• home
• current students
• graduate
• events

Presenter: Mr. Juan Carretero - Department of Mechanical Engineering
Supervisor:

Date: Wed, May 1, 2002
Time: 13:30:00 - 14:30:00
Place: EOW 430

ABSTRACT

Abstract:

Determining the minimum distance between two objects is a problem that has been solved using many different approaches. Most methods proposed so far are, in essence, limited to solve the problem amongst convex polyhedra. Thus, to deal with concave objects, these methods partition concave objects into convex sub-objects and solve the convex problem between all possible sub-object combinations. This adds a large computational expense, especially when the concave objects in the scene are complicated, or when concave quadratically bound objects are to be linearized.

In this seminar, a novel optimization-based formulation to solve the minimum distance problem without the need for partitioning concave objects into convex sub-objects is presented. In this new formulation the geometries of the objects are replaced by large sets of points arranged in surface meshes. Then, the pair of points (one on each object) that minimizes the distance between objects is sought for using a combinatorial optimization algorithm. Since the optimization problem is not unimodal (i.e., has more than one local minimum point), global optimization techniques are used.

The talk focuses on a Genetic Algorithm (GA) implementation of the minimum distance problem. Some basic principles of GAs are presented followed by their application to the minimum distance problem. Next, some advanced GA techniques such as the niche formation are described. These techniques are then applied to the minimum distance problem in order to allow the distance algorithm to track multiple minima.

The talk concludes with a series of numerical examples in which a preliminary Matlab implementation of the proposed algorithm is shown to be equivalent, in terms of robustness and computational efficiency, to some conventional approaches.

FREE AND OPEN TO THE PUBLIC

Coffee and Cookies Will be Provided

For Further Information Please Contact: Student (721-8923)