Gradient Flows as a Selection Procedure for Equilibria of Nonconvex Energies

A Variational Principle for Gradient Flows of Nonconvex Energies

We present a variational approach to gradient flows of energies of the form E = φ1−φ2 where φ1, φ2 are convex functionals on a Hilbert space. A global parameter-dependent functional over trajectories is proved to admit minimizers. These minimizers converge up to subsequences to gradient-flow trajectories as the parameter tends to zero. These results apply in particular to the case of non λ-conv...

An Efficient Neurodynamic Scheme for Solving a Class of Nonconvex Nonlinear Optimization Problems

‎By p-power (or partial p-power) transformation‎, ‎the Lagrangian function in nonconvex optimization problem becomes locally convex‎. ‎In this paper‎, ‎we present a neural network based on an NCP function for solving the nonconvex optimization problem‎. An important feature of this neural network is the one-to-one correspondence between its equilibria and KKT points of the nonconvex optimizatio...

Regularized Wulff Flows, Nonconvex Energies and Backwards Parabolic Equations

In this paper we propose a method of regularizing the backwards parabolic partial differential equations that arise from using gradient descent to minimize surface energy integrals within a level set framework in 2 and 3 dimensions. The proposed regularization energy is a functional of the mean curvature of the surface. Our method uses a local level set technique to evolve the resulting fourth ...

FLUENCE MAP OPTIMIZATION IN INTENSITY MODULATED RADIATION THERAPY FOR FUZZY TARGET DOSE

Although many methods exist for intensity modulated radiotherapy (IMRT) fluence map optimization for crisp data, based on clinical practice, some of the involved parameters are fuzzy. In this paper, the best fluence maps for an IMRT procedure were identifed as a solution of an optimization problem with a quadratic objective function, where the prescribed target dose vector was fuzzy. First, a d...

An effective optimization algorithm for locally nonconvex Lipschitz functions based on mollifier subgradients