Swarming the Kingdom: A New Multiagent Systems Approach toN-Queens

摘要

This paper introduces Swarm Queens, a Swarm Intelligence approach to the N-Queens problem. ERA is the current best Multiagent Systems solution to this classic Constraint Satisfaction Problem. Our algorithm improves upon ERA in terms of both time and space complexity. In addition, Swarm Queens has a simpler design and delivers much more consistent, time performance from run to run. We discuss the similarities and differences between these two algorithms, showing how the design decisions we made led to better performance. Our experiment demonstrates that consistent global results can be achieved 'with non-deterministic local behavior in Multiagent Systems. Furthermore, Swarm Queens, exhibits features that may be useful in designing a better general MAS algorithm for .solving Constraint Satisfaction Problems.