M.R. Lauer, Dept. of Math. & Comput. Sci., Georgia State Univ., Atlanta, GA, USA
P.A. Mitchem, Dept. of Math. & Comput. Sci., Georgia State Univ., Atlanta, GA, USA
R.A. Gagliano, Dept. of Math. & Comput. Sci., Georgia State Univ., Atlanta, GA, USA
Conway's 'game of life' is extended using Axelrod's (1984) model of the prisoner's dilemma to investigate resource utilization. Totally cooperative "birth" and "death" rules in the standard game ("selfless" nodes) are combined with competitive algorithms of strategy from the prisoner's dilemma ("selfish" nodes). In prior simulation experiments, non-zero levels of "selfish" births provided counterintuitively higher resource utilization values; i.e. a larger population on a fixed sized grid. Additional results are obtained by examining the separate processes of "selfish" nodes; e.g. the relative advantage of "killing" other nodes over a derived vitality factor. These game extensions are called 'real life', which includes both processes, and 'special life', which has the vitality factor only. In order to analyze resource utilization, stability and complexity, several metrics are employed, including Hamming distance and fractal (box) dimension. Combining two strategies appears to not only enhance overall system effectiveness but also provides insight into the relationships between notions of order, complexity and chaos.
Index Terms:
competitive algorithms; cellular automata; optimisation; chaos; simulation; resource optimization; self-interest; game of life; prisoner's dilemma; resource utilization; cooperative rules; node birth; node death; selfless nodes; competitive algorithms; selfish nodes; simulation; fixed sized grid; node killing; vitality factor; real life; special life; stability; complexity; chaos; Hamming distance; fractal dimension
Citation:
M.R. Lauer, P.A. Mitchem, R.A. Gagliano, "Resource optimization and self interest: variations on the game of life," ss, pp.136, 28th Annual Simulation Symposium, 1995