A Case For Exhaustive Optimization, Sanza Kazadi, Michele Lee, Lauren Lee
Jun 2005
A Case For Exhaustive Optimization, Sanza Kazadi, Michele Lee, Lauren Lee
Sanza Kazadi
Evolutionary algorithms have enjoyed a great success in a variety of different fields ranging from numerical optimization to general creative design. However, to date, the question of why this success is possible has never been adequately determined. In this paper, we examine two algorithms: a genetic algorithm and a pseudo-exhaustive search algorithm dubbed Directed Exhaustive Search. We examine the GA's apparent ability to compound individual mutations and its role in the GA's optimization. We then explore the use of the DES algorithm using a suitably altered mutation operator mimicking the GA's surreptitious compounding of the mutation operator. We find that …
On The Dynamics Of Puck Clustering Systems, S. Kazadi, M. Chung, B. Lee, R. Cho
Dec 2003
On The Dynamics Of Puck Clustering Systems, S. Kazadi, M. Chung, B. Lee, R. Cho
Sanza Kazadi
We examine the theoretical foundations for the dynamics of puck clustering systems. Key in this investigation is the development of methods of controlling variance in cluster size, an important precursor to swarm-mediated clustering. We derive conditions under which clustering can take place in a general framework, and demonstrate two different behavioral regimes for clustering systems.
Swarm-Mediated Cluster-Based Construction, S. Kazadi, J. Wigglesworth, A. Grosz, A. Lim, D. Vitullo
Dec 2003
Swarm-Mediated Cluster-Based Construction, S. Kazadi, J. Wigglesworth, A. Grosz, A. Lim, D. Vitullo
Sanza Kazadi
We investigate the use of swarm clustering algorithms in the design of simple robots capable of carrying out swarm-mediated construction. Methods for generating multiple clusters of predetermined size are developed. Relative cluster motion algorithms are also developed and explored. All robotic algorithms are predicated on the use of only robots utilizing no processing, gps, or explicit communication. Simple stigmergic communication and minimal sensing capabilities are used exclusively. We demonstrate swarms of minimal agents building equilateral triangles, squares, and pentagons. Future use of these methods in the design of more sophisticated construction techniques is discussed.
A Study Of Evolutionary Acceleration, Sanza Kazadi, Sparta Cheung, Cristal Ogletree, Shirley Kim, Catherine Lee, Andrew Min
Jul 2003
A Study Of Evolutionary Acceleration, Sanza Kazadi, Sparta Cheung, Cristal Ogletree, Shirley Kim, Catherine Lee, Andrew Min
Sanza Kazadi
We investigate the phenomenon of numerical evolutionary acceleration. This phenomenon is a simple consequence of numerical analysis of the probabilities of evolving independent parts of a complex system in the presence of evolutionary epochs. The epoch mechanism allows the newly evolved structure to become part of the overall system design of all elements of the population. We demonstrate that this phenomenon not only exists in real evolving systems, but that evolutionary acceleration dwarfs the group mechanism for some complex structures.
On The Convergence Of Puck Clustering Systems, S. Kazadi, A. Abdul-Khaliq, R. Goodman
Dec 2001
On The Convergence Of Puck Clustering Systems, S. Kazadi, A. Abdul-Khaliq, R. Goodman
Sanza Kazadi
Puck clustering involves the physical relocation of small objects known as ‘pucks’ from random positions to a central location, which need not be specified a priori. The evolution of systems of clusters of pucks under the action of robots capable of moving pucks from several locations is considered. A general set of conditions by which a puck collection system may be seen to evolve to a one-cluster system is developed. General conditions leading to clustering behavior in the presence and absence of non-embodied agents are derived. Conditions leading to more efficient algorithms are also derived. Several examples are given to …
Insufficiency Of Piecewise Evolution, Sanza Kazadi, Yan Qi, Isaac Park, Nancy Huang, Paul Hwu, Brian Kwan, Wayne Lue, Hubert Li
Jul 2001
Insufficiency Of Piecewise Evolution, Sanza Kazadi, Yan Qi, Isaac Park, Nancy Huang, Paul Hwu, Brian Kwan, Wayne Lue, Hubert Li
Sanza Kazadi
We describe an evolutionary design paradigm called piecewise evolution.
This evolutionary design paradigm allows the gradual evolution of a
piece of hardware using discrete functional stages. The paradigm
removes designs from a population of designs which effectively lose
functionality already discovered. Significant improvement s in the
evolution time of simple one-bit adders are reported. However,
evolution of more complex devices does not seem to share the
improvements in evolutionary speed of simple devices. These results are
discussed in the context of epistasis and deceptiveness.
Conjugate Schema In Genetic Search, Sanza Kazadi
Jul 1997
Conjugate Schema In Genetic Search, Sanza Kazadi
Sanza Kazadi
Functional optimization is profoundly affected by the use of specific encodings. In one encoding, a particular problem may be simple to undertake, while in another encoding, the problem may be intractible. Genetic algorithms solve optimization problems by making use of schema. By locating schema in a solution vector, the paradigm can settle on a solution that makes use of several schema and combines them via crossover.
We propose a generalization of this idea, conjugate schema. Conjugate schema are disjoint subsets of the basis over which the fitness function can be written as a sum of smaller dimensional functions. We find …
