Title :
Evolution of L-systems for compact virtual landscape generation
Author :
Ashlock, Daniel A. ; Gent, Stephen P. ; Bryden, Kenneth M.
Author_Institution :
Math. & Stat., Guelph Univ., Ont., Canada
Abstract :
An L-system or Lindenmayer system consists of a grammar and an interpreter. The grammar contains an axiom, usually a short string, that the grammar expands into a long, complex string. The interpreter then renders the string into an object. A midpoint L-system is a generalization of L-systems to two-dimensional arrays of characters inspired by midpoint displacement fractals. This study presents a system for simultaneously evolving the rules and interpreter for a midpoint L-system that encodes a desired landscape. Unlike a midpoint displacement fractal a midpoint L-system is deterministic and can be evolved to yield fixed, complex shapes. The fractal character of a midpoint L-system permits the storage of a large complex virtual landscape in a small data object. The level of detail rendered by an L-system can be changed rapidly and, with a fast graphics engine, dynamically. This study introduces midpoint L-systems, gives techniques for evolving them, and demonstrates those techniques on trial landscapes that resemble hills and craters. The application of this work is for virtual reality where midpoint L-systems allow a designer to select from many rugged versions of a landscape without requiring vast amounts of storage or machine time to render them.
Keywords :
evolutionary computation; graph grammars; virtual reality; L-system evolution; L-system generalization; Lindenmayer system; axiom; compact virtual landscape generation; complex string; deterministic midpoint L-system; evolving interpreter; evolving rules; fast graphics engine; fractal character; grammar; landscape encoding; large complex virtual landscape; long string; midpoint displacement fractal; rugged landscape version; short string; small data object; trial landscape; two-dimensional character array; virtual reality; Engines; Fractals; Graphics; Mathematics; Mechanical engineering; Probability distribution; Rendering (computer graphics); Shape; Statistics; Virtual reality;
Conference_Titel :
Evolutionary Computation, 2005. The 2005 IEEE Congress on
Print_ISBN :
0-7803-9363-5
DOI :
10.1109/CEC.2005.1555041