Research On The Innovative Mechanism Of Big Data Transaction Based On Combinatorial Auctio

The explosive growth of data has led to only increasing research on its pricing methods,with a variety of research methods blossoming,but less research addressing the problem of high costs to buyers due to the large-volume data.This means that many data demanders with deep processing capabilities may give up bidding because they cannot afford the ultra-high costs,which in turn leads to a potential waste of data resources.In this thesis,we start with the concept of data packets in Computer Communication Networks,and divide the large-volume data into multiple packets for sale according to data types and data sources.And based on the combination auction,we propose the optimal combination auction transaction mechanism for a single seller and the multi-unit combination auction transaction mechanism for multiple sellers.It achieves cost reduction from two perspectives of maximizing the interests of sellers and maximizing social welfare,respectively,breaks through the information asymmetry barrier of transaction participants,and provides conditions for releasing the potential value of data.The first model aims at maximizing sellers’ revenue,incentivizing data vendors to provide data on the one hand,and achieving cost reduction while ensuring buyers’ differentiated competitiveness on the other.When designing the payment rules,we use the difference between the highest contribution value and the next highest contribution value as the buyer’s revenue sum to further maximize the seller’s revenue and ensure the optimal mechanism.We demonstrate mathematically that the mechanism satisfies the individual rationality and allocative efficiency constraints,as well as the incentive compatibility constraint that the target buyer telling the truth is the dominant strategy regardless of single buyer speculation and multiple buyer speculation.Finally,we simulate real market transactions through simulation experiments to verify the stability of the incentive-compatible nature of the mechanism and to demonstrate that the mechanism brings higher returns to sellers.The second model,with the goal of maximizing social welfare,incorporates the zero-cost replicability of big data in the first model,Providing opportunities for buyers such as companies with less financial resources and less comprehensive strength or new organizations in the start-up phase.When there are multiple successful bidders for a batch of data products,the cost can be shared equally and the financial burden can be reduced.At the same time,if the organization has a breakthrough in data analysis technology,there is hope to achieve the maximum benefit with the minimum cost.Given that the combination of data non-exclusivity and portfolio auctions greatly increases the difficulty of data allocation operations,this model simplifies the problem by adopting the Vickrey-Clarke-Groves mechanism as a payment rule.This can incentivize buyers to report the true value to maximize their own interests and thus create maximum social welfare.Finally,the model’s strategic prevention and feasibility are also analyzed through mathematical arguments and arithmetic examples.