6.4.3 Logical Tests
Logical tests are designed to assure:
- Parametric Verification. The values of each parameter included in a module, sector or model must be tested for consistency with what is known about those parameters. In the absence of hard (numerical) data, parametric verification must be based at least on descriptive knowledge combined with reliable estimates of specific parametric values.
- Dimensional Integrity and Unit Consistency. All algebraic expressions, arithmetic integrations and parametric values must be consistent both in terms of the units involved and, in the case of arrays, have consistent and compatible dimensions.
- Correct Sequence of Calculation. The sequence of flows into and out of stocks (levels or accumulators) is important. This sequence must be tested for all stocks.
- Stochastic / Statistical Character. The form of output from any module, sector or model must have the same characteristics as the real-world situations they represent. The modelling output must be able to withstand statistical tests applicable to real-world data.
