Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Launch Supplement

Université de Montréal

$12,500.00

Decision-making in dynamical systems

EF 2024-25

Résumé
Sequential decision-making problems involve taking into account not only the immediate consequences of actions but also future ones. In many applications related to pollution regulation, subsidies, and portfolio optimization, contractual or cost constraints prevent agents from taking action at all time periods, thus deviating from the classical assumption of multistage decision-making. For instance, the government occasionally adjusts subsidies, while firms make production decisions daily. When intervention decisions (impulse control) of one agent can influence other agents in the system, this leads to a dynamic game problem. The computational difficulty in solving nonzero-sum dynamic games with impulse control arises from the fact that the decision to intervene (give an impulse) depends on both immediate rewards and future interventions as well as the actions of other agents. The long-term goal of my research is to create interpretable, robust, and fair decision support systems for making interventions in dynamic environments. These intervention decisions will take into consideration the response of the agents to intervention in the population to avoid unintended consequences. For example, recent empirical research shows that electric vehicle subsidies benefit wealthier households, neglecting incentives for low-income families to purchase used cars. While average performance criteria (i.e., expectation) is predominantly used to evaluate intervention policies, we will also focus on evaluating policies based on their performance on the worst-affected populations. Over the next 5 years, my goal is to develop models and algorithms for practitioners engaged in environmental conservation (subsidy programs, wildlife protection), operations management, and finance, in particular,  1.Develop interpretable models whose prescriptions for intervention can be explained using the available information.  2.Formulate game models and devise provably convergent algorithms to analyze response of agents to interventions, particularly in leader-follower interactions where leader commits to intervention policy first and then followers determine their actions. 3. Design risk-averse intervention policies that are robust to perturbations in the problem parameters.
Programme
Discovery Launch Supplement
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
DGECR-2024-00329
Name nom
Sadana, Utsav
Department département
Computer Science and Operations Research
Organization id
63
Institution établissement
Université de Montréal
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
12500
Program id
DGECR
Program name en
Discovery Launch Supplement
Program name fr
Supplément Tremplin vers la découverte
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1509
Committee name en
Civil, Industrial and Systems Engineering
Committee name fr
Génie civil, industriel et des systèmes
Area of application code
1002
Area of application group en
Northern development
Area of application group fr
Développement du nord
Area of application en
Environment
Area of application fr
Environnement
Research subject code
1601
Research subject group en
Industrial engineering
Research subject group fr
Génie industriel
Research subject en
Operations research and management science
Research subject fr
Recherche opérationnelle et science de la gestion
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
Decision-making in dynamical systems
Keywords
nash equilibria; optimal control theory; differential game theory; dynamic programming; impulse control theory; linear-quadratic games; machine learning; robust optimization; policy optimization algorithms; network interdiction
Application summary
Sequential decision-making problems involve taking into account not only the immediate consequences of actions but also future ones. In many applications related to pollution regulation, subsidies, and portfolio optimization, contractual or cost constraints prevent agents from taking action at all time periods, thus deviating from the classical assumption of multistage decision-making. For instance, the government occasionally adjusts subsidies, while firms make production decisions daily. When intervention decisions (impulse control) of one agent can influence other agents in the system, this leads to a dynamic game problem. The computational difficulty in solving nonzero-sum dynamic games with impulse control arises from the fact that the decision to intervene (give an impulse) depends on both immediate rewards and future interventions as well as the actions of other agents. The long-term goal of my research is to create interpretable, robust, and fair decision support systems for making interventions in dynamic environments. These intervention decisions will take into consideration the response of the agents to intervention in the population to avoid unintended consequences. For example, recent empirical research shows that electric vehicle subsidies benefit wealthier households, neglecting incentives for low-income families to purchase used cars. While average performance criteria (i.e., expectation) is predominantly used to evaluate intervention policies, we will also focus on evaluating policies based on their performance on the worst-affected populations. Over the next 5 years, my goal is to develop models and algorithms for practitioners engaged in environmental conservation (subsidy programs, wildlife protection), operations management, and finance, in particular,  1.Develop interpretable models whose prescriptions for intervention can be explained using the available information.  2.Formulate game models and devise provably convergent algorithms to analyze response of agents to interventions, particularly in leader-follower interactions where leader commits to intervention policy first and then followers determine their actions. 3. Design risk-averse intervention policies that are robust to perturbations in the problem parameters.