SMU Guildhall Graduate Catalog Spring 2018 — Cohort 26 - Page 92

Rebecca S. Ward Software Development « Artificial Intelligence: Choosing Between Group and Individual Goals This project is about showing how we can This project needs to have a system that will put the needs of the group above the AI and a World AI that can withstand make Artificial Intelligence (AI) agents that needs of their own. The AI used in this project is Utility Theory. This project took approximately nine months. This is a good project for learning how to use and implement Goal-Oriented Action Planning (GOAP). For this project to have interesting choices, we will need a Utility System that will allow for it. Each AI will have at least five needs that they are trying to fulfill. The World AI will allow for at least 15 goals for the AI to complete. One to two goals will fulfill each type of need. Having multiple different worlds will allow us to see how they will react in different environments. 92 SOFTWARE DEVELOPMENT can accommodate for Individual Agent different environments. These agents need a world that is big enough to house the maximum population for said world. The agents calculate their needs; based off these needs and the world’s, they will need to choose appropriate goals to pursue. These goals need to have a way of being completed—this is where GOAP comes into play. The World needs a map for the Agents to move around and gather the resources they need to use for survival as well as to help the community. The map itself will be made of several types of tiles that will help them fulfill their individual and world needs.