For my second assessment at the Academy of Interactive Entertainment, I was asked to write a description of the “proposed [AI] simulation” that I was going to undertake. In other words: what’s supposed to happen in my AI tech demo? I think what I’ll shoot for is shoot for one set of logic that will accept parameters for behavior determination.

The goal is simple: two players need to gather the largest amount of gold possible from various pools of gold that are scattered across the map. The player with the largest amount of gold by round end or gold mine exhaustion, whichever comes first. Each player will only be able to hold a certain amount of gold at any given time.

Each agent will be capable of the same actions as the other, only that each one has different conditions for switching to different states.

Decision Tasks

Movement

  • Move to Base
    • Move towards a friendly base.
  • Move to Gold Mine
    • Move towards a populated gold mine.
  • Move to Opposing Player
    • Move towards the enemy player’s avatar.
  • Move to Opposing Player’s Base
    • Move towards the enemy player’s base.
  • Arrive at T
    • Arrive at a destination.

Actions

  • Gather Gold
    • When at a location with gold, draw gold from that mine at a fixed rate after a certain amount of time has passed. Withdrawing from the location is instantaneous.
  • Steal Gold from Player
    • A random number will be generated. If the number is past a specified threshold, then an additional roll will occur for the player to steal a percentage of the currently held gold, ranging from 10 percent to 50 percent.

Locomotion / Steering Behaviors

  • Seek
    • The agent will continously move towards its destination at a fixed speed.
  • Arrive
    • When the agent is within range, the speed of the agent will decrease as it draws closer to its destination, eventually coming to a halt.

A post-mortem of the thing will come when I actually go through and write the logic for all of it!