Arrival Process Modelling in Internet Auction

Abstract

Exploring the characteristics of online auction, one of the most important applications on the Internet, is a critical issue for computing and communication capacity assessment to the success of system design in E-commerce. In this study, the self-similarity and long-range dependence of the bid arrival flows associated with various commodity items commonly traded over an Internet auction are evaluated using the Hurst parameter. Findings show that video games, books or clothes exhibit a weak self-similarity. By contrast, brand name handbags and smartphones have a strong degree of self-similarity. Furthermore, a trace generating model is proposed to mimic the aggregate bid arrival process of commodities with short-range or longrange dependence properties, respectively. The validity of the proposed model is demonstrated by comparing the synthesized bid arrival traces with sample auction traces gathered from eBay. The proposed modelling methodology shows efficacy of approaching Internet auction data system for online e-commerce system design management.