Several model inputs specify a quantity over a period, e.g., Service Traffic per Connection, Service Rental Tariff and Resource Maintenance Cost. All of these are given as annual quantities, e.g., the tariff charged for being connected for a year, or the cost arising from a year’s operation.
| Element | Input |
|---|
| Global |
Tax and Interest Rates
Slack Past-Usage Significance |
| Service |
Traffic per Connection
Cost Independent Rental Tariff
Initial Cost Dependent Rental Tariff
Max. Change (Connection/Rental/Usage)
Admin Cost
Churn Proportion
|
| Resource |
Depreciation Rate
Maintenance Cost, Churn Cost, Rental Cost, Usage Cost and Operations Cost
Redundant Unit Decomm. Prop.
Annual EOS and Past Volume Significance
|
Some inputs which specify a quantity over a period
When STEM 5.4 calculated the corresponding results for a year, the calculations were simplified and tailored to this annual interpretation. In order for the new model engine to calculate over shorter time periods, it must take account of the length of each period (measured in days). For example, the Rental Revenue result for a given period is calculated as the product of the Rental Tariff input, average connections over the period, and the length of the period in years (i.e., the number of days divided by 360). Although many issues are raised by this shorter time period development, the primary goal for the STEM 6.0 release was a model engine which would make proper allowance for all of these effects, so that the results of a shorter time period model would be broadly comparable with those of the corresponding annual model. It is likely that further refinements will be made in a future release, particularly regarding consistency between results summed over months or quarters and the corresponding raw annual results.
Demand and supply
A conventional annual STEM model is rather forgiving over the precise timing of installation, in the sense that the unrealistically tight binding of supply to demand is masked by the vagueness of exactly when in a year demand increases and equipment is installed. The scope for precision with the new model highlights this problem; and so, as a first step towards greater realism, a new supply-forecasting feature has been added. We hope to introduce a model for equipment lead times in a future release. For now, one must either interpret demand inputs as connected customers and carried traffic, or view the subsequent installation as the result of perfect, just-in-time planning.