Auctions provide an efficient way of resolving one-to-many negotiations. Th
is is particularly true for automated agents where delays and long communic
ations carry negative externalities. A properly designed auction, tailored
to the specific needs of the relevant multi-agent system, can significantly
improve its performance. In this paper, we focus on the specific problem o
f service allocation among autonomous, automated agents, within the context
of the ADEPT project, which concerns the BT (British Telecom) business pro
cess of providing a quote for designing a network for a customer.
The main contributions of this paper are threefold: First, we show how an E
nglish auction can be modifed for services, which an multi-dimensional priv
ate value objects. Second, we show how, under certain conditions, auctions
can be arranged by the service providing agents, in the cases where the ser
vice seeking agents fail to do so. We consider the incentives of all partic
ipants, and show how such an arrangement can be in their best interest. Fin
ally, by examining our results for what is, essentially, an application of
game-theory and mechanism design to an existing application, we draw some g
eneral conclusions on how such concepts can be operationalized in automated
agents. (C) 2000 Elsevier Science B.V. All rights reserved.