Over the last decade, there has been considerable development in investigating methods of symbolic mathematics in many application areas of computer science and artificial intelligence, for example, engineering design - solid and geometric modeling, robotics and motion planning, and machine vision. Most of this research, however, has been theoretical, and its application to real problems has been limited because of the slow progress made in applying the combinatorially expensive symbolic techniques to real problems. Also there are no mechanisms in computer algebra and other symbolic techniques to handle issues of data error and computational precision. Typically such problems are addressed in practice by using numerical methods to approximate the underlying mathematical models. This workshop will focus on approaches for systematic integration of symbolic techniques with numerical techniques. Symbolic methods can provide insight into the structure and stability of problem solutions while numerical methods are efficient for handling the thousands of data elements which arise in practical applications. The workshop will bring together experts from numerical computation, symbolic computation, and application areas using such methods. Some possible topics of discussion are: methods for solving nonlinear equations, homotopy methods, nonlinear optimization, exploitation of geometric invariants, differential theories.
