Controlling Computational Cost: Structure, Phase Transition and Randomization
Abstract
This report describes Cornell's contribution to the ability to build information systems that use highly decentralized and autonomous negotiation of tasks for distributed resource allocation. This effort extends phase transition analysis to structured domains and generalized constraint satisfaction tasks. The effort was devoted to connecting frameworks for multi-agent negotiation based systems with the research on analytical and empirical computational complexity. The general goal was to improve the expressiveness and scalability of complex distributed systems by exploiting computational hardness awareness in both the design and operation of the systems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 2004
- Accession Number
- ADA426243
Entities
People
- Bart Selman
Organizations
- Cornell University