New Functionality and Changes Introduced since Decision Optimizer 8.2
This section describes the new features and changes introduced with Decision Optimizer 8.2.
This release builds on the recent Decision Optimizer 8 releases as we drive the next evolution of Decision Optimizer, focused on ease of use, automation, and innovation.
Added in Decision Optimizer 8.2.2
This release is a maintenance release, resolving a number of user interface and decision tree defects.
Defect Remediation
- Some consistency constraints were being ignored.
- Performance issue regarding decision tables with string columns.
- Validation scenario with over sample weighted data could crash.
- Subscenarios with identical names lead to missing data in the tableau workbooks.
Added in Decision Optimizer 8.2.1
Defect Remediation
- User Interface elements:
- Some tables not providing scroll bars to reach additional information.
- Scenario Settings swap set baseline scenario table only available for first scenario.
- Incorrect Project Status toolbar color / status.
- Decision Trees with string variables, UTF-8 characters and special conditions
- Improved guidance on the use of ranged constraints in Exploratory Optimization Scenarios, including examples showing how different inputs impact outputs.
- Added clarification on the potential impact of using missing values in averaged global constraints, together with potential solutions.
- Improved guidance and examples on the definition of Actions and Treatments.
Added in Decision Optimizer 8.2
Swap Set Reports Enhancements
- Optimization Scenarios
Users can now define multiple baseline scenarios in the Scenario Settings view. When you view the results in the Swap Set reports, you now have the option of which Scenarios, Sub-Scenarios, and Baseline scenarios to view
- Exploratory Optimization Scenarios (Constraints and Stress Testing)
Users can now define a baseline scenario in the Scenario Settings view. When run this will create a swap set report for each Sub-Scenario.
- Improvements to the way Base Decision Trees are imported into Tree Templates.
On import these will now reuse existing binnings already in the binning library.
- Improvements to the application of Tree Constraints Consistency Criteria
Inconsistent constraints can now be automatically relaxed by the model when it detects they lead to infeasible problems, while previously if infeasible they would be disabled.
- The Global Tree Constraint Maximum Number of Tree Leaf Nodes can now be used as a soft constraint (previously this was optional criterion used to specify a goal measured on the scenario output.)
The value set must be less than the number of nodes created by the Granularity criteria. This limit will be enforced during optimization. If the combined effect of this limit and the limits from other constraints result in an infeasible problem, then the tree leaf node limit will be automatically disabled.
- Tree Aware Optimization scenario runs now provide an additional output report, detailing all the consistency constraint violations in the output decision tree.
- Decision Impact Model Execution
Improvements in the decomposition of execution steps and the generation of the math model have reduced peak memory usage by 20% to 30% on benchmark projects, whilst improvements of over 80% have been seen on some more complex scenarios. In turn this has also helped reduced scenario run times.
- Tree Aware Optimization Scenarios
Improvements in the way that data is created on each TAO scenario run, has helped reduce scenario run times by 15% to 20% in benchmark projects.
- Optimization Profiles
A new profile has been added that supports Linear Programming (LP) Relaxation. This profile will work with optimization and exploratory optimization scenarios (it does not provide good quality solutions for combinatorial problems like tree aware optimization). LP Relaxation is a standard technique to find approximate solutions to complex optimization problems faster than with the standard algorithm. The advantage in using this option is that it will help quickly find a solution with an objective function value that is not worse than the optimal solution of the problem. The drawback is the returned solutions can sometimes violate the global constraints. It is recommended for use on large, complex problems or where you need to run a lot of optimization scenarios quickly, for example during exploratory optimization. The Summary Report Solution Status section will provide information on the resulting quality of the optimization solution.
- Binning Labels should now be imported as part of DO6 project binning libraries
- The default DO6 Do Nothing treatment will be maintained if it is being used as a Treatment
User Interface Enhancements
Project Status Bar—improvements have been made to the underlying logic determining the project status and how that is displayed through the colored status bar.
Defect Remediation
- User Interface—missing custom buttons
- Project Status Bar
- DIM Graph resizing
- Base Decision Tree import
- Swap Set Reports
Documentation
- Decision Trees—creation and use of tree data files on Project and Scenario attachments
- Segmentation, with guidance on ordering segments.
- Efficient Frontier Reports
- DIM Graph Component Tracing
© 2001-2024 Fair Isaac Corporation. All rights reserved. This documentation is the property of Fair Isaac Corporation (“FICO”). Receipt or possession of this documentation does not convey rights to disclose, reproduce, make derivative works, use, or allow others to use it except solely for internal evaluation purposes to determine whether to purchase a license to the software described in this documentation, or as otherwise set forth in a written software license agreement between you and FICO (or a FICO affiliate). Use of this documentation and the software described in it must conform strictly to the foregoing permitted uses, and no other use is permitted.