Timeline
04/20/08:
- 23:08 Ticket #111 (Check if formal-actual name translation is implemented correctly in ...) created by
- 23:04 Ticket #110 (Rename SinglePointCrossover (for GP) to size-fair crossover) created by
- related to #108
- 21:36 Ticket #109 (GreaterThan function (in symmetry with LessThan)) created by
- 13:11 Ticket #108 (Homologous and size-fair crossover operators for GP) created by
- 12:46 ModularizationAndReuseWithOperatorLibraries edited by
- added description of import-operator idea (diff)
- 12:31 ModularizationAndReuseWithOperatorLibraries created by
- created a page about modularization and reuse with description and the …
- 12:03 Ticket #107 (Improve capabilities for modularization and operator reuse) created by
- OperatorLibraries are meant as a way to enable modularization and …
- 11:53 Ticket #106 (Programmable functions for GP (similar to ProgrammableOperator)) created by
- User programmable functions for GP would increase flexibility of GP …
- 11:29 Ticket #32 (Show number of available updates after 'update' action in plugin manager) closed by
- fixed: fixed with r140 in trunk and r141 in 3.0 branch
- 11:28 Changeset [141] by
-
merged new feature: "display number of installed, new, upgradeable plugins in pluginmanager" (#32, r140) from trunk into 3.0 branch
- 11:26 Changeset [140] by
-
added a statusbar for the plugin-manager and display the number of installed, available, and upgradeable plugins. (ticket #32)
- 10:59 Changeset [139] by
-
merged serialization bug-fix (r138) from trunk into the 3.0 branch
- 10:57 Ticket #102 (Investigate if all numbers are serialized with InvariantCulture ...) closed by
- fixed: fixed with r138
- 10:56 Changeset [138] by
-
fixed serialization number-format of double values in DoubleBoundedConstraint and NormalDistributedRandom (ticket #102)
04/19/08:
- 15:36 Changeset [137] by
-
merged bugfixes from r135 and r136 from the trunk into the 3.0 branch
- 15:33 Ticket #29 (Implement premature stop-criterion for evaluation operators) closed by
- fixed: Finished implementation of early stopping criterion in MSE-evaluator. …
- 15:30 Changeset [136] by
-
fixed a bug in the MSE evaluator which uses an early stopping criterion.
(ticket #29)
- 14:46 Ticket #105 (Fix quality-variable formalname translation in GPEvaluatorBase) closed by
- fixed: fixed with r135
- 14:46 Changeset [135] by
-
fixed #105
- 14:19 Ticket #105 (Fix quality-variable formalname translation in GPEvaluatorBase) created by
- 12:58 Ticket #54 (Create a HL 3.0 Release branch) closed by
- fixed: Done in r134
- 12:57 Changeset [134] by
-
#58: Created a release branch for HeuristicLab 3.0
- 12:29 Ticket #104 (Plugin for local sqlite databases) created by
04/18/08:
- 15:16 Changeset [133] by
-
changed GP 'terminal symbol' Variable to return NaN for index outside the range of the dataset (for instance when negative time-offsets are allowed and a model is evaluated for the first sample)
- 15:15 Changeset [132] by
-
added methods for GetRange and GetMean that use the full set of rows instead of start and end points (semantically coupled with r128) (ticket #29)
- 15:07 Changeset [131] by
-
changed the StructIdProblemInjector to inject only the training data (between TrainingSamplesStart and TrainingSamplesEnd). As a consequence the two variables start and end are not needed and have been removed.
- 15:05 Changeset [130] by
-
fixed the return value for early stops (estimated final MSE)
(ticket #29)
- 14:11 Changeset [129] by
-
fixed variable type of QualityLimit for the early stopping MSE-evaluator
- 13:53 Changeset [128] by
-
- created abstract base class for GP evaluators
- created a version of MSEEvaluator that implements an early stopping criterion (to be combined with offspring selection)
(ticket #29)
04/16/08:
- 12:59 Changeset [127] by
-
fixed unwanted change of local installation path
04/15/08:
- 20:56 Changeset [126] by
-
fixed two bugs in the serialization of DistributedEngine and ProcessingEngine
- 16:41 Changeset [125] by
-
created a branch that combines the XmlReader and XmlWriter branches
- 16:15 Changeset [124] by
-
fixed a problem that was caused by serializing a sub-element (constraints) before serializing attributes
- 15:57 Changeset [123] by
-
fixed more bugs (not thoroughly tested but at least it works for OSGP_NOx and OSGA_TSP)
- 14:17 Changeset [122] by
-
bug fixes to make loading of OSGA-TSP work. Some non-working code remains
04/14/08:
- 21:51 Changeset [121] by
-
created a branch that uses XmlTextReader instead of XMLDocument to load documents. Investigating ticket #103. (...work in progress!)
- 17:59 Changeset [120] by
-
fixed a typo that caused an error in save/restore of operator JSSPInjector
- 17:49 Changeset [119] by
-
created a branch that uses XmlTextWriter instead of XMLDocument to save documents. Investigating ticket #103.
- 15:42 Ticket #103 (Serialization of large engines should use less RAM) created by
- When serializing and deserializing large engines HL allocates a lot of …
- 15:30 Ticket #102 (Investigate if all numbers are serialized with InvariantCulture ...) created by
- especially: * DoubleBoundedConstraint * NormalDistributedRandom
- 10:19 Ticket #100 (PluginManager uses the active path as the location to store downloaded ...) closed by
- fixed: fixed with r118
- 10:18 Changeset [118] by
-
fixed #100 by using an absolute path for the plugin-download location
04/13/08:
- 16:52 Ticket #101 (Caching scheme to boost performance of GP evaluation) created by
- Caching the return values of tree branches could improve GP evaluation …
04/08/08:
- 22:03 Ticket #100 (PluginManager uses the active path as the location to store downloaded ...) created by
- This is a serious error because the .NET framework remembers the …
- 21:46 Changeset [117] by
-
ticket #98 also removed call to Compile() in the Populate() method since it slowed down the deserialization needed for the parallel execution horribly and it's not needed in the lazy compilation scheme anyway (also see r116)
- 16:09 Ticket #99 (Possibility to publish multiple plugins in one action) created by
- Publishing many plugins is cumbersome. It would be nice to have an …
- 15:02 Ticket #98 (Emtpy method of ProgrammableOperator is compiled on every clone operation) closed by
- fixed: fixed with r116 as suggested by pspitzli by lazily compiling the …
- 15:00 Changeset [116] by
-
fixing #98 by compiling the method in the programmable operator lazily in the apply() method. as suggested by pspitzli
- 14:49 Changeset [115] by
-
improved the comment of the OS.version check
- 14:44 Ticket #96 (GridClient / GridServer doesn't work in Vista (inconsistent result of ...) closed by
- fixed: applied patch with r114
- 14:44 Ticket #97 (Multiple instances of GridClient on one machine) closed by
- fixed: patch applied with r114
- 14:44 Changeset [114] by
04/07/08:
- 22:29 Ticket #98 (Emtpy method of ProgrammableOperator is compiled on every clone operation) created by
- This is potentially harmful (at least for efficiancy reasons) in a GP …
- 22:26 Ticket #97 (Multiple instances of GridClient on one machine) created by
- Could be useful to allocate both cores on a dual-core machine. (first …
- 22:22 Ticket #96 (GridClient / GridServer doesn't work in Vista (inconsistent result of ...) created by
- GridClient and GridServer do not work in Windows Vista (TM) …
