Legend:
- Unmodified
- Added
- Removed
-
DEV_TODO
r112 r240 1 1 DIRECT: 2 2 * Add more python optimizers: scipy, OpenOpt, PARK (snobfit) 3 * Genertate python bindings to DAKOTA; add DAKOTA optimizers4 3 * Allow for derivative and gradient capture -> use Sow? 5 * Update the parallel branch (pyina), get "handler" to work in parallel (ipd?)4 * get "handler" to work in parallel 6 5 * Better 'programmatic' interface for handler 7 6 * Add more options to handler (i.e. toggle_verbosity?, get_cost?, ...?) 8 7 * Allow sigint_callback to take a list (i.e. provide call[i]) 9 8 * Add "constraints" to models (design similar to pyre.inventory and validators) 10 * Extend parallel launcher and communication to distributed services (gsl & idd)11 9 12 10 INDIRECT: 13 11 * Build a failure test suite, and a proper test suite 14 12 * Try one of the VTF apps... or Sean's "cain" library 15 * Clean-up and package "gsl infect" core utilities16 13 17 14 REFERENCE: … … 20 17 * <-- Find OpenOpt's model & optimizer API --> 21 18 * <-- Find DAKOTA's model & optimizer API --> 22 23 (DONE or MOSTLY DONE):24 * Standardize the interfaces, also organize imports better25 * Modify Sow to accept n-inputs, instead of just two inputs
Note: See TracChangeset
for help on using the changeset viewer.