A Pre-processing Approach for Fast and Stable Allocations on Approximation-based Pricing for Multi-unit Combinatorial Auctions

Abstract

In this paper, some discussions about a pre-processing approach of fast approximation on stable pricing and allocation of resources in a combinatorial auction are presented. On the discussions, an approximate auction which has VCG-like pricing mechanism is used which considers the situation when a cancellation of winner bid(s) could be occurred after its completion of winner determination. An analysis about stable approximate pricing mechanisms against cancellation of a winner after its winner determination is also presented, where a single-unit non-combinatorial reserve price bidding on a combinatorial auction is employed on it. The pricing algorithms employ a kind of approximate allocation and pricing algorithms that are capable of handling multi-unit auctions with reserve price biddings. We consider a scenario on the allocation of electricity power usage rights while considering electricity generation costs on the power suppliers as well as external conditions such as violations of regulations done by some bidders outside the auction mechanism. An experimental analysis is presented on the scenario and the presented algorithms efficiently produced approximation allocations that are necessary in the pricing phase, while keeping the same level of stability in the case of single-winner cancellation scenario.

An improved pricing approach based on combinatorial auction method in cloud environment

“…The challenge of this method is that it requires precise scheduling of tasks. In (Borjigin et al, 2018;Fukuta, 2016), the researchers present an auction model with a combined auction method with copies of resources. Buyers bid their offer without knowing the price of their competitors.…”