Miro Lahdenmäki

Problems in simulating real life

We've learned from movies and science-fiction literature that one day scientists will succeed in developing intelligent machines that can do all things better than human beings. Let's take a step back for a moment and see if we'll soon get to meet these all-powerful beings.

The whole consept of creating artificial life is fascinating to me. I’m also interested in how the diverse nature works. Therefore I decided to find out why is simulating life so difficult. Internet was the obvious place to look for information about this. For this survey I searched the internet and found several sites about artificial life and simulating life. I've included some of these links to the end of this text so look there for more information.

The purpose of this survey is to give us a glimpse to simulating real life and the problems that are faced. I'm going to tell you what kind of problems have to be faced in simulating real life, what limitations there are and finally I'm going to try to give you an idea on how these problems could be solved. We'll find out that it is not at all straightforward to simulate real life. There are a number of issues that have to be taken into consideration.

There are those who think it is impossible to create artificial life. Some suspect life is a property of the material structure of a living system and therefore will only spring from organic matter. Others believe that life involves a soul of some sort and hence is impossible to be created artificially.

Those who do believe in the possibility of creating artificial life perceive life as an abstract form of organization that can be realised in diverse media; artificial as well as natural.

There are two main areas of life that have been attempted to simulate. Nature and intelligence. In this presentation I will concentrate on nature and artificial life. Artificial intelligence is another subject and I won't be covering it here. Rest assured artificial life is just as interesting.

What is essential in simulating life is to know how things work in real life. Simulating nature can only be done according to the existing knowledge of physics. And there are a number of things that haven’t been figured out completely. For example how proteins or genes work exactly. This raises many questions as we will see later on.

The complexity of the systems involved is a major problem. The number of elements in these systems is unimaginable as the elements are so infinitely small. For example a single cell contains billions (109) of water molecules that are essential to the operation of a cell. In a human sized creature there are about one hundred trillion cells (1014). To give you an idea of how big a number this is. It is also the number of stars in our galaxy. To simulate action we have to implement time to our system. Moreover we have to use a very short timestep of about a femtosecond (10-15) as the atoms of a liquid interact strongly with each other. At the moment quantum mechanics is practically the only theory with which we can explain the formation of molecules in a satisfactory manner. Additionally we can only estimate molecule systems using numerical methods. This can become quite time consuming.

Simulating even a single cell at this level of detail is simply impossible with the computational power of today. Which brings us to another problem. The computers of today are not fast enough to be able to calculate this kind of complex systems. We will have to wait for the quantum computers to simulate anything but the most simple systems.

One of the challenges is that you need to have an understanding of several fields of science to be able to understand the methods of creating artificial life. You need computer science, advanced mathematics, biology, psychology, physics perhaps even philosophy.

If we want to simulate life as realistically as possible we should start with the basic element of life - the cell. The most primitive life forms are formed from a single cell. Such as the amoeba or the yeast. The size of a cell can vary, but usually it’s less than one tenth of a millimeter. The number of cells in an individual can vary. As I mentioned earlier there are about one hundred trillion cells (1014) in a human.

To mimic a cell you somehow have to implement the properties that make a cell a cell. Such as:

DNA, the exact genetic instructions required to create a particular organism with its own unique traits.
Examples

The genome contains the master blueprint for all cellular structures and activities for the lifetime of the cell or organism.

Genes, carry the information required for constructing proteins, which provide the structural components of cells and tissues as well as enzymes for essential biochemical reactions.
Example

Problem: Only about 10% of the genome is known to include the protein coding sequences (exons) of genes. The balance of the genome is thought to consist of other noncoding regions (such as control sequences and intergenic regions), whose functions are unclear. So they cannot be simulated.

Multiplication of cells, the multiplication of ordinary cells is easy to simulate.

Metabolism

Inheritance of traits, to have evolution we have to be able to transfer the good traits to descending lines.

To have evolution we also need variation. This can be accomplished with crossing over and mutation. That is mixing parts of the genomes or adding random changes to the genome. The problem here is how to give the entity input of its virtual environment so that it might evolve into something useful.

There are different possibilities to simulate real life. They are all simplifications of the actual thing. For example we could mimic the way DNA controls life or we could mimic the way genes control life or we could mimic cells with their traits. There is a scientific debate going on whether it is at all possible to mimic life with these simplifications.

Let’s sum up what we have learned through this survey.

Problems

Simulating real life is quite complex because of the immense amount of components in real life systems.

There are a number of things we still can’t fully explain even in the classical sciences.

Many different fields are combined in the study of artificial life.

Limitations

The existing knowledge of physics limits what we can simulate.

Simulating life on molecular level is very complex and requires tremendous computing power which is not available today.

The writer of this text is not in anyway an authority in the field of artificial intelligence. Please take this into consideration when reading or quoting this text.

This text was part of a presentation for a language course - not a scientific publication.

Miro Lahdenmäki, 2nd year student of computer sciences, 5.11.2001.

http://www.alife.org/ (general information)
http://mitpress.mit.edu/journals/ARTL/Bedau.pdf (Open Problems in Artificial Life)
http://bioinfo.mbb.yale.edu/sciam/sa-water.html (simulating water and molecules)
http://dmoz.org/Computers/Artificial_Life/Artificial_Worlds/ (Computer programs)
http://www.webslave.dircon.co.uk/alife/intro.html (introduction)