ABL-Assignment
Description
For this assignment, you will be authoring rudimentary, intermediate, and advanced ABL behaviors for a single agent that controls multiple bots.
Starter Tasks
- Add a move action that takes a (int xdir, int ydir, int id) as input. This action should set the direction of movement for the agent and not automatically position the agent at a coordinate.
- Make a MoveTo(int id, int xpos, int ypos) behavior that moves the bot that matches the id argument to a target location over time. Note that this is not an action but a behavior that handles moving an agent over time from its current position to a destination. This behavior should succeed when the bot reseaches the destination.
- Add a SetColor action that takes (int r, int b, int g, int id).
- Add color sensing to the BotWME (this includes changes to BotSensor.java and BotWME.java).
Intermediate Tasks
- Add on action that creates a new Bot.
- Create 4 bots and have them keep a formation around the player. Your MoveTo behavior will be beneficial here.
Advanced Tasks
Design and create a multi-bot social interaction that communicates to the player via movement and bot color. The player should be able to interact with your multi-bot behavior set through movement, shooting, or some other user interaction that you design.
Here are some spaces you can design multi-bot social interactions for but feel free to create your own:
- loneliness - how would the bots make the player feel lonely?
- factions - split the bots into two factions and put the player in the middle.
- stigma - if the character or a bot had a stigma (i.e. an extremely undesirable social quality), how would the bots react? If a bot or player interacted with the stigmatized character, how should the others react?
- building trust - The player starts as not being trusted by the bots. How would they gain their trust and how would the bots performance change?
Group Work Policy
You may work in teams of up to 4 people.
Grading
This assigment will be graded on a 100 point scale.
- 30 Starter Tasks
- 30 Intermediate Tasks
- 40 Advanced Tasks
Turn-in
Push your code to your group's GitHub Classroom repository. The master branch will be graded.
Each task should be in its own ABL agent file:
- Starter.abl Demonstrates the move action, the MoveTo Behavior, the setColor action, and color sensing from the BotWME.
- Intermediate.abl Implements and executes an action that creates new Bots. Demonstrates four bots in a formation around the player.
- Advanced.abl Performs your implementation of multi-bot social interaction. Please add a description of your playable experience at the top of your the README.md file that includes a script or walkthrough on how to interact with your agent and what social interaction you have modelled.
Questions and Answers
Question: How do I build and run the project?
Answer: The process runs in two steps: 1) compile the .abl code into .java classes and 2) run the GameEngine with the newly-generated code. To help you with this, there are two classes that have a entry point for execution (i.e. they have a '''static void main(String args[])''' function declared):
- abl.build.AgentCompiler.java - This class invokes the compile in abl.jar to transcompile .abl code into .java code.
- game.GameEngine.java - This launches the game. The code for your abl agent must be transpiled by abl.jar before you can run this class.
Question: I have changed my code and recompiled the project but the change isn't seen when I run the GameEngine class. What do I do?
Answer: Periodically, Java doesn't recognize that the .java files generated by abl have change and fails to update the .class files. Run Project->Clean... on the project and try again.
Question: There is an debugger window that opens and my bots aren't moving. What's is happening?
Answer: You have the debug variable set to true in AgentCompiler.java. This tells ABL to run your agent in debug mode which opens the debugger window and automatically pauses your agent before its first decision cycle. Press the Continue button to proceed.
Screenshot of the ABL debugger
Quesion: I added multiple bots the the Chaser.abl agent but the are mostly moving vertically when they should be moving horizonatally. What's the problem and how do I fix it?
Answer: Part of the answer can be found in Ben Weber's comment above the block of move behavior declarations: moving vertically takes priority as those behaviors are of higher specificity when compared to the horizontal movement variants of the move behavior. Another confounding factor is the move behaviors in the Chaser.abl agent are made for a single Bot and do not handle multiple bots and their ids. If you wanted to fix this, you could modify the move behaviors to take an integer argument that holds the id of the bot to move. You could call these id-senstive behaviors via subgoaling one time for each agent. Alternatively, you could add a data member to the bot keeps the state of movement for the bot. With this new information, you could only move bots that have not recently be subject to a move action.