Equilibria in Bundle-Reducing Strategies for Combinatorial Auctions

We search for equilibria on a reduced (parameterized) subset of the set of strategies a bidder may use in a combinatorial auction. The strategies allow the bidders to reduce their preferred bundle in response to adverse auction results which are known after each round. Extensive experimentation using an agent-based computational simulator was conducted after which it is recognized that efficiency of the auction can be increased when bidders are allowed to reformulate their bids during the auction.