What is simulation?

Simulation is defined by Sheldon Ross as a "method of empirically determining probabilities by means of experimentation" and it is with this sense of simulation that we approach the problem of ecosystem modeling. The "experiments" we will conduct are computer representations of the ecosystems. For illustrative purposes, the models will be simplified but they should convey some notion of the computational power available to ecologists through computers.

Below is a demonstration of a very basic simulator. Notice that the Simple Discrete Event Simulator (SDES) display is divided into three large sections. The top section demonstrates the way in which the simulator schedules "events." The leftside is a list of the actions performed by the current "event." The right side is an animation of the simulation.

This simulation mimics the motion of four abstract entities residing in some geographical region. The region is divided into a 22-by-33 grid of "cells." Each cell represents a square in the ecosystem we are modeling; therefore, each cell has the physical characteristics of a quadrat in the modeled ecosystem.

The rules for movement are as follows:

1) The entity will rest for some random amount of time determined by an exponential distribution

2) When the entity moves, it will move to one of the eight adjacent cells with a uniform random distribution.

3) When an entity moves off an edge in the simulation space it is reintroduced on the opposite side, thus the space is acutally a flattened torus.

If you click on the "STEP" button you will see a demonstration of SDES in action. Notice that the event in the front of the list is removed, the entity affected by the event reacts to the event, generates a new event and the new event is inserted into the event list. These simple actions are the basis for our approach to the problem of ecological simulation.


[SECTION 1] PREVIOUS [SIMULATION PITFALLS] [SECTION 2]