Difference between revisions of "Extending"
From SUMOwiki
Jump to navigationJump to search| Line 16: | Line 16: | ||
* [[Add Distributed Backend|Add a distributed backend to run simulations in parallel]] | * [[Add Distributed Backend|Add a distributed backend to run simulations in parallel]] | ||
* [[Add Profiler|Add a profiler to monitor some aspect of the modeling]] | * [[Add Profiler|Add a profiler to monitor some aspect of the modeling]] | ||
| − | * [[ | + | * [[Optimizer|Add a new general purpose optimization algorithm]] |
== Tips == | == Tips == | ||
* Note that for some problems the outputs may be complex, so if your model type does not support complex numbers you need to ensure complexHandling is not set to complex in the config file and that you give an error if this is not the case (see ANNModel for an example). | * Note that for some problems the outputs may be complex, so if your model type does not support complex numbers you need to ensure complexHandling is not set to complex in the config file and that you give an error if this is not the case (see ANNModel for an example). | ||
* Please use the logging framework for all output! Look at the source code to see how we used it. You can find more details on the logging framework in the java.util.logging package documentation. | * Please use the logging framework for all output! Look at the source code to see how we used it. You can find more details on the logging framework in the java.util.logging package documentation. | ||
Revision as of 15:40, 23 February 2009
The SUMO Toolbox can be extended in many ways.
If you simply want to model your own problem see the Adding an example page.
Prerequisites
First make sure you have read and understand the coding guidelines and the toolbox internals pages. You may also want to refer to the Object Oriented Programming in matlab page .
Component
What do you want to do?
- Add a new model type
- Add an optimization algorithm to optimize the model parameters
- Add a new initial experimental design
- Add a sample selection algorithm
- Add a new model selection criteria
- Add a distributed backend to run simulations in parallel
- Add a profiler to monitor some aspect of the modeling
- Add a new general purpose optimization algorithm
Tips
- Note that for some problems the outputs may be complex, so if your model type does not support complex numbers you need to ensure complexHandling is not set to complex in the config file and that you give an error if this is not the case (see ANNModel for an example).
- Please use the logging framework for all output! Look at the source code to see how we used it. You can find more details on the logging framework in the java.util.logging package documentation.
