The Scenario Manager allows editing of a small portion of a RiverWare model without affecting the integrity of a baseline model. With the Scenario Manager, the values of a subset of slots can be altered independently of the baseline model. Models can be viewed and run using these various scenario values without corrupting the values in the baseline model. The Scenario Manager provides a convenient interface to load, edit, and save values for these slots, run the model using these different slot values, and compare the baseline values with the scenario values.

This documentation describes the basic features of the Scenario Manager and how to construct scenarios.

The following definitions are used in this documentation.

A baseline model must be created before any scenarios can be developed. The baseline model is generated by saving an existing model, with or without a ruleset, as a baseline model. Once a model has been saved as a baseline model, model topology, methods, rulesets, and non-scenario slots become protected from editing. The baseline model cannot be re-saved once it is created unless it is resaved as a non-baseline model. To maintain the integrity of the baseline model, we do not recommend re-saving the model as non-baseline. Following are the steps to create a baseline model from a regular model; these steps are described in detail in the remainder of this chapter:

To create a baseline model, open the desired model (and ruleset). Open the Scenario Manager from the Utilities, then Scenario Management menu on the main RiverWare menu.

This will open the Scenario Manager dialog. At this point, the dialog is empty because no slots have been specified. The title bar of the Scenario Manager says “[Not baseline model]” because no baseline scenario slots have been specified or saved.

You will need to decide which slots are key to generating different scenarios. This subset of slots will be the only slots that scenario users are allowed to modify values or data to test various scenarios against the baseline scenario. To add the slots that form the baseline scenario select Configuration, then Add Input Slots.

Use the selector to choose the Scenario Input Slots. The added slots appear under the Slots to Modify (Input Slots).

As creator of the baseline, you can specify a minimum or maximum (or both) value that the slot is not allowed to violate in any scenario.

To specify a minimum or maximum slot value, highlight the slot and select Configuration, then Set Min Value or Configuration, then Set Max Value from the menu bar. A dialog window will appear. Enter a value and select the OK button. You can also use the Configuration, then Compute Min/Max Values from Baseline Slots to give you the existing bounds. Once this value is set, it cannot be altered by scenario users.

Comments can also be added to the individual slots in the comment field to the right of the slot name and to the scenario in the scenario comment field. To enter a slot comment, highlight the desired slot and single select the Comment field next to the slot name. A text cursor will appear. Enter the desired comment.

The list of slots can be ordered in several ways. The current ordering of slots is saved with the model.

Next, specify which output slots you would like to compare against the baseline and other scenarios. When a scenario run is made, a snapshot will be taken of each of these slots. Define the list of slots in the Results to Compare (Output Slots) slot list. This list is saved in the baseline model and is not modifiable by the stakeholder.

To enter the baseline scenario comment, single select the Baseline Comment field and enter the desired text.

A Scenario Sandbox provides a convenient way for the scenario users (i.e. the stakeholders) to easily create and run scenarios and view results.

The configuration of the sandbox is defined by the baseline creator (i.e. scenario engineer) and the scenario user cannot change the configuration. The current implementation of the Scenario Sandbox supports the following functionality:

Figure 8.1 shows a sample completed sandbox.

Use the Create Sandbox button in the Scenario Manager (available only in non-baseline models) to create a sandbox.

Configure the Scenario Sandbox by selecting up to three scenario input slots to be adjusted by the stakeholder in the sandbox. The configuration for each of these three slots consists of the following:

After configuring the Scenario Manager and the Scenario Sandbox, it is critical that you run the model. This generates the BASELINE output and a new snapshot object is created. These are the data against which all scenario results are compared.

When you are finished with the configuration, save the entire RiverWare model as a baseline model. This will preserve the baseline scenario slots in the Scenario Manager and ensure that other users are not able to modify the model further, unless they specifically resave the baseline model as non-baseline.

To help document the baseline model, you will be prompted with the Model File Info dialog. A comment about the date, time, and username saving the model as a baseline is added to the end of the comment by default. This comment field can be freely edited by the user before saving the model as a baseline.

Once the model has been saved as a baseline model, “[Scenario: BASELINE]” will appear next to the model name in the main RiverWare workspace. The Scenario Manager will also indicate that the model has been saved as a baseline model.

You can now distribute the model to stakeholder scenario users.

After the creation of a baseline model, no slots can be added or deleted from the scenario slot list. This is done to preserve the integrity of the baseline model.

You have the following options:

The Scenario Sandbox is accessible from the RiverWare Workspace and from the Scenario Manager. If a Scenario Sandbox has been configured,

The Scenario Sandbox presents all the functionality necessary to create and run a scenario within a baseline model.

Next adjust values using the sliders for up to three scenario input slots. Changes to the sliders effect those input slots of the currently selected scenario.

Each input slot panel contains the following controls:

The Scale sliders are exponential. The range is based on the upper value set by the scenario engineer. The lower limit of the range is the reciprocal of the specified upper value. The mid-point is always “1”, which results in no difference from the baseline slot values.

The Offset sliders are linear, and range from the negative value to the positive value of the adjustment limit value set by the scenario engineer. The midpoint is always zero, which results in no difference from the baseline slot values. The Offset values are in the slot display units.

The computed scenario slot values are clipped at the configured minimum and maximum values for the baseline slot, if those have been configured.

The bottom section of the Scenario Sandbox shows generated scenario results (outputs). Output controls consist of three components:

The plot is dynamically updated as the selections in the two lists are changed. A curve is shown for each of the selected scenarios' value of the single selected output scenario slot. A curve for the baseline slot is always shown.

Only scenarios for which a scenario snapshot object has been generated are shown with a “Generated” timestamp. Double-clicking an item on the Generated Scenarios list shows the Open Object Dialog for that snapshot. Figure 8.2 illustrates.

Once a new scenario has been created, it must be saved. Use the File, then Save Scenario As and save the scenario in the desired directory. This can be loaded into subsequent BASELINE model runs for additional analysis.

For any particular scenario, full editing of Scenario input slots can be made available by unchecking the checkbox to the left of the Sandbox button on the Scenario Manager. When re-enabling the Sandbox for a scenario, if doing so would effect the values of those input slots, this warning message popup is displayed, providing the option to keep the Sandbox disabled (and preserve the scenario’s input slot values).

When a sandbox is not configured or you would like to change more than 3 input slots, you can create a scenario and have more control over the inputs. New scenarios to test permutations of the model can only use the slots specified in the scenario slot list, and all input values must fall within the minimum or maximum constraints (if any) set on the slot.

A tab with the name of the new scenario will be added next to the baseline tab in the Scenario Manager. Selecting the tabs allows you to toggle between the baseline scenario and other scenarios.

The name of the scenario corresponds to the scenario file’s name. A newly created scenario will have the default name “Unsaved Scenario.”

Finally, a convenient way to compare the differences between a slot in various scenarios is to plot the slot against either the baseline slot values or against the values of the slot in other scenarios. You can also plot the slot for a single scenario only. All the plotting options are located in the menu bar of the Scenario Manager under Slots, then Plot Slot, Slots, then Plot Slot + Baseline, or Slots, then Plot Slot on All Scenarios. The slots are plotted in the familiar RiverWare Plot dialog with all features of the Plot dialog active. A sample graphic is provided below, which compares the selected scenario slot with the baseline slot values. The scenario slot name is distinguished from the baseline slot name by the convention: Slot Name_Scenario Name_.

Running a scenario is fairly straight forward. The scenario with which you wish to run the model must be active. Open scenarios by selecting File, then Open Scenario.

There are two options for running an active scenario:

The Scenario Manager allows for comparisons between slots in the baseline scenario run and various scenario runs or between slots in the various scenario runs. Comparisons are managed through the Snapshot Manager and the Output Manager, which can be accessed through the Scenario Manager.

A snapshot object, or Scenario Output Snapshot Object, is automatically created after each scenario run (and also for the baseline run) to capture the values of a prescribed list of slots which generally represent outputs from the run. These new snapshot objects include copies of slots prescribed by the Scenario Engineer.

The association between a scenario and its output snapshot object is based only on the scenario and the object having the same name. This snapshot object will be saved with the scenario when the scenario is saved as a scenario file.

The background color of the snapshot object icon distinguishes the two types of snapshot objects:

If the Results to Compare (Output Slots) slot list described above is not empty, then, each time a run finishes or aborts, a Scenario Output Snapshot Object having the same name as the active scenario (or else, “BASELINE” in a non-scenario run) is created (or re-created) with the following data:

The automatically generated text lines within the snapshot object’s description begin with a special character (“>”) to distinguish them from user-entered comments. All such lines are deleted from the description before generating new comment lines.

The Open Object Dialog for the scenario’s output snapshot object can be shown by selecting the Output Slots button on the scenario’s tab in the Scenario Manager dialog or in the Scenario Sandbox; see “Scenario Sandbox” for details. These buttons are disabled if that snapshot object has not yet been generated.

Also, all existing Scenario Output Snapshot Objects are listed in the Snapshot Manager Dialog, with the snapshot items having a green-background snapshot icon.

Scenarios are saved in their own files, apart from the baseline model. The scenario’s output snapshot object is also saved in that file. When loading a scenario file, the snapshot object is loaded. This will replace any existing snapshot object on the workspace having the same name.

To save the current contents of the model after running a scenario, the model must be saved as a non-baseline model. The resulting non-baseline model will only contain the current input/output values in the model; thus the baseline data will not be saved unless the baseline is the currently active scenario.