DECENTRALIZED ALGORITHMS FOR NASH EQUILIBRIUM PROBLEMS – APPLICATIONS TO MULTI-AGENT NETWORK INTERDICTION GAMES AND BEYOND

Nash equilibrium problems (NEPs) have gained popularity in recent years in the engineering community due to their ready applicability to a wide variety of practical problems ranging from communication network design to power market analysis. There are strong links between the tools used to analyze NEPs and the classical techniques of nonlinear and combinatorial optimization. However, there remain significant challenges in both the theoretical and algorithmic analysis of NEPs. This dissertation studies certain special classes of NEPs, with the overall purpose of analyzing theoretical properties such as existence and uniqueness, while at the same time proposing decentralized algorithms that provably converge to solutions. The subclasses are motivated by relevant application examples.

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Industrial Engineering

Year

2015

Author dc:creator

• Sreekumaran, Harikrishnan

Contributors dc:contributor

• ANDREW LU LIU
• OMID NOHADANI
• MOHIT TAWARMALANI
• SATISH UKKUSURI

• Distributed optimization
• Game theory
• Network interdiction
• Parallel algorithms

Repository record dc:identifier

https://docs.lib.purdue.edu/open_access_dissertations/1432

OAI identifier oai:identifier

oai:docs.lib.purdue.edu:open_access_dissertations-2648