Adversarial Hierarchical-Task Network Planning for Complex Real-Time Games

Authors: Santiago Ontanon, Michael Buro

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We also present empirical results for the µRTS game, comparing it to other state of the art search algorithms for RTS games. ... We evaluated the performance of AHTN using the free-software µRTS (https://github.com/santiontanon/microrts), which has been used by several researchers to validate new algorithms for RTS games [Onta n on, 2013; Shleyfman et al., 2014]. ... Table 1 shows the results we obtained.
Researcher Affiliation Academia Santiago Onta n on1 and Michael Buro2 1 Drexel University, Philadelphia, USA santi@cs.drexel.edu 2 University of Alberta, Edmonton, Canada mburo@ualberta.ca
Pseudocode Yes Algorithm 1 HTNPlanning(s0, N0), Algorithm 2 AHTNMax(s, N+, N , t+, t , d), Algorithm 3 AHTNCD(s, N+, N , t+, t , d), Algorithm 4 AHTNMax CD(s, N+, N , t+, t , d)
Open Source Code No The paper states 'We evaluated the performance of AHTN using the free-software µRTS (https://github.com/santiontanon/microrts)', which is the game environment, but does not explicitly state that the source code for their AHTN methodology itself is provided or linked.
Open Datasets Yes We evaluated the performance of AHTN using the free-software µRTS (https://github.com/santiontanon/microrts)... The three maps we used for our experimentation are: M1 (8x8 tiles), M2 (12x12 tiles), and M3 (16x16 tiles).
Dataset Splits No The paper describes a tournament-style evaluation on game maps but does not provide specific dataset split information (percentages, sample counts, or citations to predefined splits) for training, validation, or testing.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper mentions the use of 'µRTS' and the 'A* algorithm' but does not provide specific version numbers for any software dependencies.
Experiment Setup Yes We gave every player a budget of 200 playouts per game frame to decide on actions (but players can split the search process over multiple frames... All playouts were limited to 100 game frames using Random Biased as playout policy... The implementation of the AHTN algorithm that we used employs alpha-beta search...