28 March 2005 Evolutionary learning through replicator dynamics in continuous space games
Author Affiliations +
Abstract
Many situations involving strategic interaction between agents involve a continuos set of choices. Therefore it is natural to model these problems using continuous space games. Consequently the population of agents playing the game will be represented with a density function defined over the continuous set of strategy choices. Simulating evolution of this population is a challenging problem. We present a method for simulating replicator dynamics in continuos space games using sequential Monte Carlo methods. The particle approach to density estimation provides a computationally efficient way of modeling the evolution of probability density functions. Finally a resampling step and smoothing methods are used to prevent particle degeneration problem associated with particle estimates. Finally we compare and contrast the resulting algorithm for simulation with genetic programming techniques.
© (2005) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Julide Yazar, "Evolutionary learning through replicator dynamics in continuous space games", Proc. SPIE 5803, Intelligent Computing: Theory and Applications III, (28 March 2005); doi: 10.1117/12.604232; https://doi.org/10.1117/12.604232
PROCEEDINGS
8 PAGES


SHARE
Back to Top