Objectives and visions

A long-term goal of this open-source project is to provide a high-performance particle-based simulation platform in order to tackle various questions in the field of Artificial Life. ALIEN has been specifically designed to facilitate open-ended evolution simulations. Networks of particles can form arbitrarily complex physical structures and behavior patters controlled by neural networks. Besides that it can also be seen as a kind of game where one can play, experiment and explore digital worlds and ecosystems without a specific aim. The main design goals of the simulator are as follows:

  • Utilization of CUDA and OpenGL for simulation and rendering to support large worlds with millions of particles.
  • No hard-coded behavior patterns. Everything should emerge from the interaction of particles.
  • An enjoyable user interface in spite of the complexity of the simulation model and the wide range of adjustment options.
  • Visually appealing post-processing of the rendered simulation data.

In the following some visions targeting particular aspects of Artificial Life are highlighted, where ALIEN might be useful as a simulation platform.

This refers to the aim of simulating life on a digital and non-carbon basis. The underlying assumption is that it is not crucial for living structures to be composed of physical entities of the real-world. ALIEN utilizes a physics engine where the interacting particles can be programmed to execute tasks. It allows one to create complex structures, which can interact in a common environment.

A goal in this respect is to construct artificial worlds that exhibit self-organizing and/or self-repairing structures maintaining a higher degree of order in comparison to their environment. More specifically, the search for machines with the following characteristics is strived for:

  • control their movement,
  • search for nutrients,
  • interact with their vicinity,
  • replicate and grow if enough resources are available,
  • possess self-repairing capabilities,
  • and possibly show intelligent behavior.

It is still an unsolved question how complex system which we would consider as life can emerge from ancestry structures of less complexity. One possibility is that some kind of prebiotic evolution without RNA and DNA comes into play. In this spirit the emergence of life can be considered as a gradual process and not as a spontaneous miracle.

One of our objectives is to construct an environment where such conditions are simulated.

It is not fully understood how highly-organized systems such as biological cells are assembled from component which follow simple rules. From theoretical point of view, at least two types of self-replicating machines are conceivable: replication by self-inspection as well as replication by using some information storage such as DNA. While the first one requires intricated self-analyzing techniques and suffers some fundamental limitations (difficult for large structures and vulmerability to errors), the second one paves the way to open-ended evolution.

A goal in this context is to create machines which use (variants of) these replication techniques in order to get an impression of how life can be.

It has become a well-known fact that evolution driven by natural selection is the crucial force causative for the complexity and diversity of our ecosystem reaching from relatively simple prokaryotes such as bacteria to most involved multicellular organisms.

In ALIEN everything consists of particles. Merely the connections and the internal states of the them determines the life-like behavior of compositions. By designing machines that can self-replicate, one can conduct evolutionary studies just by simulating them in a common environment exposed to mutations. No external selection based on fitness calculation and genotype-phenotype mapping is necessary.

Evolution by natural selection and mutation seems to be the cornerstone of increasing complexity. As soon as a digital ecosystem containing self-replicating machines has been built up, long time studies can be made. The effect of mutation rates, environmental noises, physical parameters, evolutionary arms races and symbiosys are aimed to be investigated.

Further questions about stability of such systems in comparison to certain parameters can also be tackled in this context.

Nach oben