Cooperative games and stable matchings in networks

Overview & Synopsis

In this thesis, we propose new solutions to matching problems in networks. We use cooperative games, particularly stable matchings, classically used in economy to analyze two-sided markets and design matching mechanisms. In the first part, we introduce bargaining and stable matching games. In the second part, we propose a new stable matching mechanism for user association in WiFi reducing the impact of the anomaly in the protocol. Furthermore, we analyze a video caching problem and show a new algorithm enumerating stable structures. In the third part, we analyze conditions for the stability of some fairness schemes in terms of risk aversion indicators. In the fourth part, we analyze the stability of a two-sided crowdsourcing marketplace with scheduling constraints on the tasks. The classical substitutability condition does not hold in this case. We introduce new conditions and show the existence of stable matchings. We also solve the crowdsourcing problem as a non-cooperative game in extensive form.