- A new multi-agent game solver software is presented that allows for the investigation of N-person games with arbitrary payoff functions defined by the user.
- The software models agent interactions in a two-dimensional cellular automaton where agents update their strategies based on their neighbors.
- Demonstration experiments show the total cooperation ratio fluctuating wildly over iterations in some cases, and remaining constant in other cases, depending on the initial conditions and payoff functions used. The results emphasize the complex and unpredictable nature of N-person games.