Task 2: Making decisions for a population of cases

You learn about using Decision Designer to implement decisions for multiple cases.

When customers rank sensor formats, their top choice might not always be practical. A full-frame camera might rank highest but prove infeasible due to weight constraints, budget limitations, or skill level. Each customer must identify feasible alternatives and select the best option from this reduced set.

While individual customers perform this feasibility check once, consider an entire customer population. Assume all customers share the same sensor format ranking but differ in budget level (low, medium, or high), photography type (landscape, portrait, sports, or night photography) and experience level (beginner, advanced, professional). These factors determine which alternatives remain feasible for each customer profile.

The challenge here is to recommend the optimal sensor format based on each customer's specific constraints. When the top-ranked option is infeasible, the next best alternative should be suggested.

Exercise 1: Evaluating camera sensor formats by using a decision spreadsheet

You build a spreadsheet that helps evaluate which camera sensor format is most appropriate under different user scenarios. The goal is to systematically apply constraints and make a best recommendation for each case.

  1. Create a spreadsheet with the following three categorical characteristics, each represented as its own column:

    • Budget: Low, Medium, High
    • Subject: Landscape, Portrait, Sports
    • User level: Beginner, Professional
  2. Add 18 rows representing all possible combinations of the three characteristics. Each row should fully describe a specific photographic scenario. For example: medium budget, landscape photography, beginner.
  3. Add one column for each camera sensor format: Micro Four Thirds, APS-C, and full-frame.
  4. In each of these columns, indicate whether that sensor format is feasible for the case that is described in the row.

    Apply the following constraints:

    • Low budget eliminates APS-C and Full-frame.
    • Medium budget eliminates Full-frame.
    • Sports photography eliminates Micro Four Thirds.
    • Landscape photography eliminates Full-frame.
    • Beginner level eliminates Full-frame.

    A sensor format is feasible for a given case only if any applicable constraint does not eliminate it.

  5. Add a final column labeled Sensor format. For each row, identify the highest-ranked feasible option. If no options remain feasible, indicate: No recommendation.

The following row represents a beginner photographer with a medium budget who is interested in landscape photography:

Level Subject Budget Full-Frame APS-C Micro-Four Thirds Decision
beginner landscape medium infeasible feasible feasible APS-C

The beginner with a medium budget who is interested in landscape photography has two feasible options (APS-C and Micro Four Thirds). Since Full-Frame is eliminated, the decision defaults to the second-ranked choice: APS-C.

Exercise 2: Implementing in Decision Designer

Note: The solution for this exercise is available through the New decision service wizard under Practice tutorials. For more information about importing samples, see Building decision services.

You define a decision table that directly maps each case to a decision.

  1. Sign in to Decision Intelligence Client Managed Software using your instance credentials.
  2. Create a decision automation.
  3. Click New decision +.
  4. Click Create + in the Data tab.
  5. Create a data model and define an enumeration type:
    1. Open the Data tab and click Create +.
    2. Enter a name for the data model and click Create.
    3. Click the Add button Add button next to Data model, and select Enumeration type to create the four following enumeration types: budget, subject, level, and sensor format.
  6. Go back to the decision service to create a decision model:
    1. Click Create + in the Models tab and select Decision model.
    2. Add three input data nodes to the decision model:
      • budget of output type budget
      • subject of output type subject
      • level of output type level
    3. Add a decision node sensor format and associate it with the output type sensor format.
    4. Make sure that the sensor format decision node depends on the three input data nodes you defined, as shown in the following screenshot:
      A decision diagram with one decision node and three input data nodes
  7. Add a decision table to the sensor format decision node:
    1. Click the decision node and open the Logic tab.
    2. Click Create rule artifact and select Decision table.
    3. Complete the decision table based on the spreadsheet that you created in Exercise 1.
The following row represents a beginner photographer with a medium budget who is interested in landscape photography:
A decision table with one row and four columns

Lessons learned

Unlike the spreadsheet, the decision table does not include columns showing the feasible sensor formats. That information is necessary to determine the best decision for a specific case. However, it is not necessary to define a decision service that returns the best decision when a case is provided as input.

The ranking that is established in Task 1 is also not shown in the decision table. Even so, you can still identify the best decision for each case and verify that the decision logic is consistent with that ranking.

Decision tables should not be confused with consequence tables, which are used to evaluate alternatives against different criteria (see Task 1). As illustrated, a decision table provides a shortcut for identifying the best feasible alternative for a specific case. Therefore, it captures information that constrains the set of alternatives, whereas a consequence table describes the consequences of those alternatives.

Limitations

This approach is appropriate when the number of cases is relatively small. Because each case requires its own row, completing the table can be labor-intensive. The next step explores how this approach can be improved.