US 6,983,233 B1 | ||
Combinatorial parameter space experiment design | ||
Marco Falcioni, San Francisco, Calif. (US); C. Eric Ramberg, Sunnyvale, Calif. (US); and Youqi Wang, Atherton, Calif. (US) | ||
Assigned to Symyx Technologies, Inc., Santa Clara, Calif. (US) | ||
Filed on Apr. 19, 2001, as Appl. No. 9/840,003. | ||
Claims priority of provisional application 60/198208, filed on Apr. 19, 2000. | ||
Int. Cl. G06F 17/50 (2006.01); G06G 7/48 (2006.01) |
U.S. Cl. 703—12 | 74 Claims |
1. A computer program product on a computer-readable medium for designing a set of experiments, the program comprising instructions
operable to cause a programmable processor to:
receive input defining a set of experimental parameters, one or more sampling patterns defining a sampling for each parameter
of the set of experimental parameters, and one or more constraints limiting the set of experiments to a particular volume
or volumes of a hyperspace defined by the set of experimental parameters, at least a plurality of the set of parameters being
grouped according to a parameter type such that the grouped parameters are constrained to perform a common role in the set
of experiments;
generate a first estimate of the practicability of a first set of experiments defined by the parameters, the sampling patterns
and the constraints, the first estimate including a count of the set of experiments defined by the set of experimental parameters,
the sampling patterns and the constraints, and provide the first estimate to a user;
receive input modifying at least one of the set of parameters, the sampling patterns or the constraints in response to the
first estimate;
in response to the input modifying at least one of the set of parameters, the sampling patterns or the constraints, generate
a second estimate of the practicability of a second set of experiments defined by the set of experimental parameters defined
by the parameters, the sampling patterns and the constraints, including the modified parameters, sampling patterns or constraints,
the second estimate including a count of the second set of experiments, and provide the second estimate to the user, and
in response to input approving of the second estimate, generate an experiment design defined by the parameters, the sampling
patterns and the constraints, the experiment design including data representing a final set of experiments, the data including
a plurality of sets of coordinates in the hyperspace defined by the set of experimental parameters, each of the sets of coordinates
defining an experiment in the final set of experiments.
|