Date of Award


Document Type


Degree Name

Master of Science


Department of Electrical and Computer Engineering

First Advisor

Gary B. Lamont, PhD


The protein folding problem consists of attempting to determine the native conformation of a protein given its primary structure. This study examines various methods of hybridizing a genetic algorithm implementation in order to minimize an energy function and predict the conformation (structure) of Met-enkephalin. Genetic Algorithms are semi-optimal algorithms designed to explore and exploit a search space. The genetic algorithm uses selection, recombination, and mutation operators on populations of strings which represent possible solutions to the given problem. One step in solving the protein folding problem is the design of efficient energy minimization techniques. A conjugate gradient minimization technique is described and tested with different replacement frequencies. Baidwinian, Lamarckian, and probabilistic Lamarckian evolution are all tested. Another extension of simple genetic algorithms can be accomplished with niching. Niching works by de-emphasizing solutions based on their proximity to other solutions in the space. Several variations of niching are tested. Experiments are conducted to determine the benefits of each hybridization technique versus each other and versus the genetic algorithm by itself. The experiments are geared toward trying to find the lowest possible energy and hence the minimum conformation of Met-enkephalin. In the experiments, probabilistic Lamarckian strategies were successful in achieving energies below that of the published minimum in QUANTA.

AFIT Designator


DTIC Accession Number