Optimizing Hearthstone Agents using an Evolutionary Algorithm

TL;DR

This paper introduces an evolutionary algorithm-based approach to develop Hearthstone game agents that learn and optimize decision-making, achieving competitive results without relying on external training data or complex future state modeling.

Contribution

It presents a novel coevolutionary method for creating Hearthstone agents that self-learn and outperform some state-of-the-art techniques in AI competitions.

Findings

Agent achieved top 6% in international competition

Method outperformed Monte-Carlo Tree Search in some scenarios

Self-learning coevolutionary approach proved effective

Abstract

Digital collectible card games are not only a growing part of the video game industry, but also an interesting research area for the field of computational intelligence. This game genre allows researchers to deal with hidden information, uncertainty and planning, among other aspects. This paper proposes the use of evolutionary algorithms (EAs) to develop agents who play a card game, Hearthstone, by optimizing a data-driven decision-making mechanism that takes into account all the elements currently in play. Agents feature self-learning by means of a competitive coevolutionary training approach, whereby no external sparring element defined by the user is required for the optimization process. One of the agents developed through the proposed approach was runner-up (best 6%) in an international Hearthstone Artificial Intelligence (AI) competition. Our proposal performed remarkably well, even when it faced state-of-the-art techniques that attempted to take into account future game states, such as Monte-Carlo Tree search. This outcome shows how evolutionary computation could represent a considerable advantage in developing AIs for collectible card games such as Hearthstone.