You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The general idea is to have a table where each row specifies a particular parameter and a value for it to take on, and then each column corresponds to the other parameters in the model. When a test condition has implications for the immediate values of one of the other parameters, these are specified in a cell. In the example from the above paper, the test matrix could look like:
Value
Infection
Recovery
Susceptible
0
0
-
Total Population
0
0
0
Infected
0
0
0
Contact Rate
0
0
-
Recovery Time
inf
-
0
Infectivity
0
0
-
To make this type of testing easier, we should also have a function which creates a template for this type of matrix, and saves it to a file that the modeler can edit and add their constraints to. The most basic template would just take extreme conditions for each value and make them into rows.
A more advance templating function would also:
remove rows for extreme conditions that had no downstream impact (ie, were only called in a stock's ddt function and did not influence any other variable)
remove rows representing table functions (the table functions themselves, not the result of calls to them)
remove columns for variables that are exogenous, and for table functions
block off cells that are not influenced by test condition - ie, they are upstream in a feedback loop, etc.
The text was updated successfully, but these errors were encountered:
One of the problems with this method is that if you rename a variable, it won't be automatically renamed in the excel file. If you've done a lot of work filling in values for the tests, this could cause a problem.
Expanding on ideas in Lessons from software testing for developing
behavioral tests of dynamic models, create a simple way to specify extreme conditions tests for a model.
The general idea is to have a table where each row specifies a particular parameter and a value for it to take on, and then each column corresponds to the other parameters in the model. When a test condition has implications for the immediate values of one of the other parameters, these are specified in a cell. In the example from the above paper, the test matrix could look like:
To make this type of testing easier, we should also have a function which creates a template for this type of matrix, and saves it to a file that the modeler can edit and add their constraints to. The most basic template would just take extreme conditions for each value and make them into rows.
A more advance templating function would also:
ddt
function and did not influence any other variable)The text was updated successfully, but these errors were encountered: