Teaching a computer game simulation to think for itself.
This year Samantha Lok has spent her 4th year, honours project working on the ‘Tartarus Problem’. Samantha, an international student from Malaysia, is a Bachelor of Engineering student specialising in Software Engineering.
She has been developing software to allow a computer simulation to learn from its mistakes. The simulation learns through a process of trial and error, like any animal or person does.
In Samantha’s words: “The AI agent will first explore the simulations world then proceed to exploit the best action to be taken.”
For example if the simulation approaches a wall and crashes into it, the next time it will know better and turn before it hits the wall.
It is this kind of artificial intelligence that is used for solving multistep problems as needed in computer games. The simulation would learn to optimize the space in the game and could be applied to a wide range of games, from simple phone based games to open world adventures. Using the simulation would reduce the time needed for game developers to design games.
Samantha has had the chance to try her hand at game design throughout her degree. One project that sticks out in her mind was a team challenge to design an online game for kids to learn why it’s important to insulate your house.
“The game was for the Sustainable Trust, once we completed the trial game they were loaded up on their website and kids could test them out. The opportunity to work directly with an industry client was very rewarding.”
Samantha and her group were one of the top teams in the challenge with their game which tasked kids with directing a cartoon cat around a house to try and fix it up and make it more energy efficient. Samantha said the opportunity to work with an industry client was very rewarding.