TRIAL DESIGN WITH SIMULATED ANNEALING
A method for trial design with simulated annealing includes obtaining, via at least one processor of a trial design platform, initial trial design simulation results data for a set of trial designs. The method may further include identifying an initial design from the design simulation results data, predicting, performance data of designs for variations of parameters corresponding to the set of trial designs, and identifying a new design for simulation by varying parameters of the initial simulated design based on the predicting. The method may also include simulating the new design, determining performance data of the new design, and identifying a second new design for simulation by varying parameters of the new design based on the simulated performance of the new design.
1 . A method comprising:
obtaining, via at least one processor of a trial design platform, initial trial design simulation results data for a set of trial designs and generated, in part, via the trial design platform;
identifying, via the at least one processor, an initial design from the design simulation results data;
predicting, via the at least one processor, performance data of designs for variations of parameters corresponding to the set of trial designs;
identifying, via the at least one processor, a new design for simulation by varying parameters of the initial simulated design based on the predicting;
simulating, via the at least one processor, the new design and determining, via the at least one processor, performance data of the new design; and
identifying, via the at least one processor, a second new design for simulation by varying parameters of the new design based on the simulated performance of the new design.
2 . The method of claim 1 , wherein the initial design is a Pareto design.
3 . The method of claim 1 , wherein the initial design is a convex hull design.
4 . The method of claim 1 , wherein the predicting is based on historical data.
5 . The method of claim 1 , wherein the predicting is based on Delaunay triangulations.
6 . The method of claim 1 , wherein variations of different parameters are simulated in parallel.
7 . The method of claim 1 , further comprising:
determining, via the at least one processor, a magnitude of variations based on historical data.
8 . The method of claim 1 , wherein varying of the parameters is directed to gaps in the simulated designs.
9 . The method of claim 1 , wherein varying of the parameters is based on a parameter priority rating.
10 . The method of claim 1 , wherein the initial trial design simulation results data provides a coarse representation of design options.
11 . The method of claim 1 , further comprising:
evaluating historical trial design selections to identify one or more trial design parameters based at least in part on one or more trial design criteria determined from a user via an interactive interface, wherein obtaining the initial trial design simulations results data is based at least in part on a quick search data structure and the one or more trial design parameters;
generating a substitute for at least some of the initial trial design simulations results data based at least in part on a relationship between the initial trial design simulations results data and supplemental data;
generating a performance surface based at least in part on a set of trial designs corresponding to the initial trial design simulations results data;
evaluating one or more trial designs based at least in part on the performance surface; and
calculating a score based on normalized score component values corresponding to the initial trial design simulations results data.
12 . An apparatus comprising:
an analysis circuit structured to receive data for an initial design;
an evaluation circuit structured to vary parameters of the initial design to generate a second design; and
a simulation circuit configured for simulating the second design and determining a performance of the second design;
wherein the evaluation circuit is further structured to vary parameters of the second design to generate a third design based at least in part on the performance of the second design.
13 . The apparatus of claim 12 , wherein the initial design is a Pareto design.
14 . The apparatus of claim 12 , wherein the initial design is a convex hull design.
15 . The apparatus of claim 12 , wherein the evaluation circuit is structured to predict the performance of the second design based on historical data.
16 . The apparatus of claim 12 , wherein the evaluation circuit is structured to predict the performance of the second design based on Delaunay triangulations.
17 . The apparatus of claim 12 , wherein variations of different parameters are simulated in parallel.
18 . The apparatus of claim 12 , further comprising:
determining one or more magnitudes of variations of parameters based on historical data.
19 . The apparatus of claim 12 , wherein variations of parameters are directed to gaps in the simulated designs.
20 . A system comprising:
an electronic device having an electronic display; and
a server in electronic communication with the electronic device and having at least one processor and a memory device that stores an application that adapts the at least one processor to:
obtain initial trial design simulation results data for a set of trial designs;
identify an initial design from the design simulation results data;
predict performance for designs for variations of parameters;
identify a new design for simulation by varying parameters of the initial simulated design based on the predicting;
simulate the new design and determine the performance of the new design;
identify a second new design for simulation by varying parameters of the new design based on the simulated performance of the new design; and
transmit, to the electronic device, data structured for display on the electronic display and corresponding to the second new design.