The authors would like to thank the National Science Foundation, which provided support for part of this work under grant BCS-9211914. We would also like to thank the rest of the UNH Cooperative Distributed Problem Solving research group for their helpful comments and suggestions.