CUNYBot

From Liquipedia StarCraft Brood War Wiki
This article is a Bot stub. You can help Liquipedia by expanding it.
[e][h]Zerg CUNYBot
Bot Information
Programmer:
Bryan S. Weber
Affiliation:
City University of New York, College of Staten Island
BWAPI Version:
4.4.0
Language:
C++
AI Techniques:
Cobb-Douglas Economic Model
Strategies:
Various
Country:
Race:
Links
Website

Bryan Weber, CUNYBot (City University of New York) is a C++ bot that plays a full game of SC:BW as the Zerg race. Its central design feature is a focus on historical economic models. In particular, it focuses on the capital/labor ratio (k), and technology/labor ratios (t), where it responds to percieved opponent's choice of k and t. The resulting emergent behavior can be characterized as a greedy tit-for-tat style response. Link to a publication summarizing the bot in its initial state is here, and some direct gameplay links are organized with the initial release of the bot here.

Since its creation, the bot has been a base for student work, and projects like Dolphinbot are student projects to work with AI and C++. A paper describing the bot is available here.

Machine Learning Techniques[edit]

Currently, the bot uses Monte Carlo simulation to determine what units and upgrades are best to build (by including each unit/upgrade individually in an FAP sim of all observed units) and builds based on a moving average of those simulations. The bot also uses a genetic algorithm to improve between games.

Libraries[edit]

CUNYBot initially used no open libraries. It now uses the FAP combat simulator for Monte Carlo Simulation of combat and unit choices, BWEM community library for standard mapping functions, and BWEB with some minor modifications to control building locations.

Notable Placements[edit]

18th out of 27 in IEEE:COG 2019 (Full Results) Conference on Games Tournament Wiki