Difference between revisions of "ITUBot"
(Created page with "{{Infobox bot |id=ITUBot |author=Volkan Ilbeyli |affiliation= |country= |country2= |race=T |elo= |ladderID= |version= |type=DLL |internalDownload= |externalDownload= |language...") |
|||
Line 2: | Line 2: | ||
|id=ITUBot | |id=ITUBot | ||
|author=Volkan Ilbeyli | |author=Volkan Ilbeyli | ||
− | |affiliation= | + | |affiliation=Istanbul Technical University |
− | |country= | + | |country=Turkey |
|country2= | |country2= | ||
|race=T | |race=T | ||
Line 30: | Line 30: | ||
}} | }} | ||
− | + | ITUBot is a StarCraft Broodwar AI Bot that implements and modifies [http://arxiv.org/abs/1306.4460 Certicky's wall-in algorithm] created using logic programming. | |
+ | |||
+ | The problem is formulated as a [http://en.wikipedia.org/wiki/Constraint_satisfaction_problem constraint satisfaction problem] and solved by the Potsdam University's [http://en.wikipedia.org/wiki/Answer_set_programming ASP (Answer Set Programming)] solver: [http://www.cs.uni-potsdam.de/clasp/ ''clasp'']. The solver is tested with two optimization modes: gap minimization and resource cost minimization. A [https://www.dropbox.com/s/l5rgp5jtbnww03e/Walling%20in%20at%20StarCraft.pdf '''detailed report'''] about the analysis is generated. The report contains: | ||
+ | * Basic bot architecture | ||
+ | * A step by step explanation of Certicky's modified algorithm | ||
+ | * Algorithm run times for each optimization criteria on several different maps | ||
+ | * Differences in the layout of the buildings in the solution between different optimization criteria | ||
+ | * Related studies and future work | ||
+ | |||
+ | |||
+ | [http://www.youtube.com/watch?v=WdhIv_yxIbM Here is a video of ITUBot walling in]. |
Revision as of 12:38, 15 July 2015
Author(s):
Volkan Ilbeyli
Affiliation:
Istanbul Technical University
Race:
BWAPI version:
Type:
DLL
Download:
Language:
C++
ITUBot is a StarCraft Broodwar AI Bot that implements and modifies Certicky's wall-in algorithm created using logic programming.
The problem is formulated as a constraint satisfaction problem and solved by the Potsdam University's ASP (Answer Set Programming) solver: clasp. The solver is tested with two optimization modes: gap minimization and resource cost minimization. A detailed report about the analysis is generated. The report contains:
- Basic bot architecture
- A step by step explanation of Certicky's modified algorithm
- Algorithm run times for each optimization criteria on several different maps
- Differences in the layout of the buildings in the solution between different optimization criteria
- Related studies and future work