Is this Chapter will find information how to make sure if your parameter works properly. Tester is a tool with which you may check if your function or parameter give you result as you expected. When your parameter or function has thousands of arguments and you are looking for few wrong values then Sweep Analysis may help you. Basically it creates a chart on which one may see all the inconsistencies in distribution.

Tester #

To check if parameter works, go to main toolbar Simulations and Tester. Then screen will show.

Tester screen is divided on two sections Tested elements and Test input data and above that operations useful in adding elements process. In 'Available elements' type or choose from the list which function/parameter would like to test. Then click ‘Add to test’ button. If you want to add many elements at once click on the list icon [Add from domain].

Check the checkboxes of the elements you wish to add to test and click on 'Choose' button in right down corner. Then all your choices will be added to tester and will appear on 'Tested elements' section. When functions and parameters that need to be tested are already on the Tested elements list then may click run and see how your parameter works. Not always it will work the first time. Usually there are few additional information to be filled for tester to simulate properly, those information appear in section 'Test input data'. For example let's test parameter: and click RUN test.

In section 'Test input data' tester shows us what is needed to simulate how parameter works. This parameter needs to know 'driver gender' and 'driver age' then it may run again. When IN values are filled (example Gender: F, Age: 31) then OUT Value should occur.

Now in section 'Tested elements', column 'Result value' will see parameter level OUT value. Results may be compared with parameter matrix data Parameterization ->Parameters and go to our example parameter detail:

Tester gave us value 0,582 as it is shown in the matrix (on OUT level).

Test results can go through verification according to set verification rules, for this field 'Result verification' must be set as 'Yes', all verification rules possible to configure are:

  • No - choose if no comparison is needed
  • Yes – 'without tolerance' if result value must be exactly the same as expected value
  • Yes – 'max. absolute error' if result value may be different from expected value by inputted factor
  • Yes – 'max. percent error' if result value may be different from expected value by inputted percent of expected value.

For the comparison to be possible field 'expected result' must also be filled.

Any test case, with provided test data and result verification settings, can be exported to excel file. For export purpose column delimeter might be changed to ',' from default : '|', but we do not recomment changing the delimeter to ',' if any of the tested emements returns list of values or any value with ',' withinit.

Batch tester #

Batch tester allows you to test more than one element at the time. You may create a list of parameters/functions and run test simultaneously to check all of them. To run batch test go to Simulations then Batch tester.

In Batch tester toolbar will find operations like:

  • Import/Export XLS file
  • Remove all / selected – deleting elements from test
  • Add test case – Add selected function to the mass test
  • Save package – save full test settings and results
  • Open package – open previously saved package

First are basic information about test subject like:

  • Package name
  • Result verification -if test result should be compared with some expected value and with what tolerance.
  • No - choose if no comparison is needed
  • Yes – 'without tolerance' if result value must be exactly the same as expected value
  • Yes – 'max. absolute error' if result value may be different from expected value by inputted factor
  • Yes – 'max. percent error' if result value may be different from expected value by inputted percent of expected value.

What’s important only test cases may be added to mass tester, what means that if we want to test some function in Batch Tester first we need to test it in Tester and save project. For example, if I want to add function to Batch tester first I need to go to Tester and add to test that function by typing its name in field 'Available elements' or choosing from the list, then clicking 'Add to test' button. When the function is in the test user should check if everything works properly. Then fill Test Name field (top right corner) – ex. BiCalc. Now one may 'Save' that test.

For Batch tester it is really important to input expected value because in most cases in Batch tester user needs to make a comparison between value expected and result to tell if function/parameter is working correctly. Then go to Simulations ->Batch tester then in batch tester toolbar choose Add test case. From tests list choose your test:

Then click 'Accept' button to accept your choice. When accepted 'Choose test case' window will disappear and your test case should show on 'Test case list' section. This way may add as many test cases as you think is needed to Batch tester. When all are in 'Test cases list' choose 'Run' button to start testing. When tester is done checking elements gives you information like:

  • Status - green for success (no errors detected) or red for failure (error occur) may be yellow if there is no expected value and comparison was expected
  • Time [ms] – time spend on testing the element
  • Result – Success or failure

When list of 'Test cases' is long may use function 'Show failures only' (checkbox on the right side) to show only those elements which failed the test.

May also expand your test. Lets’ assume we want to test all Male in age from 30 to 45. The easiest way to do so, is to export our element as it is and then make some changes in XLSx file and then import it back to Hyperon.

In exported table have to input ‘age range’ and ‘driver gender’. Example changes were limited to blue column were are typed random numbers as expected results. Of course those expected results will product errors in test and most of our test subjects will come as failures.

Colored column is ‘expected value’, since those are random vales our expectation is to have inconsistency in ‘value expected’ and ‘result value’. All that’s left is to import file to Hyperon and click ‘Run’ button. Status of most of elements will be failure but that is just the effect of random numbers as expected values in XLSx file. Then effect of second Run export again to XLSx file, in table will see that column ‘Evaluated…’ is filled, those are result values for our function/parameter. Copy those to blue column and save that values for future tests. If there is new parameter or function created it is recommended to save its values for future comparisons. Erase value cells in 8th column, save XLSx file and Import it to Hyperon again. Run test. Now all status should be green since all result values are the same as expected values.

Sweep analysis  #

As mentioned Sweep analysis is the best tool to find mistakes in parameters or functions definitions.

Sweep analysis screen divides into few parts:

  • Data visualization where chart will show
  • Data series – where functions or parameters that are about to be visualized will be listed
  • Parameters Check window – where needed information has to be added to run simulation
  • Horizontal axis (categories) where may change axis view

There are tree available operations on the simulation: ‘Save’, ‘Open’ and ‘Chart type’ – to change linear chart into bar graph. For example if parameters value depends on persons’ gender and age like demo.motor.coverage.pd.premium and there is a need to make sure that all values meet the standards then we use Sweep analysis. In Data series section type name of your function/parameter or find it on the list. If you want to add more than one function click on button 'Add', new line will appear and another function/parameter may be input to analysis.

Then go to next section, on the right: 'Parameter' – click Check button to find out if your function need any information to run simulation. There may be a need to declare age of client but that does not mean you can’t check selected function by age range –declaration is obligatory for simulation to start. In example needed are Clients age and Age when driver license was obtained. Lets’ assume that our client is 30 years old women.

When asked parameter were given, click Check button again to make sure that’s all what is needed to run simulation. If no new parameter was asked go to next section: Horizontal axis.

  • Parameter - Here may choose what type of parameter sweep will be made through -that parameter will be your functions’ argument.
  • Samples type :
  • List
  • Range

In example 'Age' range was 'from' 18 'to' 55 and 'step' should be 1 unit cause variable 'Age' may take only integer values. Then after clicking Show button, application Hyperon will show:

That is correct characterization of that parameter. Now if any change was made in source matrix characteristic will show it. For example:

Scale will adjust automatically. Now in a moment you can localize wrong value in parameter and correct it as needed. What’s more we may compare values for female and male of that parameter. Just Click Add button in Data series section, then in new line input the same parameter:

Click 'pencil' icon on the right side of selected parameter. Then enter 'Series Name', then 'Add' (overwritten parameters for series). In column 'Parameter' choose gender to overwrite (since we want to show on chart female and male characteristic). Then 'Accept'. 'Series name' Shows in legend above chart.

First parameter is name Female (to name it click pencil icon and input name). Then click button Show. Now you will have characteristic for both genders in age range from 18 to 55.

Overwriting parameter is needed when default information are not suitable for all parameters added to analyze or when you need to make a comparison of the same parameter for two different values (for example when you want to see how parameters values distribution looks like for Male and Female) . User will be asked about every value necessary for simulation in Parameter section– think about it as a test case which has no influence for whole simulation. Everythink about distribution is declared in section Horizontal axis where user has to declare by which Parameter value sweeping will be made by.