notes for research gathering on 2008-11-28 at FoAM related » project groworld

1. form → (…)
2. function → more involved

“The power of L systems goes beyond their capability to generate realistic images of plants […] they also provide a model of their operation, including processes such as tropisms, abscission, signal propagation, or watering.”

–Roberto S. Ferrero

There is a number of applications in which L systems plays an important role as a biological model:

1.- structural models of trees integrated in more complex forest ecosystem simulations,
2.- identifying plant response to insect attack,
3.- design of new varieties of plants,
4.- reconstruction of extinct plant species,
5.- crop yield prediction,
6.- classification of branching patterns in inflorescences,
7.- simulation of fungal growth, or
8.- computer aided learning for farm managers."

via. Roberto S. Ferrero

Following the structure of “algorithmic beauty of plants”…

• deterministic & context free (simple) l-systems

• variations & randomness

==== Context-sensitive L-systems

• symbol replacement dep. on context, or previous states
• can be used to model signal propogation

==== Parametric L-systems

• elements can be parameterised, eg. segment lengths
• continous development, motion, growth or diffusion

==== Developmental models (ABOP 3.1 →)

• L-systems provide good structural models, how can we model growth and changes over time?
• multi-level models
  • partial l-system (structural - non deterministic)
  • l-system schemata (control mechanisms, resolve n.d., temporal aspects)
  • complete l-system (geometric info, growth rates/branching/appearance)
• compound flowering structures (inflorescences)
• Phyllotaxis. In order to describe the pattern of florets (or seeds) in a sunflower head,

Vogel proposed the formula φ=n∗137.5◦, r=c√n {cf. ABOP 4.1}

• surface models (ABOP 5)
  • the shape as well as size of plant organs may change over time
  • leaf types

• The original formalism of L-systems provides a model of development that is discrete both in time and space.
  • the model states are known only at specific time intervals.
  • spatial organisation is finite
  • Parametric L-systems remove the limits imposed by discrete spatial representation
    • assign continuous attributes to model components
    • model states are still known only in discrete time intervals.

==== Timed DOL systems

• expression of production rules is still discrete, yet each cell has its own 'lifetime' as specified in teh production rule
• symbols represent cells that elongate during their lifetime and divide upon reaching terminal age.
• effect of aging, and gradual development can be modelled
• 'young' cell can be replaced by mature form → fruits

==== open L-systems

• what advantages/disadvantages over l-systyems?
• why another abstraction?

• benefits
• tradeoffs

• http://www.blprnt.com/processing/cherrytree/
• http://www.blprnt.com/processing/birchtree/
• karl sims - panspermia (1990) http://www.youtube.com/watch?v=AgeuRukfZLE ~1:00→