This project, a structured attempt to solve spider solitaire and similar games, originated in Stuart Russell's introduction to Artificial Intelligence class. This page contains a project description, demonstrations, and relevant links including one to the solitaire agents wiki.
Spider solitaire is a particularly challenging card-based solitaire game requiring both luck and clever play. From an artificial intelligence perspective, spider solitaire is a partially observable markov decision process (POMDP) that appears to be an appropriate domain for reinforcement learning techniques.
As a continuation to Stuart Russell's course on artificial intelligence, we plan to apply a variety of approaches to building a spider solitaire agent. This domain should allow for a systematic investigation of reinforcement learning techniques.
Currently, our project has focused solely on spider solitaire, a particularly challenging game that requires both luck and skill.
To be released