Optimization Seminar
Location
Mathematics/Psychology : 401
Date & Time
November 3, 2016, 10:30 am – 12:00 pm
Description
Title: Approximating an inequality constrained optimization problem in function space
Speaker: Tom Seidman
Abstract:
For many problems in continuum mechanics the forces are obtainable as a gradient of a potential energy functional so minimization of this potential gives a balance of forces and so an equilibrium (steady state) configuration. Formally, this is much like finite-dimensional optimization, but some new considerations may arise e.g., even to define existence of a minimum requires selection of an appropriate function space as context. We consider here the modeling of an elastic rod flexible enough to be concerned with the possibility of self-contact and, in particular, want to know that a penalty function approach to approximation of the constraint leads to convergence of the configuration and the local contact force.
Speaker: Tom Seidman
Abstract:
For many problems in continuum mechanics the forces are obtainable as a gradient of a potential energy functional so minimization of this potential gives a balance of forces and so an equilibrium (steady state) configuration. Formally, this is much like finite-dimensional optimization, but some new considerations may arise e.g., even to define existence of a minimum requires selection of an appropriate function space as context. We consider here the modeling of an elastic rod flexible enough to be concerned with the possibility of self-contact and, in particular, want to know that a penalty function approach to approximation of the constraint leads to convergence of the configuration and the local contact force.
Tags:
