Research On Bitcoin Blockchain Transaction Pattern Based On Game Theory

Blockchain is a novel application model that combines the uniqueness and innovation of computer technology.Researches in related fields such as economy,politics,society and law show an explosive growth trend,among which the most prominent one is the design and implementation of digital currency in the economic field.This dissertation takes transaction mode of Bitcoin as the research object,and transaction pricing problem of Bitcoin as the main line,exploring the modeling of transaction mode and revealing principle of transaction mechanism from two aspects of transaction packaging sorting mechanism and transaction pricing model.Specific research results and innovations can be summarized as follows:(1)Based on the single-round static game model,this section establishes a payment model and revenue model respectively from the perspective of dynamic game to analyze bitcoin transaction pricing formulation based on GSP mechanism.In order to optimize the continuance and enhance the ability of self-adjustment of the proposed model,transaction ranking problem is considered from the perspective of successive rounds.This dissertation puts forward a method to analyze transaction mode from the perspective of dynamic game based on the existing static game model under GSP mechanism,and establishes a payment and revenue model respectively to analyze user’s income and expenditure in during dynamic game,so as to obtain a more flexible and complete transaction system;Demonstrating the proposed model from a mathematical point of view and verifying the real usability of the proposal by deducing reasoning;A new transaction ranking result is obtained according to the payment standard from the proposed dynamic game model,also transaction ranking result is compared with that under static model according to the revenue formulation.(2)This dissertation proposes a Bitcoin transaction fees prediction model with multiplayer dynamic game model combining GSP mechanism,from which the impact of historical transactions on new transactions is visualized from the perspective of users.Users in this model will comprehensively take five factors of transaction size,transaction location,historical transaction cost,transaction accumulation and transaction delay into account when trading.The fee of each transaction uploaded not only refers to the historical records in the block but the impact on subsequent transactions,then the proposed model in this paper takes maximization of user income as the key point to predict transaction fee,so that guaranteeing the clarity of each step and allowing tweak on the prediction model in time with the change of situation,fully improving the adaptability of model to surroundings;This paper also proposes an expression on user satisfaction and establishes user benefit model along the way,finding an optimal bidding strategy from the perspective of profit maximization,in which users can get a reasonable and acceptable bidding price and also ensuring successful packaging to an extreme.The users cost can be reduced as far as possible to improve users benefit as well,so that participants can have a good trading experience in Bitcoin system;The article also compares this proposal with deep learning method from the perspective of bidding and machine learning algorithm evaluation indicators,and fully reflects the efficacy and advantages of this model in Bitcoin transaction price prediction.