Objective of this coursework is to design and implement one or more PacMan controller using AI techniques.

The AI PacMan needs to go through decision making process to decide path it should take to complete the game by eating all the food while avoiding the ghosts.

Features:

• Additional window to show
• AI generated PacMan path
• AI generated Ghost worst path
• Configurable AI Settings

Technologies and techniques used:

• C++ programming language
• A* path finding
• Breath - first goal searching
• Mini / Max decision making

To download demo, please click here.

By default, the demo runs in AI mode where user has no controls on the pacman character. To enable user playable mode, please refer to ReadMe.txt included in the zip file.

Additional information:

Implementation of this AI PacMan controller separated into 4 sections. First, it will find the worst path each ghost towards PacMan and adds additional NEGATIVE value to those path's grid. Secondly, Breath - first seaching algorithm has been used to search goals that are available, then MiniMax algorithm used to decide the best goal. Finally, A* path finding techniques used to find its best and optimal path towards the goal.

By the time this demo has been uploaded, winning rate for this AI PacMan is around 80%.

Screenshots: