CS-175
Daanya Anand, Jatin Sethi, Ryan Feigenbaum CS 175, Project Proposal
Summary of our Project:
For our project this quarter, we will be choosing option 1, Malmo Minecraft. Our problem for the AI to solve will be to locate, find a path to, and obtain the diamond on the map. It must then make its way back to the starting location. In its way will be obstacles such as fences, enemy mobs, and doors. This project can be changed and adapted to different goal states, in that we could set the goal to be a variety of different things, such as going through a nether portal, or finding gold instead of diamonds.
Evaluation Plan:
The algorithm will be measured in the following way: the AI will start off with 100% health and in its first life. Each movement and/or attack will have a percentage of “energy loss” associated with it, and the AI’s goal will be to maximize its energy level by the end of its course to the diamond. As mentioned below, our stretch goal is to implement multiple lives, so once we get to that stage, an added metric will be the amount of lives lost as well. Another metric we can consider using is time, to measure how long the AI takes to complete its mission and return to its starting point.
Our moonshot case is that the AI will reach the diamond in the most efficient way, i.e. It will adapt to the current environment and if need be, avoid some monsters to reach the diamond in the least time. Efficiency will be measured by the amount of health left at the end of its journey.
Goals and Milestones:
Goal 1) Our minimum goal is to get our AI to be able to find a randomly-placed diamond without any obstruction in its path. The diamond can be placed anywhere on the map, and our AI will have to find the most efficient path to the diamond. Milestone 1) Create Minecraft kingdom where diamond is randomly placed on the map. The AI must be able to perceive the diamond and get to it. Milestone 2) The AI must now be able to identify the most efficient path to the diamond and take that route to the diamond. The AI must also be able to get back to its initial location.
Goal 2) Our realistic goal is to create an AI that can fight off monsters, get past fences, open doors, and more to reach the diamond and retreat back to its starting point. Milestone 3) The AI must be able to identify the difference between a fence and door and appropriately handle the obstruction. Milestone 4) The AI must now be able to handle moving monsters and attack them before the monster reaches the AI.
Goal 3) Our stretch goal is to create an AI that can detect multiple diamonds and retrieve all of them, without losing all 3 lives. The obstructions the AI will need to pass will have varying effects on the AI, and the algorithm will have to choose the path that causes the least damage to its score. Milestone 5) Now we must implement the AI having multiple lives. The AI needs to identify the path with least obstruction and, on occasion, choose to lose one life if that maximizes its score at the end. This will require the AI to look ahead of its current stage and anticipate