An Economic Framework for Resource Allocation in Ad-hoc Grids An Economic Framework for Resource Allocation in Ad-hoc Grids

I n this dissertation, we present an economic framework to study and develop different market-based mechanisms for resource allocation in an ad-hoc Grid. Such an economic framework helps to understand the impact of certain choices and explores what are the suitable mechanisms from Grid user/owner perspectives under given circumstances. We focus on resource allocation in a Grid-based environment in the case where some resources are lying idle and could be linked with overloaded nodes in a network. In such networks, the resources are neither necessarily dedicated nor have predictable availability at any point in time. We call such networks ad-hoc Grids. Selfinterested nodes in ad-hoc Grids are considered as consumers (buyers) and producers (sellers) of resources within the economic framework. Consumers and producers of resources are autonomous agents that cooperate through a simple, single metric namely the price that summarizes the global state of a network in a number. The price represents all the available information that may reside at the level of the individual nodes and that is not necessarily shared among them. A middle agent, named the matchmaker, sets up a mutual agreement between consumer and producer agents based on the price by employing economic mechanisms such as auctions. The transaction is established when the consumer and producer constraints such as resource quantity, time, and budget are met. In this dissertation, we study market-based resource allocation mechanisms at macro and micro levels. Macroeconomics addresses the behavior of an economy at the aggregate level and microeconomics describes the individual behavior. At the macro level, we compare different economic models as the matchmaking mechanisms. We study the impact of choosing particular auction mechanisms in the framework. At the micro-level, we study different pricing mechanisms and investigate the effect of introducing money and budget constraints. Furthermore, we analyze different bidding strategies that help agents to better achieve their objectives under varying constraints.