Computing Desirable Outcomes in Specific Multi-Agent Scenarios

Coalition formation and Schelling segregation are important scenarios in algorithmic game theory. While the former considers the strategic behavior of agents gathering in coalitions, the latter is a setting in which agents of two groups seek to surround themselves with like-minded agents. In each case, the quality of outcomes can be measured in form of axioms of optimality and stability. The thesis investigates how to compute such desirable outcomes efficiently and how to deal with computational intractability by means of approximation algorithms, randomization, or domain restrictions.