Floatspace

 Boats


^ Click this image to see the animated demonstration.
This simulation originally started as an exercise to find the optimal placement of a number of boat campers in a recreational anchorage.   The methods used to find this placement centered on establishing rules for each boat to find their satisfaction with any particular arrangement based on the suitability of their location in the anchorage, and how much they are aggravated by the boats with which they share the anchorage.  Since the rules that define the mutual aggravation of any two boats is largely a function of the distance between the two boats (boats who are farther apart will annoy each other less, all else being equal), the overall situation in the anchorage can be improved by altering the distances between the individual boats.

In this simulation, boats (red dots) are given rules by which they can assess their level of annoyance with their neighbors.  At each time step, each boat will try to move away from the most annoying of the neighbors, while trying to stay in suitable habitat (area good for camping).   Suitability is defined by depth of the water in the anchorage (shown as shades of blue); in future simulations, other factors such as bottom type, land ownership, and presence of seagrass beds will be considered.

The eventual goal of this simulation is to find the maximum number of boats that can coexist in an anchorage with the minimum level of aggravation between them.  For the present discussion, this is presented as a way to have agents moving around in a GIS-generated landscape.  At this stage, agents are simply reading the features of a static, unchanging landscape.

In this simulation, the landscape is represented in a raster format.  However, rather than have a series of raster layers (as is done in normal raster-GIS models), each pixel is an object with a list of variables describing its depth, bottom type, ownership, etc.

In the next page, we discuss  grass agents.
 


[Top of the page]
[Previous] [Floatspace Home] [Next]

 [Kenge home page]