The Kungsängen site is 9 km2. The Settlement Simulation was developed to run on a 1km2 field size ensuring all production for those residents is within a walk-able distance.
While the simulation will run on larger sites, (a test for this site was run on a 2 km2 field) we found that large continuous productive plots, and associated Wholesale Nodes, begin to aggregate away from the denser dwelling clusters, which form around the attractor seeds. Also, computation time grows considerable with each km2 added, as the various evaluations that are done across the simulation field must check more and more cells at every run count.
Therefore, our strategy for scaling will be to tile the site, sequentially adding 1 km2 tiles, taking initialisation data from the boundary of each previous tile.
Working in tiles will ensure that the required production capacity and associated Wholesale Nodes are distributed within 15min walking distances of residents.
On our Kungsängen site, tiles will be assigned different Density (D) and Friendliness (F) values
providing residents a choice of 'neighbourhoods' with varying spatial characters city-wide. The most compact neighbourhoods will be nearest the train station and another at the south end of the shoreline with views back to the city of Stockholm. The western-most tiles near the wooded patches will be the most dispersed.
The distribution of Densities and Friendliness values across the tiles should really be negotiated by the tiles as they develop, in a larger scale cellular automata type calculation (as happens cell by cell within the simulation.)
Also to be investigated is how the outputs of the early tiles may be re-informed by the results of the latter tiles.