A Market Protocol for Decentralized Task Allocation

We present a decentralized, asynchronous market protocol for allocating and scheduling tasks among agents that contend for scarce resources, constrained by a hierarchical task dependency network. Agents trade tasks and resources at prices determined by an auction mechanism, e ectively constructing supply chains through bottom-up interactions. We specify a simple set of bidding policies that, in conjunction with the auction mechanism, exhibits desirable convergence properties. The protocol always reaches quiescence. If it reaches quiescence below some end consumer's value for a high level task, the system will have found a solution consistent supply chain serving that consumer. If the system reaches a solution state such that all consumers are either winning or have given up bidding, it will achieve quiescence in such a state. Experimental evidence supports our conjecture that the system converges to a solution when one exists and a consumer values a good su ciently high. In the w...

• Type: Conference
• Year: 1998
• Where: ICMAS
• Authors: William E. Walsh, Michael P. Wellman
• Keywords: Auction Mechanism, ICMAS 1998, Intelligent Agents, Market Protocol, Supply Chain