## Modelling places of employment

The number of
employment places in a location can be obtained from census information. It is
then possible to randomly distribute points within mesh blocks to represent
each of the employment opportunities. The travel time across the network to
each of these points can then be calculated. Repeating this process with multiple
iterations of the random distribution (Monte Carlo simulation) would yield a
statistically robust result.

However, the
computational power and time required to calculate such a result for a city the
size of Christchurch rapidly expands to prohibitive levels. For example, there are
approximately 145,000 jobs in Christchurch, which could mean a potential
network calculation of up to 3000 * 145,000 = 435 million calculations.
Conversely, by modelling places of employment at the centroids of mesh blocks
they reside within, the maximum potential calculations can be reduced to 3000 *
3000 = 9 million calculations. This is achieved by solving the network path
cost to the meshblock centroid, applying the negative exponential equation to
the path cost and then multiplying this result by the number of jobs in the
meshblock. Taking the sum of all these values yields the number of accessible
co-located job equivalents.