Efficient Object Search in Game Maps

Authors: Jinchun Du, Bojie Shen, Shizhe Zhao, Muhammad Aamir Cheema, Adel Nadjaran Toosi

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our extensive experimental study, conducted on standard game maps benchmarks and real-world keywords, demonstrates that our approach has up to 2 orders of magnitude faster update times for moving objects compared to stateof-the-art approaches such as navigation mesh and IR-tree.
Researcher Affiliation Academia Jinchun Du , Bojie Shen , Shizhe Zhao , Muhammad Aamir Cheema , Adel Nadjaran Toosi Faculty of Information Technology, Monash University, Melbourne, Australia {jinchun.du, bojie.shen, shizhe.zhao, aamir.cheema, adel.n.toosi}@monash.edu
Pseudocode Yes Algorithm 1: Boolean k NN query processing
Open Source Code Yes 1https://github.com/goldi1027/GT-EHL
Open Datasets Yes We run experiments on widely used game map benchmarks [Sturtevant, 2012] of four popular games: Dragon Age II (DA); Dragon Age Origins (DAO); Baldur s Gate II (BG) and Star Craft (SC).
Dataset Splits No The paper describes generating objects, keywords, and queries, but does not specify formal training, validation, or test dataset splits or percentages.
Hardware Specification Yes We run our experiments on a 3.2 GHz Intel Core i7 machine with 32 GB of RAM.
Software Dependencies No All the algorithms are implemented in C++ and compiled with -O3 flag. The paper mentions using 'nltk, an NLP library' but does not specify its version. It also mentions 'Chat GPT (Jan 9 version)' but this is a tool used for keyword generation, not a software dependency for the algorithm itself.
Experiment Setup Yes We vary the density from 0.1% to 10% and the default density is 1%. We define mobility of an object set as the percentage of objects that move between two timestamps. We vary the mobility from 10% to 100% and the default mobility is 70%. We evaluate the effect of k which is varied from 1 to 10 where the default value of k is 3. We also evaluate the effect of number of query keywords by varying the number of query keywords from 0 to 3 where the default number of keywords is 2. For each experiment, we generate 100 queries per timestamp.