FILTERING DESIGNS USING BOUNDARIES DERIVED FROM OPTIMAL DESIGNS
A method, according to some implementations, receiving, for each trial design of a set of trial designs, a simulated performance and presenting, for at least two criteria, a plot of the simulated performance. The method may further include presenting, for at least one additional criteria, a plurality of contour lines on the plot and selecting a subset of the set of trial designs based on the plurality of contour lines.
1 . A method for trial design analysis, the method comprising:
receiving, for each trial design of a set of trial designs, simulated performance;
presenting, for at least two criteria, a plot of the simulated performance;
presenting, for at least one additional criteria, a plurality of contour lines on the plot; and
selecting a subset of the set of trial designs based on the plurality of contour lines.
2 . The method of claim 1 , wherein the plurality of contour lines connect Pareto optimal designs.
3 . The method of claim 1 , wherein the plurality of contour lines connect convex hull designs.
4 . The method of claim 1 , wherein the plurality of contour lines are step contours.
5 . The method of claim 1 , wherein selecting the subset comprises selecting a set of designs between two contour lines of the plurality of contour lines.
6 . The method of claim 1 , wherein selecting the subset comprises selecting a set of designs bounded by at least two contour lines.
7 . The method of claim 1 , further comprising: storing the simulated performance according to a grouping defined by the plurality of contour lines.
8 . The method of claim 1 , wherein the plurality of contour lines connect Pareto optimal designs of equal values for the least one additional criteria.
9 . The method of claim 1 , wherein the plurality of contour lines connect convex hull designs of equal values for the least one additional criteria.
10 . The method of claim 1 , wherein a density of the plurality of contour lines is configured to partition a design space into groups of less than a threshold of designs.
11 . An apparatus comprising:
a boundary defining circuit configured to:
receive, for each trial design of a set of trial designs, simulated performance;
present, for at least two criteria, a plot of the simulated performance;
present, for at least one additional criteria, a plurality of contour lines on the plot; and
select a subset of the set of trial designs based on the plurality of contour lines.
12 . The apparatus of claim 11 , wherein the plurality of contour lines connect Pareto optimal designs.
13 . The apparatus of claim 11 , wherein the plurality of contour lines connect convex hull designs.
14 . The apparatus of claim 11 , wherein the plurality of contour lines are step contours.
15 . The apparatus of claim 11 , wherein selecting the subset comprises selecting a set of designs between two contour lines of the plurality of contour lines.
16 . The apparatus of claim 11 , wherein selecting the subset comprises selecting a set of designs bounded by at least two contour lines.
17 . The apparatus of claim 11 , wherein the boundary defining circuit is further configured to store the simulated performance according to a grouping defined by the plurality of contour lines.
18 . The apparatus of claim 11 , wherein the plurality of contour lines connect Pareto optimal designs of equal values for the least one additional criteria.
19 . The apparatus of claim 11 , wherein the plurality of contour lines connect convex hull designs of equal values for the least one additional criteria.
20 . The apparatus of claim 11 , wherein a density of the plurality of contour lines is configured to partition a design space into groups of less than a threshold of designs.