A class of trust-region methods is presented for solving unconstrained nonlinear and possibly nonconvex discretized optimization problems, like those arising in systems governed by...
Serge Gratton, Annick Sartenaer, Philippe L. Toint
Many real-world search and optimization problems naturally involve constraint handling. Recently, quite a few heuristic methods were proposed to solve the nonlinear constrained op...
—In this paper we examine a technique by which fault tolerance can be embedded into a feedforward network leading to a network tolerant to the loss of a node and its associated w...
In this paper, we present a feature-based free-form shape modelling technique based on solving a fundamental problem of reconstructing the depth information from 2D sketch planes....
This paper describes a Parallel Space Decomposition (PSD) technique for the Mesh Adaptive Direct Search (MADS) algorithm. MADS extends Generalized Pattern Search for constrained n...