It is aimed at interactive development of qualitative multi-attribute decision models and the evaluation of options. A multi-attribute model is a hierarchical structure that represents the decomposition of the decision problem into subproblems, which are smaller, less complex, and potentially easier to solve than the complete problem.
Availability
It is useful for supporting complex decision-making tasks, where there is a need to select a particular option from a set of possible ones to satisfy the goals of the decision maker.
Functionality
DEXi differs from most conventional multi-attribute decision modeling tools in that it uses qualitative (symbolic) attributes instead of quantitative (numeric) attributes. It also has better graphical and reporting capabilities and facilitates the use of weights to represent and evaluate qualitative utility functions.
Applications
On the other hand, DEXi is slightly less powerful than DEX in dealing with incomplete option descriptions: DEX uses probabilistic and fuzzy value distributions, while DEXi only allows the use of clear or unknown option values.
Development and history
Until 1999, DEX had been used in dozens of real-life decision situations (see publications on the DEX web page). On the other hand, more emphasis has been placed on graphical and reporting options in DEXi.
Versions
- New components
- New program features
- Changed program features
- Bug fixes
The initial development was financially supported by the Ministry of Education, Science and Sport of the Republic of Slovenia as part of the Ro program (computer literacy). Import and export option data via File menu commands and Copy/Paste operations on the evaluation page.
Credits
Acknowledgments
Decision Analysis
Decision Problem
Decision Process
Participants of the Decision Process
Stakeholders (also called decision problem owners): Individuals or organizations who have a legitimate interest in the decision problem. Experts: People who know the field so they can provide information and advice relevant to the decision.
Decision Problem Identification
They can contribute to the identification of the overall decision problem, to the definition of options, objectives and criteria, and to the development of decision model. It is important that the decision-making problem can be decomposed into smaller, less complex subproblems, and that the options can be described by their basic characteristics corresponding to the problem decomposition.
Decision Model
Primarily, it must be about options that must be assessed, analyzed and compared with each other.
Multi-Attribute Model
Qualitative Multi-Attribute Model
Attribute
Tree of Attributes
Interpretation
Linked Attributes
Recommendations
Scale
Example scales
Recommendations
Utility Function
- Intervals
- Complex Rules
- Weights
- Combinatorial Explosion
In DEXi, a utility function maps all combinations of lower-level attribute values to a value of Y. Consider a utility function f that maps the values of attributes X1, X2,.., Xn to the value of an aggregate attribute Y.
Options
Evaluation of Options
Analysis
In all cases, you should regroup the lower-level attributes and introduce one or two new aggregate attributes, marked with . How sensitive the evaluation is to small changes to the model (such as adding or deleting an attribute, modifying some decision rules). In other words, analysis is a creative and perhaps iterative application of decision models aimed at better understanding the decision problem, better understanding of options, their characteristics and consequences, and better justification of the decision. .
On the Options page, you can duplicate an option description and prepare it for 'what-if' analysis. To explain and justify the results, you can focus on particularly bad or good evaluations. On the same page, you can change individual option values and immediately see the effect on evaluation results.
On the Model page, you can change any model component and then retry the evaluation by opening the Evaluation page. Various charts and reports can provide further insight into the evaluation process and achieved results.
DEXi Model
DEXi File
Main Toolbar
File Menu
Option Data File
Option data files contain option data, which is imported and exported using commands on the File menu. On Charts Page/Options Sub page, where you can select options that should actually be exported (all options are selected by default). In the 'Import Options' and 'Export Options' dialog boxes used to specify file names and the base file format, which is either 'Tab Delimited' or 'Comma-separated (CSV)'.
Tabbed limited options data file containing both options from the Car Evaluation model and exported using the default settings: using 'base 1' values, displaying all attributes using indentation, normal orientation. Same data as above, but using comma-separated format, 'base 0' values, and including only non-indented base attributes in normal orientation.
Function Data File
Edit Menu
Window Menu
Tile horizontally: Resize and move all currently open windows so that they are displayed on top of each other, occupying the full width of the main window. Tile vertically: resizes and moves all currently open windows so that they are shown one behind the other, occupying the full height of the main window. Minimize All: Minimize all currently open windows so that they are shown only as small dashes ('icons') in the main window.
Help Menu
Model Window
Model Page
- Workspace
- Commands
- Remarks
- Tree View
Paste: Inserts previously cut or copied subtree into the model, positioning it as a descendant ('child') of the currently selected feature. Duplicate: Makes a copy of the currently selected sub-tree and adds it as a new top-level tree in the model, so it can be easily moved around. Move Up: Moves the currently selected attribute up one place as shown in the tree view.
Move Down: Moves the currently selected attribute down one place according to what is displayed in the tree view. Scaling: Calls the scaling editor to create or edit the scale for the currently selected attribute. Helper function: Calls the function editor to create or edit the helper function for the currently selected aggregate attribute.
Tree view is an important part of the DEXi model window and its model page. Aggregate property whose utility function cannot be constructed due to undefined scales of its own and/or its descendants. Sometimes and with good reason, it is acceptable to leave a partially defined utility function and/or to have a linked property.
Scale Editor
Workspace
Input field for scale order at the top: here you can define scale order: unordered, ascending (recommended) and descending.
Commands
Remarks
Function Editor
- Table
- Toolbar
- Pop-up Menu
- Status bar
- Utility Function Status
- Weight Editor
- Handling Non-Entered Function Values
- Function Editing
Use scaling order and Use weights are two checkboxes that determine how to handle non-entered values. Note that this is generally achievable with less than 100% entered rules due to DEXi's treatment of non-entered values. In the Function Editor, utility function values are either entered by the user or not entered.
Non-entered values are shown in normal fonts and, by default, they are recalculated when the table changes. It is easy to see that monotonicity generally narrows the intervals of values that can be assigned to non-input cells. Likewise, the entered bad value of line 7 implies that the value of non-entered lines 4 and 1 is also bad.
This strategy calculates the values of unentered rules using a hyperplane (linear function), which is constructed using weights, as defined in the Weight Editor, and other rules already entered. Using these weights and the already defined rules 3, 5, 7 and 9, DEXi constructed a hyperplane and used it to determine the values of unentered rules and 8. But even in this case you should be aware that non-entered rules are more volatile than entered ones.
Options Page
Workspace
As long as your function is '100% certain', the proportion of actually entered rules is not that important and may be less than 100%. Entered values are never changed by DEXi, but non-entered values can be inadvertently changed later, for example by changing weights in the Weight Editor. To protect your finished function from such changes, you can use the Enter values command.
For a final check on your function, you may also want to review its complex rules and weights.
Commands
All commands refer to the option (column) whose cell is currently selected in the table. Copy: Copies the currently selected setting (column of cells) to the clipboard for further use. Insert: Inserts the previously cut or copied column into the table and places it to the right of the currently selected column.
Remarks
Evaluation Page
Workspace
When you select a cell that corresponds to a basic attribute, a data entry field appears in the toolbar, allowing you to change the value of that cell in the same way as on the Options page.
Commands
Charts Page
Workspace
To choose which option is displayed, select it ('check' its name) on the Options subpage.
Commands
Report
Report elements
For each scale in your model, the scale name, associated attribute description, and all scale values are printed along with their own values. Note that only options selected on the Charts page are included in this report element.
Remark
Preview
The first saves the entire report, the other two only the current page.
Settings
Report Page
Import/Export Page
Attribute Values: Specifies the format of attribute values, which can be text strings or ordinal numbers starting at 0 (base 0) or 1 (base 1).
Advanced Page
This is a very small and simple model used to illustrate the key concepts of multi-attribute modeling and DEXi, and is not intended to address the problem of auto-evaluation at a realistic level.
Tree of Attributes for Car Evaluation
Interpretation
Attribute Types
Attribute Descriptions
This model has traditionally been distributed together with all previous versions of the DEX and DEXi programs. In DEXi reports, this tree is printed along with attribute descriptions and appears as follows:
Scales for Car Evaluation
Utility Functions for Car Evaluation
Interpretation
Each row returns a value of CAR for one combination of the values of PRICE and TECH.CHAR.
Elementary decision rules
Complex rules and weights
Description and Evaluation of Cars
Interpretation
At the bottom, the table shows two options, Car1 and Car2, described by the qualitative values assigned to the six basic characteristics of the tree. These values are then aggregated from the bottom to the top of the tree of properties according to the structure of the tree and defined utility functions. In this way, intermediate evaluation results are first obtained and assigned to the attributes PRICE, COMFORT and TECH.CHAR.
Some Car Evaluation Charts
Bar Chart
Scatter Chart
Radar Chart
Documentation in Slovene
Selected Publications
Bohanec, M., Messéan, A., Angevin, F., Žnidaršič, M.: SMAC advisor: a decision support tool for the coexistence of genetically modified and conventional maize.