Artificial Intelligence is a controversial but growing field of study. My project (problem) was to design an agent that would survive in the hostile Wumpus world. This virtual world has been used to validate many types of agents including a doctorial dissertation on A.I..
The Wumpus world is a very challenging environment in which to live--there are many obstacles to navigate, a few percepts that give the slightest hint of those obstacles, and, finally, the goal which is to find gold. My agent was designed in such a way that it will maximize its score. That is, go into the world, grab as many pieces of gold as it deems safe, and exit the world.
I have achieved the overall goal of my project. I did this through various A.I. techniques. My agent is primarily a Reflex Agent, that is, it follows a set of rules that incorporates a notion of greediness and potentials. Greediness is not only how much gold should the agent get, but also, whether it is reasonable to continue searching. Potentials are numeric values of goodness or badness associated with an individual room. The combination of these attributes and various techniques provides my agent with superior success in the Wumpus environment.