Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

The University of British Columbia

$59,000.00

Structured Blackbox Optimization

EF 2023-24

Résumé
Optimization is the study of minimizing or maximizing a function. From minimizing construction costs to maximizing image quality, optimization arises naturally in virtually every field of modern research. Thus, there is a need for fundamental research in optimization. Moreover, HQP trained in the proper use of optimization are highly-valued members of any research team. A blackbox is a function that provides output without explanation of how the output was constructed. A structured blackbox optimization (SBBO) problem is an optimization problem where some, or all, of the functions defining the problem are given by blackboxes, but the problem retains some visible structure. For example, a computer simulation may return two values and the objective is to minimize their product. In this proposal, we continue our long-term research program in SBBO focusing on Analysis, Algorithm Design, and Applications. Our long-term objectives are to improve tools and algorithms for SBBO, and ensure that those assets are available to the broadest research community. Our Analysis research focuses on Numerical and Variational Analysis. In the past 5 years, we have: advanced active set complexity; analyzed approximate gradients; and expanded knowledge of subspace decompositions. In this proposal, we will further our research by developing analysis of p-bases and advancing techniques to approximate Hessians. These advances will be instrumental in Algorithm Design for SBBO. Our research in Algorithm Design focuses on blackboxes that only provide function values. Our recent research has: developed algorithms using subspace decompositions; analyzed algorithms using complex variables; and presented practical expositions on the proper comparison of optimization algorithms. In this proposal, we will expand our research by developing methods for high-dimension SBBO and developing an SBBO method that dynamically selects between direct-search and model-based strategies. These algorithms will be inspired from and applied to real-world optimization Applications. In Applications, we maintain an open-minded philosophy and apply our expertise in wide range of problems. In the past five years, we have modelled and optimized problems in: road design; seismic retrofitting; image denoising; and solid-state tank design. In this proposal, we will continue to connect our fundamental research with real-life problems. Our work will improve solution time and quality, thereby have immediate positive impacts on the engaged research areas. This interaction also iteratively inspires and directs our research in Analysis and Algorithm Design. In all activities, we will train HQP in valuable research skills. They will learn mathematical tools of optimization, master numerical and software development techniques, and gain skills in mathematical modelling and applied optimization. Many HQP will be trained through multidisciplinary co-supervision that will lead to skill acquisitions beyond SBBO.
Programme
Discovery Grants Program - Individual
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2023-03555
Name nom
Hare, Warren
Department département
Okanagan Campus, Mathematics
Organization id
2
Institution établissement
The University of British Columbia
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
59000
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
1509
Committee name en
Civil, Industrial and Systems Engineering
Committee name fr
Génie civil, industriel et des systèmes
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
1601
Research subject group en
Industrial engineering
Research subject group fr
Operations research and management science
Research subject en
Génie industriel
Research subject fr
Recherche opérationnelle et science de la gestion
Application title
Structured Blackbox Optimization
Keyword
Applied optimization; Blackbox optimization; Derivative-free optimization; Hessian approximations; Mathematical Modelling; Numerical analysis; Numerical approximation errors; Numerical Optimization; Positive Basis; Variational analysis
Application summary
Optimization is the study of minimizing or maximizing a function. From minimizing construction costs to maximizing image quality, optimization arises naturally in virtually every field of modern research. Thus, there is a need for fundamental research in optimization. Moreover, HQP trained in the proper use of optimization are highly-valued members of any research team. A blackbox is a function that provides output without explanation of how the output was constructed. A structured blackbox optimization (SBBO) problem is an optimization problem where some, or all, of the functions defining the problem are given by blackboxes, but the problem retains some visible structure. For example, a computer simulation may return two values and the objective is to minimize their product. In this proposal, we continue our long-term research program in SBBO focusing on Analysis, Algorithm Design, and Applications. Our long-term objectives are to improve tools and algorithms for SBBO, and ensure that those assets are available to the broadest research community. Our Analysis research focuses on Numerical and Variational Analysis. In the past 5 years, we have: advanced active set complexity; analyzed approximate gradients; and expanded knowledge of subspace decompositions. In this proposal, we will further our research by developing analysis of p-bases and advancing techniques to approximate Hessians. These advances will be instrumental in Algorithm Design for SBBO. Our research in Algorithm Design focuses on blackboxes that only provide function values. Our recent research has: developed algorithms using subspace decompositions; analyzed algorithms using complex variables; and presented practical expositions on the proper comparison of optimization algorithms. In this proposal, we will expand our research by developing methods for high-dimension SBBO and developing an SBBO method that dynamically selects between direct-search and model-based strategies. These algorithms will be inspired from and applied to real-world optimization Applications. In Applications, we maintain an open-minded philosophy and apply our expertise in wide range of problems. In the past five years, we have modelled and optimized problems in: road design; seismic retrofitting; image denoising; and solid-state tank design. In this proposal, we will continue to connect our fundamental research with real-life problems. Our work will improve solution time and quality, thereby have immediate positive impacts on the engaged research areas. This interaction also iteratively inspires and directs our research in Analysis and Algorithm Design. In all activities, we will train HQP in valuable research skills. They will learn mathematical tools of optimization, master numerical and software development techniques, and gain skills in mathematical modelling and applied optimization. Many HQP will be trained through multidisciplinary co-supervision that will lead to skill acquisitions beyond SBBO.