Market-based Task Allocation for Dynamic Information Processing Environments
Conference Proceeding Article
Flexible and large-scale information processing across enterprises entails dynamic and decentralized control of workflow through adaptive allocation of knowledge and processing resources. Markets comprise a well-understood class of mechanisms for decentralized resource allocation, where agents interacting through a price system direct resources toward their most valued uses as indicated by these prices. The information-processing domain presents several challenges for market-based approaches, including (1) representing knowledge-intensive tasks and capabilities, (2) propagating price signals across multiple levels of information processing, (3) handling dynamic task arrival and changing priorities, and (4) accommodating the increasing-returns and public-good characteristics of information products. A market gaming environment provides a methodology for testing alternative market structures and agent strategies, and evaluating proposed solutions in a realistic decentralized manner.
decentralized resource allocation, dynamic processing environments, information products, knowledge-intensive tasks, large-scale information processing, market gaming environment, market-based task allocation, price signals propagation
Artificial Intelligence and Robotics | Operations Research, Systems Engineering and Industrial Engineering
Intelligent Systems and Decision Analytics
International Conference on Integration of Knowledge Intensive Multi-Agent Systems : KIMAS '03 : September 30-October 4, 2003, Cambridge, MA
City or Country
WELLMAN, Michael P. and CHENG, Shih-Fen.
Market-based Task Allocation for Dynamic Information Processing Environments. (2003). International Conference on Integration of Knowledge Intensive Multi-Agent Systems : KIMAS '03 : September 30-October 4, 2003, Cambridge, MA. 109-114. Research Collection School Of Information Systems.
Available at: http://ink.library.smu.edu.sg/sis_research/1216