Evolutionary complexity for protection of critical assets.

Battaile, Corbett Chandler ; Chandross, Michael Evan

Sandia National Laboratories

关键词: Evolutionary Programming (Computer Science);    Genetics;    Genetic Algorithms.;    59 Basic Biological Sciences;    Algorithms;   

DOI  :  10.2172/919199
RP-ID  :  SAND2004-5372
RP-ID  :  AC04-94AL85000
RP-ID  :  919199

美国|英语

来源: UNT Digital Library

【 摘 要 】

This report summarizes the work performed as part of a one-year LDRD project, 'Evolutionary Complexity for Protection of Critical Assets.' A brief introduction is given to the topics of genetic algorithms and genetic programming, followed by a discussion of relevant results obtained during the project's research, and finally the conclusions drawn from those results. The focus is on using genetic programming to evolve solutions for relatively simple algebraic equations as a prototype application for evolving complexity in computer codes. The results were obtained using the lil-gp genetic program, a C code for evolving solutions to user-defined problems and functions. These results suggest that genetic programs are not well-suited to evolving complexity for critical asset protection because they cannot efficiently evolve solutions to complex problems, and introduce unacceptable performance penalties into solutions for simple ones.