Fail-Stop Distributed Combinatorial Auctioning Systems with fair resource allocation

A Distributed Combinatorial Auctioning System (DCAS) is a multi-agent system (MAS) where each auctioneer has a set of local bidders and all of them adopt a cooperative strategy in conducting the auction. Any occurrence of failure in the system leads to an unpredictable behavior and hence forcing abandonment of the auction. We systematically analyze the failures occurring in the agents and propose a fail-stop design of the DCAS. We present a set of Event Based Failure Handlers (EBFHs) that are triggered upon receiving specific messages from other agents in the network. We provide a proof of correctness of our approach and analyze the efficiency in terms of the message complexity. We also discuss the different parameters that influence the performance of a given DCAS configuration. We use the framework in which the concepts of basic and extended fairness are used in the bidding process.