Offline matching approximation algorithms in exchange markets

Motivated by several marketplace applications on rapidly growing online social networks, we study the problem of efficient offline matching algorithms for online exchange markets. We consider two main models of one-shot markets and exchange markets over time. For one-shot markets, we study three main variants of the problem: one-to-one exchange market problem, exchange market problem with short cycles, and probabilistic exchange market problem. We show that all the above problems are NP-hard, and propose heuristics and approximation algorithms for these problems. Experiments show that the number of items exchanged will increase when exchanges through cycles are allowed. Exploring algorithms for markets over time is an interesting direction for future work. Categories and Subject Descriptors F.2.2 [Theory of Computation]: Algorithms & Complexity--Nonnumerical Algorithms; H.m [Information Systems]: [Miscellaneous] General Terms Algorithms, Performance Keywords Social Networks, Excha...

• Type: Conference
• Year: 2008
• Where: WWW
• Authors: Zeinab Abbassi, Laks V. S. Lakshmanan
• Keywords: Exchange Market Problem, Internet Technology, Offline Matching Algorithms, Online Exchange Markets, WWW 2008