Developmental models for emergent computation

  • Authors:

  • Keith L. Downing

  • Affiliations:

  • The Norwegian University of Science and Technology, Trondheim, Norway

  • Venue:
  • ICES'03 Proceedings of the 5th international conference on Evolvable systems: from biology to hardware
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

The developmental metaphor has clear advantages for the design of physically-realizable artifacts, particularly when coupled with evolutionary algorithms. However, the embodiment of a developmental process in a purely computational system appears much more problematic, largely because embryogenesis evolved for the purpose of synthesizing 3-dimensional structure from a linear code, not for growing Universal Turing Machines. This research considers possible models of computational problem-solving based on the 5 primary developmental stages: cleavage division, patterning, cell differentiation, morphogenesis, and growth. A specific developmental approach to the NP-Complete problem, vertex cover (VC), is discussed, as well as a general model of developmental computation based on a multicellular extension of PUSH [12], a new stack-based language designed specifically for evolutionary computation.