Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of Toronto

$53,000.00

Learning, Incentives and Control in Strategic Games

EF 2024-25

Résumé
Understanding and coordinating cooperation and competition between multiple intelligent decision-makers with shared resources is perhaps the most crucial challenge of our connected society. This is even more important nowadays as artificial intelligence (AI) and learning agents are permeating our lives. Game theory provides powerful tools to analyze strategic decision-making among a set of agents. It is especially attractive for decision-making in autonomous systems and networks. Such systems are ubiquitous around us - they are spanning engineered, physical, biological, and social systems. Many problems in these systems can be modelled as multi-player games. Examples range from control of autonomous vehicles, congestion control in communication networks, ad-hoc (peer-to-peer) wireless networks, to demand-side management in smart grid/power networks and autonomous robots engaged in search and rescue missions. The problems we aim to study are fundamental to interconnected multi-agent systems where information between agents/players may be incomplete or noisy, hence they must learn. Multi-agent systems may be the next big area where learning will have a big breakthrough. Learning and equilibrium-seeking algorithms can create network-wide stability by making only local decisions to benefit each agent. However, the theory in this domain has remained very little explored. An open research problem is how to design learning algorithms with minimum information requirements, since typically agents  have limited capabilities for observation and computation. Furthermore, under sufficient information sparsity, the network becomes prone to adversarial attacks that can undermine system integrity for the self-interest of a single agent. The overall goal of the proposed research is to advance the fundamentals of learning algorithms in game theory over networks, as well as providing the tools for a principled design of such algorithms. In this research program three new theoretical directions are proposed under the umbrella of distributed multi-agent systems and control: (I) advancement of learning in strategic interactions under partial information, exploiting network structure/sparsity, while being resilient to adversaries, (II) learning about the environment and adapting to it, and (III) achieving performance guarantees and convergence by design. The fourth direction (IV) is to advance the development of these theories and their applications to information and communication technologies (ICT), smart grids, cyber-physical networks, or autonomous vehicle networks. By strengthening our theoretical foundations in learning in games, this research will enhance our ability to predict and influence emergent behaviours in systems comprised of multiple learning agents. It will also empower algorithm designers to create more effective, safe, ethically responsible, and socially beneficial multi-agent learning applications.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2024-04958
Name nom
Pavel, Lacra
Department département
Electrical and Computer Engineering Department
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
53000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1510
Committee name en
Electrical and Computer Engineering
Committee name fr
Génie électrique et informatique
Area of application code
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
2500
Research subject group en
Electrical and electronic engineering
Research subject group fr
Génie électrique et électronique
Research subject en
Electrical and electronic engineering
Research subject fr
Génie électrique et électronique
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Learning, Incentives and Control in Strategic Games
Keywords
control systems; game theory; largescale decision systems; multi-agent systems; nonlinear dynamical systems; system theory; algorithms; networks; distributed optimization; learning and evolutionary systems
Application summary
Understanding and coordinating cooperation and competition between multiple intelligent decision-makers with shared resources is perhaps the most crucial challenge of our connected society. This is even more important nowadays as artificial intelligence (AI) and learning agents are permeating our lives. Game theory provides powerful tools to analyze strategic decision-making among a set of agents. It is especially attractive for decision-making in autonomous systems and networks. Such systems are ubiquitous around us - they are spanning engineered, physical, biological, and social systems. Many problems in these systems can be modelled as multi-player games. Examples range from control of autonomous vehicles, congestion control in communication networks, ad-hoc (peer-to-peer) wireless networks, to demand-side management in smart grid/power networks and autonomous robots engaged in search and rescue missions. The problems we aim to study are fundamental to interconnected multi-agent systems where information between agents/players may be incomplete or noisy, hence they must learn. Multi-agent systems may be the next big area where learning will have a big breakthrough. Learning and equilibrium-seeking algorithms can create network-wide stability by making only local decisions to benefit each agent. However, the theory in this domain has remained very little explored. An open research problem is how to design learning algorithms with minimum information requirements, since typically agents  have limited capabilities for observation and computation. Furthermore, under sufficient information sparsity, the network becomes prone to adversarial attacks that can undermine system integrity for the self-interest of a single agent. The overall goal of the proposed research is to advance the fundamentals of learning algorithms in game theory over networks, as well as providing the tools for a principled design of such algorithms. In this research program three new theoretical directions are proposed under the umbrella of distributed multi-agent systems and control: (I) advancement of learning in strategic interactions under partial information, exploiting network structure/sparsity, while being resilient to adversaries, (II) learning about the environment and adapting to it, and (III) achieving performance guarantees and convergence by design. The fourth direction (IV) is to advance the development of these theories and their applications to information and communication technologies (ICT), smart grids, cyber-physical networks, or autonomous vehicle networks. By strengthening our theoretical foundations in learning in games, this research will enhance our ability to predict and influence emergent behaviours in systems comprised of multiple learning agents. It will also empower algorithm designers to create more effective, safe, ethically responsible, and socially beneficial multi-agent learning applications.