Algorithmic feasibility of entity recognition in artificial life

  • Authors:
  • Janardan Misra

  • Affiliations:
  • HTS Research, Bangalore, India

  • Venue:
  • ECAL'09 Proceedings of the 10th European conference on Advances in artificial life: Darwin meets von Neumann - Volume Part II
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Observations are an essential implicit component of the simulation based artificial-life (ALife) studies by which entities are identified and their behavior is observed to uncover higher-level "emergent" phenomena. Building upon the axiomatic framework of Henz & Misra [2], we analyze computational complexity bounds for the algorithmic implementation of an observation process for an automated discovery of the life-like entities in arbitrary ALife models. Among other results of such analysis is the conclusion that the problem of entity recognition in a simulation using syntactic constraints is a NP-hard problem and therefore cannot always be solved in polynomial number of steps. The computational complexity bounds are established distinguishing further between those ALife models which allow entities with overlapping structures to coexist in a state and others which do not.