Files
Abstract
Genetic algorithms, and other evolutionary mathematical algorithms, are important tools to finding
approximate solutions to complex problems. These are in the category of NP-Hard
problems, which can not be solved by direct searches. In this dissertation genetic
algorithms are used to find (optimal, perhaps) solutions in different areas of science.
These problems are explained in the introduction and in the subsequent chapters.
Detailed use of the genetic algorithms is presented in several chapters, from real-time
system scheduling analysis in sensitivity analysis, to nuclear magnetic spectral assignment,
and in a classic NP-Hard problem, the
maximally spanning backbone $k$-tree problem.
The use of genetic algorithms is demonstrated to produce better results than earlier works
in these fields. For example, in real-time systems the processor utilization is higher,
in NMR an automated assignment package is presented in both large and small proteins, and
the last project, the maximally spanning $k$-tree problem, more and better solutions are
found.
The presentation in this dissertation doesn't cover all of the work completed during the course of
Ph.D. completion. However, additional work is described in an appendix.
approximate solutions to complex problems. These are in the category of NP-Hard
problems, which can not be solved by direct searches. In this dissertation genetic
algorithms are used to find (optimal, perhaps) solutions in different areas of science.
These problems are explained in the introduction and in the subsequent chapters.
Detailed use of the genetic algorithms is presented in several chapters, from real-time
system scheduling analysis in sensitivity analysis, to nuclear magnetic spectral assignment,
and in a classic NP-Hard problem, the
maximally spanning backbone $k$-tree problem.
The use of genetic algorithms is demonstrated to produce better results than earlier works
in these fields. For example, in real-time systems the processor utilization is higher,
in NMR an automated assignment package is presented in both large and small proteins, and
the last project, the maximally spanning $k$-tree problem, more and better solutions are
found.
The presentation in this dissertation doesn't cover all of the work completed during the course of
Ph.D. completion. However, additional work is described in an appendix.