Asymmetric Action Abstractions for Multi-Unit Control in Adversarial Real-Time Games
Authors: Rubens Moraes, Levi Lelis
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirical results on combat scenarios that arise in a real-time strategy game show that our search algorithms are able to substantially outperform state-of-the-art approaches. |
| Researcher Affiliation | Academia | Rubens O. Moraes, Levi H. S. Lelis Departamento de Inform atica, Universidade Federal de Vic osa, Brazil {rubens.moraes, levi.lelis}@ufv.br |
| Pseudocode | Yes | Algorithm 1 Portfolio Greedy Search |
| Open Source Code | No | The paper provides a link to Spar Craft (github.com/davechurchill/ualbertabot/tree/master/Spar Craft), which is a testbed/simulation environment used, not the open-source code for the authors' specific methodology (GAB/SAB) described in the paper. |
| Open Datasets | No | The paper uses a simulation environment (Spar Craft) for experiments, detailing combat configurations and unit types. It does not use or provide access to a pre-existing, publicly available dataset in the traditional sense, but rather defines the simulated environment's parameters. |
| Dataset Splits | No | The paper describes simulation setups and runs 'matches' to evaluate performance, but does not specify training, validation, or test dataset splits as it's a simulation-based study rather than one using a pre-collected dataset. |
| Hardware Specification | Yes | All experiments are run on 2.66 GHz CPUs. |
| Software Dependencies | No | The paper mentions using 'Spar Craft' as a testbed but does not specify any version numbers for Spar Craft or any other software dependencies/libraries used for the implementation. |
| Experiment Setup | Yes | We use P = {NOKAV, Kiter} and a time limit of 40 milliseconds for planning in all experiments. We use the Ψ function described by Churchill et al. (2012). Instead of evaluating state s directly with LTD2, our Ψ simulates the game forward from s for 100 state transition steps until reaching a state s ; we then use the LTD2-value of s as the Ψ-value of s. The game is simulated from s according to the NOKAV script for both players. |