE, CNRS, France 3Safran Electronics and Defense, France 4Hawkswell Studios, France |
| Pseudocode | Yes | Procedure 1 Build Chain(vi, s, g, D, ED, A) Input: vi M, s, g Dvi, s = g and ag vi is white; D, a set of yellow actions; ED, a set of orders between the actions of D; Parameters: x, y, two values of Dvi. Output: Each browsed action a is colored yellow, added to D and the order (Npre(a), a) is added to ED. 1: x ; y 2: if ag vi is a ghost action then fail 3: end if 4: Color(ag vi) yellow; D D {ag vi}; y pre(ag vi); ED ED {(ay vi, ag vi)} 5: if Next(ay vi) = {ay vi can be a ghost action.} then 6: Next(ay vi) ag vi 7: end if 8: while y = s do 9: if ay vi is a ghost action then fail 10: end if 11: if Color(ay vi) = yellow then fail 12: end if 13: Color(ay vi) yellow; D D {ay vi}; x y; y pre(ay vi); ED ED {(ay vi, ax vi)} 14: if Next(ay vi) = {ay vi can be a ghost action.} then 15: Next(ay vi) ax vi 16: end if 17: end while |
| Open Source Code | No | The paper mentions 'Our C++ implementation of Topo Plan' but does not provide a direct link or explicit statement about the public availability of the source code for their methodology. |
| Open Datasets | No | The paper states, 'We have carried out experiments on abstract settings to test Topo Plan. Given the description of three different realistic SAS-PUC 2 problems (Different NPCs from Red Dead Redemption 2 (including the Horse Breeder), citizens from Assassin s Creed: Origins [Ubisoft, 2017] and the acquisition machines from Horizon Zero Dawn [Games, 2017])'. While it references games, it refers to descriptions of problems from these games, not a specific, publicly available dataset in the format required for research. There's no link or citation for a dataset. |
| Dataset Splits | No | The paper does not provide specific dataset split information (percentages, counts, or references to predefined splits) needed to reproduce the data partitioning. It describes the experimental setup as 'abstract settings'. |
| Hardware Specification | Yes | With the following configuration: AMD Ryzen 7 2700X (8-Core) CPU (3.7GHz), 32Gb of RAM and Windows 10 (64 bits) |
| Software Dependencies | No | The paper mentions 'Our C++ implementation' but does not provide specific version numbers for any software dependencies or libraries used. |
| Experiment Setup | No | The paper does not contain specific experimental setup details, such as hyperparameters or system-level training settings. It mentions running experiments but lacks these specific configuration details. |