darcs-hash:20100602003632-c8de0-b834db82c3ae5ccebd59fb656dd52f6d13f3bdb1.gz

darcs-hash:20100601224759-c8de0-551289eff93a76b24740d2a8e5bea2147202e1ab.gz

darcs-hash:20100601151731-c8de0-094eca40a2ca0b04dbb029cb111b2284ede822da.gz

darcs-hash:20100531214914-c8de0-0bd388d69350800bad22d18211e5def40a0121e1.gz

darcs-hash:20100531165910-c8de0-347b8ba2ff62d825467d4baf6b21508f5364d1e0.gz

use thread-local storage for Mersenne twister to make thread-safe (for compilers supporting __thread or __declspec(thread))

darcs-hash:20100528214112-c8de0-4ca1313eec0a24eb1bc4e0b4aea0bc0c93c75788.gz

darcs-hash:20100528164440-c8de0-870eb4a31ef2c3a9f0633b8be912b0f2c3156b70.gz

darcs-hash:20100528163737-c8de0-c889ae8574dbe08c7bd3e5af247574f8a3fc0c3e.gz

darcs-hash:20100528163600-c8de0-aa1d026191abea2ffd7870265933673fe18ab890.gz

darcs-hash:20100512151812-c8de0-67f64550b5b5fa892638c3f1111dc1179755b71e.gz

darcs-hash:20100512150153-c8de0-b50f3e7a70cd3b834eef51b8b3123df0ee83dacb.gz

darcs-hash:20100512040009-c8de0-edd99ecc6904eca7c0fbb4d42a263b0ddb8cd00e.gz

darcs-hash:20100512034848-c8de0-7a7f45a53656b351d4372ed22104bab823569cf7.gz

replace Octave nlopt_minimize_constrained with wrapper around new nlopt_optimize, and document the latter

darcs-hash:20100512021454-c8de0-d8c1513f66476aa316e32561e5dc648aca70e06f.gz

darcs-hash:20100512021430-c8de0-95df1ca4c461239c6ddbd8cc0e87104d0101039f.gz

darcs-hash:20100512021351-c8de0-47f32de74b4d6dbe029d07fd0f1356b4408b5dd3.gz

darcs-hash:20100508001255-c8de0-0cffc32a0d6e50ab65715676ad2b97df1f41f5d3.gz

darcs-hash:20100507234805-c8de0-7419b76ac7a7dde36b4c7cc380ea3be74ba583ef.gz

darcs-hash:20100406025026-c8de0-653797e7b13d5e004082e273187d950c8db7e436.gz

darcs-hash:20100406024933-c8de0-f727831d86278473c905153e5bb105082705804d.gz

darcs-hash:20100405191749-c8de0-2bb5cb5b80befa74a8a666be82866d91817a9153.gz

darcs-hash:20100405072811-c8de0-307c0f8c22e707816cbe2da178b2ff92fc6a596f.gz

darcs-hash:20100405071953-c8de0-807cc38ca631b331de47a1b10fc839a5ad273889.gz

darcs-hash:20100405071650-c8de0-c7d3b05da048406f5004a4c9c46dccd6d2a41100.gz

darcs-hash:20100405063016-c8de0-5560baf07043e1495a4e1315acbc280f6105af39.gz

darcs-hash:20100405055831-c8de0-639bc4d71be55d9df49cc96a5df79352d3c7259e.gz

darcs-hash:20100405054136-c8de0-a106e736e68aa365986a1cc91b9faf8492dd6ea1.gz

darcs-hash:20100405044954-c8de0-9c3737452317f7a0619d9acff42fd48a847ef894.gz

darcs-hash:20100405044917-c8de0-be7df864ba060befce69c335e9116b01a2f82f10.gz

darcs-hash:20100405013443-c8de0-811bb252f87584b2af1919a53eab9566096b5829.gz

darcs-hash:20100405005632-c8de0-ed92251161dc73aa8ac61a40e0cd2395f266c351.gz

darcs-hash:20100405005422-c8de0-9fee12e2bca611ba6c4662af0cf2ab49b14a2749.gz

pseudo-randomize simplex steps in COBLYA, to avoid accidentally taking steps that don't improve conditioning (which seems to happen with bound constraints where the optimum is against the bounds, sometimes); note that the algorithm remains deterministic

darcs-hash:20100405003452-c8de0-af9df1f48a6f2adb1d3990aa66320235ee556cab.gz

allow COBYLA to increase trust-region radius (rho) if predicted improvement was approximately right and simplex is ok, following suggestion in SAS manual for PROC NLP

darcs-hash:20100405003344-c8de0-d288279ade2dfafe7b9bcf9facf7807288f28010.gz

darcs-hash:20100405003328-c8de0-cc8cf4dc7e7e5d46ec7e939f38317cca27c59f34.gz

darcs-hash:20100404222432-c8de0-d8e9e542f9b3e8add61ec28c5346858e37bd3577.gz

darcs-hash:20100404221835-c8de0-748279ced777a37c6a5ada00892bba75da7cdd54.gz

darcs-hash:20100404215125-c8de0-c139af2024d66a76717073417d5c15c7db314647.gz

darcs-hash:20100404214851-c8de0-5e39ad9719b7c5068a704a83e4907e98950c4065.gz

darcs-hash:20100404203147-c8de0-fb83a312980103cc419a5542b2b20faa9a224518.gz