|
|
|
|
|
|
|
Figure 1. Parallel axis plot of candidates for a PhD position. Only one solution (alternative or data point) out of nine is displayed. This candidate Lim has 80 marks for science, 50 for science and so on. One sequence of lines in a parallel axis plot represents one data point (solution). You can have mixture of numeric and symbolic attributes in RRPExplorer. This is the best compromise solution obtained through RRPARETO3 algorithm.
Figure 2. Parallel axis plot of candidates for a PhD position. All the solutions are shown. The data for candidate Lim is currently selected.
Figure 3. Parallel axis plot of candidates for a PhD position. A window has been selected containing only the top half of the axis, Science. All other candidates have been removed upon selection of this window.
Figure 4. Parallel axis plot of candidates for a PhD position. The solutions that are dominated after applying traditional Pareto filtering. (Use the command “Selection/Invert” after applying “Selection/Pareto filter”)
Figure 5. Examining why Jim is removed by Pareto filtering. Tabular view of raw data. Click on the solution Jim. Select “Selection/Dominating points”. Then press Shift and click on the points you want to select. After that click on “Selection/Compare”. Lam is better than Jim under all the criteria except publications and prizes for which they are equal.
Figure 6. Screen for defining the Problem (Click on Setup menu). You can change the order of objectives and specify other parameters required for filtering.
Figure 7. About RRPExplorer.
Figure 8. Parallel axis plot of the solutions for the design of a façade. Out of 6079 solutions, 3962 that satisfied the constraints are displayed.
Figure 9. Setting up constraints. Linear constraints which are unary or binary can be input. The general form is y >= m *x + c. Here, the constraint says that the variable W1_H should be greater than or equal to +1.5.
Figure 10. Best compromise solution displayed using a VRML viewer.
Figure 11. Specifying your own viewer for displaying the solution point when you double click.