Advancing Strategic Decision Science Since 2014
At the Nevada Institute of Game Theory, a significant portion of research is dedicated to cooperative game theory, which studies how groups (coalitions) of players can form and allocate collective payoffs. Unlike non-cooperative theory, which focuses on individual strategies, cooperative theory assumes binding agreements can be made. Researchers at NIGT investigate fundamental solution concepts like the Core, the set of allocations where no coalition has an incentive to break away, and the Shapley value, a method for fairly distributing payoffs based on each player's marginal contribution to all possible coalitions. The Institute's mathematicians work on extending these concepts to dynamic settings where coalitions form and dissolve over time, and to games with fuzzy or probabilistic coalition memberships.
A major applied research stream involves fair division problems. How should joint costs or benefits from a collaborative project—like a shared satellite launch or a regional infrastructure development—be divided among participating entities? NIGT researchers have developed customized allocation rules based on cooperative game principles that are provably fair, efficient, and resistant to manipulation. These models are rigorously tested in the Institute's experimental lab to see how real people perceive and react to different allocation schemes. This work has been directly implemented in public utility pricing models and international joint venture agreements, ensuring collaborations are sustainable and equitable for all parties involved.
Another critical area is the study of matching markets, a branch of cooperative game theory popularized by the Nobel Prize-winning work on the stable marriage problem and kidney exchange. NIGT scholars design and analyze algorithms for matching students to schools, medical residents to hospitals, and donors to recipients in organ exchange networks. The key is creating matches that are stable—no pair of participants would prefer to be matched to each other over their current assignments. Institute research has pushed the boundaries by incorporating complex constraints, such as geographic preferences, family ties, and multi-category compatibilities, into scalable matching algorithms used by national clearinghouses.
Modern collaboration often occurs within network structures. NIGT's research in cooperative network games examines how the architecture of communication or interaction networks influences the formation of coalitions and the value they can create. A project might study how research collaborations form in a scientific community or how supply chain alliances stabilize in an industry. By modeling these as games on graphs, researchers can predict which network structures foster efficient cooperation and which lead to fragmentation. This work provides valuable insights for organizational designers and policy makers aiming to build more collaborative ecosystems in technology, science, and business.
The Institute is pioneering the integration of cooperative game theory with machine learning. One project involves using learning algorithms to estimate the characteristic function of a cooperative game—the value each coalition can generate—from historical data on past collaborations. Another explores how AI mediators can suggest fair and stable coalition structures in complex, multi-party negotiations. As the world faces grand challenges like climate change and pandemic response that require unprecedented international and inter-organizational cooperation, the Nevada Institute's research in cooperative game theory provides the mathematical backbone for designing institutions and agreements that make large-scale collaboration not just possible, but rationally compelling for all involved.