EMERGENCE FROM LOCAL EVALUATION FUNCTION

This paper presents a new look on emergence from the aspect of locality and globality of evaluation functions for solving traditional computer problems. We first translate the Constraint Satisfaction Problem (CSP) into the multi-agent system, and then show how a global solution emerges from the system in which every agent uses a local evaluation function to decide its action, while comparing to other traditional algorithms, such as Local search and Simulated Annealing which use global evaluation functions. We also give some computer experimental results on large-scale N-queen problems and k-Coloring problems, and show that emergence only depends on problem instance, not details of agent settings, i.e. in some CSPs, the system can self-organize to a global solution, but can not in some other CSPs no matter what settings of agents have.