# Simulated Evolution

An algorithm for simulating evolution in a web browser

## Simulating Evolution

Evolution, is the change in heritable phenotype traits of biological populations over successive generations. Evolutionary processes give rise to diversity at every level of biological organization, including the level of species, individual organisms, and at the level of molecular evolution [1].

This article deals with the simulation of artificial life by using basic principles of evolution. The simulation presented here is a modified versions of a program first described by Michael Palmiter [2]. and later popularized by A. K. Dewdney in an article series published in the Scientific American magazine [3]. The simulation demonstrates the evolution of hunting behavior in an predator/prey situation. The simulation domain represents a small patch of a lake bottom. Microbes move across the screen living off a supply of bacteria. The bacteria are the only source of energy. The motion of a microbe is determined by a set of genes. The genes control the probability with wich a microbe will change its direction at a given time step.

## Simulation Rules

To create an evolution simulator we need a species with an easily observable behavioral pattern that is determined by a small set of genes. In the simulation laid out here the genes determine the preferred motion of a microbe. An individual competes with other members of the species for a food supply. If enough food was consumed by an individual it may create offspring after a certain amount of time. The offspring will have a slightly modified set of genes based on the parent genes. This implies that organisms which are successful in their search for food have a higher chance of creating offspring an thus passing on their genes.

### The Grid and Simulation time

The simulation deals with microbes living off an unlimited but randomly distributed food supply. The microbes live on a two dimensional grid. They can move freely into any of the eight neighboring cells. The edges of the grid are connected. If a microbe moves across the edge of the grid it immediately enters the grid back again on the opposing side. Microbes are rendered in blue, whilst food is rendered in green. For the sake of clarity microbes are rendered larger than the food but in terms of the simulation they still only occupy a single cell. There is no limit as to how many microbes a grid cell can contain. If multiple microbes are present they will be rendered above each other appearing as a single one.

The simulation runs in discrete time steps. In each time steps the microbes move along a straight line following their current direction of motion. If a microbe enters a cell occupied by food the food is consumed and the microbe gains a bit of additional energy. If multiple microbes enter the same cell in the same time step the first one to enter gets to eat the food.

## Microbes

Microbes constitute the active part of the simulation they move along the grid and they are subject to evolution during the course of the simulation.

### Basic Motion

Microbes can move freely to their neighboring cells. They travel alongside one of eight possible directions of motion. The motion direction is stored as an integer value. This value is used as an index to look up the specific ΔX and ΔY step sizes for moving over the grid. Motion is then implemented by adding these delta values to the current microbe grid position. The motion table is set up in a way that only motions to neighboring cells are possible.

### Genome

At every time step there is a certain probability with which a microbe will change its direction into another direction. The probabilities of all possible transitions represent the genome of the microbe. The genes will be labeled with the letter p and an index. The sum of all probabilities (genes) is one: $\sum_{n=0}^7 p_n = 1$

Each gene represents a segment of the interval [0,1] whose length is proportional to the gene value. The following example shows a random genome in which the gene sizes are rendered proportional to their value. Each gene represents the likelihood of a certain change in direction:

In order to compute the new direction with a probability proportional to the corresponding gene value a random value r between 0 and 1 is computed:

$r = rnd(1)$

Then the index i of the gene is computed. This index is the smallest possible i for which the following inequality holds true:

$r \lt \sum_{n=0}^i p_{n}$ I the example above i equals 3. This is the index of the directional change be applied. The direction d of the microbe in the next time step is then computed by adding i to the current direction and determining the remainder of the division by 8 (the total number of possible changes): $d_{t+1} = (d_{t} + i)\mod 8$

With this equation we can easily deduce the meaning of each of the genes:

### Energy management and the Microbe life cycle

Each microbe has a finite amount of energy and with each time step a small fixed amount of energy is subtracted to account for the "energy cost" of living and moving. In addition changing the direction also requires energy. The more drastic the directional change is the more energy is required. A microbe can boost its energy by consuming food. It needs a constant supply of food in order to continue living since it will die if it has exhausted its energy supply. A microbe stops eating once it has gained a certain amount of energy.

Once a microbe has reached a certain age it will create offspring if it has enough energy. The energy is evenly split between the parent and the child. The child will start with a genome that is based on the parents genome but contains a single slightly modified gene. The following table provides an overview on the parameters related to the microbes energy management:

## Food Supply

The food supply is unlimited. Each grid cell can either be empty or contain food. Food does not have a life cycle of its own. If a cell contains food it remains there until a microbe eats it. In order to put the population under evolutionary stress The simulation can spawn the food in three distinct patterns:

• Uniformly distributed
• More food spawns in a rectangular area in the center
• Food Spawns primarily along horizontal and vertical lines

You might also like: